Wednesday

Wednesday, 8:00am - 9:30am 

■ WA01 

C - Ballroom D1, Level 4 

Multi-stage Stochastic Optimization Applied to 

Energy Planning 

Sponsor: Energy, Natural Resources and the Environment/ Energy 

Sponsored Session 

Chair: Steffen Rebennack, Assistant Professor, Colorado School of Mines, 

Division of Economics & Business, Boulder, CO, United States of America, 

steffen@ufl.edu 

1 - Quasi-Monte Carlo Methods for Hydroelectric Energy Planning 

Tito Homem-de-Mello, University of Illinois at Chicago, Chicago, IL, 

United States of America, thmello@uic.edu, Erlon Finardi, 

Vitor de Matos 

We study a multi-stage stochastic programming model for hydroelectric energy 

planning in Brazil. Sampling techniques are used to generate scenario trees from the 

stochastic process defining the water inflows, and also to select scenarios within that 

tree. We analyze the use of Quasi-Monte Carlo methods for the problem. We also 

discuss statistical performance measures that allow us to compare methods, and 

present numerical results to evaluate the effectiveness of the proposed approach. 

2 - Reservoir Hydropower Operations: Valuing Flexibility 

Stein-Erik Fleten, Professor, NTNU Norway, Department of Industrial 

Economics and Technology Management, Trondheim, Norway, 

Stein-Erik.Fleten@iot.ntnu.no, Martin Prokosh, Camilla Kolsrud 

In this talk we provide insights into how storage flexibility impacts the expected 

revenue of hydropower plants. Using daily data going back ten years from 14 

different hydropower plants with significant seasonal reservoir capacity, and who 

operate in a well-functioning market, we conduct an empirical analysis of the 

different factors affect the ability of hydropower producers to exploit high prices. 

Storage flexibility on average accounts for 22% of actual revenues, ranging from 0 

to 40%. 

3 - Optimal Control of Energy Storage using the Knowledge Gradient 

with Nonparametric Beliefs 

Warren Powell, Professor, Princeton University, Sherrerd Hall, 

Princeton, NJ, 08544, United States of America, 

powell@princeton.edu, Emre Barut 

We consider different natural gas and pumped hydro storage problems as stochastic 

control problems which can be solved using tunable policies governed by tunable 

parameters. We present a novel stochastic search algorithm using the knowledge 

gradient adapted to nonparametric beliefs, which produces policies that are easy to 

implement. We demonstrate that the logic produces policies that slightly outperform 

actual performance. 

4 - Decomposition Approach for G-T Expansion Planning with Implicit 

Multipliers Evaluation 

Fernanda Thome, PSR, Rio de Janeiro, Brazil, fernanda@psr-inc.com, 

Marcia H.C. Fampa, Luiz Carlos da Costa Jr., Silvio Binato 

Algorithms solving stochastic hydrothermal operation problems usually take 

computational advantages in the elimination of constraints whose explicit 

representation does not affect the problem’s optimal solution. This work presents a 

new methodology for solving generation-transmission expansion planning problems 

based on Benders decomposition technique and the evaluation of the Lagrange 

multipliers associated to those non-explicit constraints and required in the 

construction of the Benders cuts. 

■ WA02 


Refinery Operations with Spot and Forward Markets 

Cluster: Energy: Modeling the Interface Between Markets and 

Operations 

Invited Session 

Chair: Simin Huang, Tsinghua University, Department of Industrial 

Engineering, Beijing, China, huangsimin@mail.tsinghua.edu.cn 

1 - Considering the Effect of Outside Options in the Capacity Planning 

for Hydrogen Fueling Station 

Ruwen Qin, Assistant Professor, Missouri University of Science and 

Technology, Department of Engineering Management, Rolla, United 

States of America, qinr@mst.edu, Scott Grasman, Kevin Martin 

We model and analyze the effect of an outside option in determining the optimal 

capacity for hydrogen fueling stations. Through assessing the economic consequence 

of the decision, this study suggests opportunities for gaining additional profits. 

INFORMS Austin – 2010 WA03 

361 

2 - Financial Engineering for Refinery Operations: Challenges 

and Opportunities 

Li Zheng, Professor, Tsinghua University, Department of Industrial 

Engineering, Beijing, 100084, China, lzheng@tsinghua.edu.cn 

Over the past decade, the world has witnessed the extreme price volatility from 

both crude supply and final product market. Financial engineering tools have been 

used to hedge the financial risk in refinery operations recently. In this talk, we 

present the challenges of the problem and provide some potential ideas for future 

research. 

3 - Integrated Financial and Operational Model for Crude Oil 

Procurement in Refineries 

Zhen Liu, Assistant Professor, Missouri University of Science & 

Technology, United States of America, zliu@mst.edu, Simin Huang 

The world seems to have entered into an era of higher crude oil price volatility. As 

the crude oil cost is about 90% of the refinery input cost, there has been significant 

volatility in the margins and profitability of any petroleum refinery. An integrated 

financial and operational model is developed to hedging the financial risk. 

4 - Financial Engineering Model for Crude Transportation in Refineries 

Zhihai Zhang, Associate Professor, Tsinghua University, Department 

of Industrial Engineering, Beijing, 100084, China, 

zhzhang@tsinghua.edu.cn 

The crude transportation cost in refineries could be up to several billion dollars and 

the volatility of freight rates can be substantial. A decision-making model for 

optimal oil tanker selection procedure is developed to help refining companies 

manage the freight market risk. 

■ WA03 


Challenges for the US Biofuels Industry: Economic and 

Technological Uncertainties 

Sponsor: Energy, Natural Resources and the Environment//Forestry 


Chair: Hayri Onal, University of Illinois at Urbana-Champaign, Dept 

Agricultural and Consumer Economics, Urbana, IL, 61801, United States 

of America, h-onal@illinois.edu 

1 - Modeling Uncertainty in Biomass Greenhouse Gas Emissions with 

the Calculating Uncertainty in Biomass 

Aimee Curtright, Physical Scientist, RAND Corporation, Pittsburgh, 

PA, 15213, acurtrig@rand.org, Henry Willis, David Johnson, 

Costa Samaras 

The greenhouse gas (GHG) intensity of biofuels depends on how feedstocks are 

produced, transported, and processed. This paper will describe the Calculating 

Uncertainty in Biomass Emissions (CUBE) model, a tool developed for the National 

Energy Technology Laboratory to examine uncertainties in bio-feedstock GHG 

estimates. The limits on the precision of results, the value of additional emissions 

information, and the sources and magnitude of uncertainty will be discussed. 

2 - Ecosystem Costs in a Logistical Model of Cellulosic 

Ethanol Production 

David Lambert, Professor and Head, Department of Agricultural 

Economics, Kansas State University, lambertd@k-state.edu, 

Jason Bergtold, Elizabeth Canales 

Network models of biomass use for cellulosic ethanol production often ignore 

ecosystem opportunity costs. Building upon an existing MIP model, we incorporate 

ecosystem opportunity costs arising from soil erosion, loss of organic and inorganic 

matter, and carbon sequestration values associated with crop residue and energy 

crop harvest for ethanol production. 

3 - Strategic Biofuel Supply Chain Planning Under Supply, Demand, 

and Technology Uncertainties 

Yueyue Fan, University of California-Davis, Davis, CA, 95616, United 

States of America, yyfan@ucdavis.edu 

This talk focuses on modeling and computational challenges in strategic biofuel 

supply chain planning under supply, demand, and technology uncertainties. Using a 

case study based on California settings, the economic feasibility, infrastructure 

requirements, and the environmental impact of converting biowastes to fuel are 

analyzed.

WA04 

4 - Projections for US Flex-fuel Vehicle Structure and Renewable 

Fuel Standards 

Xirong Jiang, Senior consulting decision analyst, Lumina Decision 

Systems, Inc, 26010 Highland Way, Los Gatos, CA, 95033, United 

States of America, xirong@lumina.com, Surya Swamy, Max Henrion, 

Costa Samaras 

We use ATEAM (Analytica Transportation Energy Assessment Model) to explore a 

variety of scenarios to see how rapidly the US needs to adopt flex-fuel vehicles to 

consume the volume of biofuels productions set by the 2010 revision of the 

Renewable Fuel Standard, including high-blend E15 or E20 options. 

■ WA04 


Decision Analysis II 

Contributed Session 

Chair: John Mamer, UCLA Anderson Grad. School of Mgmt., 

110 Westwood Plaza, D518, Los Angeles, CA, 90095-1481, 

United States of America, jmamer@anderson.ucla.edu 

1 - Value of Information in Spreadsheet Monte Carlo Simulation Models 

Mike Middleton, Decision Toolworks, 2105 Buchanan St, San 

Francisco, United States of America, Mike@DecisionToolworks.com 

For a spreadsheet planning model with uncertain inputs, value of information about 

each input is useful for evaluating information-gathering efforts and for comparing 

their importance. This paper describes non-macro computation methods for 

spreadsheet Monte Carlo simulation, calculation of value of information for each 

uncertain input, charts for presenting the results, and insights to be gained. 

2 - The Role of Supply Chain Structure in the Food vs. Biofuel Tradeoff 

Adaora Okwo, Georgia Institute of Technology, 765 Ferst Dr., 

Atlanta, United States of America, aokwo@gatech.edu 

We present a micro model of the food vs biofuel tradeoff. Most macro models 

informing the policy discussion fail to incorporate contracting and downstream 

market power in the agricultural supply chain; both of which can significantly 

influence a farmer’s decisions on which crops to supply and in which quantities. We 

present a two-echelon decentralized supply chain model to illustrate the impact of 

contract parameters and market power on the equilibrium supply response for food 

and energy crops. 

3 - Corn Ethanol Plant Investment using Real Options Analysis 

Dexin Luo, Geogia Institute of Technology, 765 Ferst Drive, Atlanta, 

GA, United States of America, dexin.luo@gatech.edu, Valerie Thomas 

We apply real options analysis of entry-exit decision to corn ethanol plants with two 

different processes: dry-milling and wet-milling. We incorporate uncertainties 

regarding ethanol and corn prices and technical change. The study shows the 

influence of technical change on the investment decision of individual firms with 

comparison to the increase of dry-milling plants in the U.S. 

4 - Dealing with the Growth of Knowledge 

Norimasa Kobayashi, Assistant Professor, Tokyo Institute of 

Technology, 2-12-1 Ookayama, Meguro-ku, Tokyo, 152-8552, Japan, 

nkoba@valdes.titech.ac.jp 

Binmore (2009) criticizes that our real decision environments are so essentially 

“large” that the complete state space assumption of Bayesianism does not hold. I 

model the growth of knowledge formalizing Popper (1935), and discuss how well 

the decision on incomplete decision analytic models perform in different 

circumstances. Particularly, I discuss that in non-cooperative games, incomplete 

mental models may result both in inefficiency and efficiency. 

5 - Fire Sales and Search 

John Mamer, UCLA Anderson Grad. School of Mgmt., 110 

Westwood Plaza, D518, Los Angeles, CA, 90095-1481, United States 

of America, jmamer@anderson.ucla.edu, Steven Lippman 

We study a model of asset sales via search with semi-rational buyers. A seller has a 

finite (or infinite) number of items to sell, independent buyers arrive according to a 

Poisson process. Each potential buyer knows the price of the last sale, and offers the 

minimum of his reservation value and the price of the last sale. As a result, the 

seller faces a falling offer distribution, each sales price setting the maximum offer for 

subsequent sales. 

INFORMS Austin – 2010 

362 

■ WA05 


Multicriteria Decision Making 


Chair: Judit Lienert, Dr., Eawag: Swiss Federal Institute of Aquatic 

Science and Technology, Ueberlandstrasse 133, P.O. Box 611, 

Duebendorf, CH-8600, Switzerland, judit.lienert@eawag.ch 

1 - A Method for Multiobjective Optimization using Trust 

Region Method 

Jong-hyun Ryu, Purdue University, 315 N. Grant Street, West 

Lafayette, IN, 47907, United States of America, ryuj@purdue.edu, 

Sujin Kim 

We propose a method for approximating the Pareto front in a blackbox 

multiobjective problem. At each iteration, each objective function on a certain 

region (trust region) is approximated by a quadratic function, and a scalarization 

method is applied to collect points to approximate the Pareto front. The region is 

iteratively updated so as to maintain the spread of solutions. Numerical results are 

presented to demonstrate the effectiveness of the proposed algorithm. 

2 - Evolutionary Computation-based Multi-objective Approach to 

Rehabilitate Interconnected Infrastructure 

Avery White, Student, Texas A&M University, 3136 TAMU, College 

Station, TX, United States of America, sacredfaith@tamu.edu, 

Emily Zechman, Lufthansa Kanta, Alex Sprintson 

In the event of urban fires, cascading failures between water and electrical 

distribution infrastructure systems may exacerbate municipal losses. Under-designed 

water distribution systems may be further disabled through power loss. This 

research takes an evolutionary computation-based approach to identify pipe 

replacement strategies for urban fire scenarios to provide fire flows, maintain water 

quality for normal operating conditions, and minimize pipe replacement costs. 

3 - Application of Dynamic Multiobjective Programming to Supply 

Chain of Crude Oil 

Moses Olusola Okesola, Student, University of South Africa, 

9,Solomon Okonkwo Street, Unity Estate, Egbeda, Lagos, 234, 

Nigeria, okesolaj@yahoo.com 

The Dynamic Multiobjective Programming (DMP) has been largely deficient in 

decision problems related to integrated supply chain of crude oil in literature.We 

tend to review developed methods for solving MOPP using DP technique and 

identify their deficiencies.The research will focus on upstream supply chain of crude 

oil in Nigeria by treating each activity along the supply chain separately with the 

grand objective to optimize return function from one stage to another. 

4 - Reducing Pharmaceuticals in Hospital Wastewater - MCDA Multistakeholder 

Elicitation Challenges 

Judit Lienert, Dr., Eawag: Swiss Federal Institute of Aquatic Science 

and Technology, Ueberlandstrasse 133, P.O. Box 611, Duebendorf, 

CH-8600, Switzerland, judit.lienert@eawag.ch, Peter Reichert, 

Nele Schuwirth 

Pharmaceuticals in water bodies are of concern; they can be reduced by point 

source measures. In a project with engineers, natural, and social scientists, we 

studied two exemplary hospitals. We used MCDA to support the decision between 

68 technological and organizational alternatives to reduce medicals. We elicited 

preferences from 26 stakeholders. Our elicitation procedure reduces the time 

demand, but remains methodologically satisfactory. We present elicitation challenges 

and main results.

■ WA06 

C - Ballroom E, Level 4 

Tutorial: New Developments for Solving Real World 

Optimization Problems by Marrying Simulation 

and Optimization 

Cluster: Tutorials 


Chair: Fred Glover, CTO, OpTek Systems, Inc., 1919 Seventh Street, 

Boulder, CO, 80302, United States of America, glover@opttek.com 

1 - New Developments for Solving Real World Optimization Problems 

by Marrying Simulation and Optimization 

Manuel Laguna, Professor, University of Colorado at Boulder, 

419 UCB, Boulder, CO, 80309, United States of America, 

laguna@colorado.edu, Jay April, Marco Better 

Companies invest billions of dollars each year in applications that can be handled by 

combining simulation and optimization. These notably include problems that 

involve uncertainty and complex nonlinearities, as in the areas of capital 

investment, workforce composition and management, energy resource and 

transmission planning, health care systems, financial portfolio optimization, 

production and inventory systems, security and emergency response planning. We 

identify latest advances and applications in these areas from combining simulation 

and optimization, together with opportunities for future applications. 

■ WA07 

C - Ballroom F & G, Level 4 

Supply Chain Optimization I 


Chair: Shuang Chen, PhD Candidate, University of Florida, 285 Corry 

Village Apt 13, Gainesville, United States of America, scljj@ufl.edu 

1 - A News-Vendor Model With External Fund Availability 

Benjamin Melamed, Professor, Rutgers Business School - Newark 

and New Brunswick, 94 Rockafeller Rd., Piscataway, 08554, 

United States of America, melamed@rbs.rutgers.edu, Junmin Shi, 

Michael N. Katehakis, Ben Sopranzetti 

The classical news-vendor (NV) problem is to find the optimal order quantity which 

maximizes the expected profit in a probabilistic demand framework. In this talk we 

present studies when there is external funding available. We treat the corresponding 

optimization problem as a capital-asset portfolio problem, and obtain the optimal 

ordering strategy. In addition, some risk issues, such as bankruptcy risk, have are 

discussed. 

2 - A Model for Planning and Operating the Norwegian Seafood 

Value Chain 

Peter Schütz, SINTEF Applied Economics, S.P. Andersens vei 5, 

Trondheim, Norway, peter.schutz@sintef.no, Kristin Uggen, 

Kjetil Midthun 

We discuss a model for planning the slaughtering and processing of farmed salmon. 

The model also includes the operations of the well boats, such as loading conditions 

at the cages and cleaning before assigning them to new regions. The problem is 

subject to uncertainty, as the number of fish in a cage, their weight and size can 

only be estimated, but is unknown until the fish is slaughtered. 

3 - The Optimum Base-Stock Levels in a Two-Echelon Supply Chain 

with Service Level Constraints 

Yat-wah Wan, National Dong Hwa University, Institute of Logistics 

Management, Shou-Feng, Hualien, 974, Taiwan - ROC, 

ywan@mail.ndhu.edu.tw, Tsung-Shung Chang 

In a two-echelon supply chain of non-zero replenishment lead times, retailers adopt 

base-stock policies and set minimum service levels. The objective is to find globally 

optimal base-stock inventory levels of the whole chain. Such levels are found from 

(i) the sample-path monotone properties of ordered quantities and of inventories on 

hand with respective to base-stock levels, and (ii) in some cases, the local optima of 

the objective function are decreasing with respect to base-stock levels. 

4 - Optimizing the Kenya Coffee Supply Chain 

Rose Karimi, Rutgers University, 1 Washington Park, Newark, NJ, 

07102, United States of America, kiwanuka@pegasus.rutgers.edu, 

Yao Zhao 

We compare the profitability of two models: the inventory control model and the 

selling-through model.In the first model based on current price information, a 

decision is made to either sell all or a portion of the coffee, or to hold the coffee in 

expectation of a more favorable price in the future while in the second model all 

coffee is sold. 


363 

5 - Case Pack Configuration and Procurement Planning 

Shuang Chen, PhD candidate, University of Florida, 285 Corry 

Village Apt 13, Gainesville, United States of America, scljj@ufl.edu, 

Joseph Geunes 

We consider a retail planning problem where retailers must order in case packs 

containing multiple individual products. We simultaneously consider case pack 

configuration and procurement decisions. We first solve this quadratically 

constrained quadratic program with integer variables using an exact linear method. 

Then we propose an integrated approach combining module design and lot sizing 

decisions, as well as an iterative heuristic. Our approaches perform very well 

compared to solver Baron. 

■ WA08 

C - Room 11A, Level 4 

Joint Session Location Analysis/ MIF: Public-Sector 

Facility Location 

Sponsor: Location Analysis/ Minority Issues 


Chair: Michael Johnson, Associate Professor, University of Massachusetts 

Boston, Department of Public Policy/Public Affairs, 100 Morrissey Blvd., 

Boston, MA, 02125-3393, United States of America, 

michael.johnson@umb.edu 

1 - A Multi Objective Available Coverage Model 

Hari Rajagopalan, Assistant Professor, Francis Marion University, 

School of Business, P.O. Box 100547, Florence, SC, 29501, 

United States of America, hrajagopalan@fmarion.edu, Cem Saydam, 

Elizabeth Sharer, Kay Lawrimore 

Demand for ambulances fluctuates spatially and temporally. Recent advances in 

computing and spatial data have enabled EMS managers to practice dynamic 

redeployment plans. In this paper we address the issue of redeployment by explicitly 

considering the number of redeployment trips to be made while meeting the 

coverage requirements with nearly minimal fleet size and develop fast heuristics. 

We present computational statistics using real data from Charlotte, NC. 

2 - Equity Across Groups in Facility Location 

Tammy Drezner, Professor, California State University, Fullerton, 800 

State College Blvd., Fullerton, CA, 92834, United States of America, 

tdrezner@Exchange.fullerton.edu, Zvi Drezner 

An equity model between groups of demand points is proposed. The set of demand 

points is divided into two or more groups. For example, rich neighborhoods and 

poor neighborhoods, urban and rural neighborhoods. We wish to provide equal 

service to the different groups by minimizing the deviation from equality among 

groups. The objective function, to be minimized, is the sum of squares of differences 

between all pairs of service distances between demand points in different groups. 

3 - Solving a Multi-Period School Location Problem with Capacity 

Constrains using Tabu Search 

Eric Delmelle, Assistant Professor, University of North Carolina at 

Charlotte, Department of Geography and Earth Scienc, Charlotte, 

NC, 28223, United States of America, Eric.Delmelle@uncc.edu, 

Jean-Claude Thill 

In rapidly expanding areas, it may be necessary to build additional schools to meet 

anticipated demand. A tabu search algorithm is used to solve a multi-period 

capacitated p-median model, applied to a a school network location problem. The 

model is flexible as it allows facility closure. 

4 - Foreclosed Housing Selection using Multi-Criteria Decision Models 

Michael Johnson, Associate Professor, University of Massachusetts 

Boston, Department of Public Policy/Public Affairs, 100 Morrissey 

Blvd., Boston, MA, 02125-3393, United States of America, 

michael.johnson@umb.edu, David Turcotte, Rachel Drew 

Acquisition of foreclosed housing for redevelopment is a key element of U.S. 

housing policy. This task requires balancing multiple criteria and assessing decisionmaker 

preferences, which can be difficult to quantify. We describe a multi-criteria 

decision model for acquisition of real-estate-owned foreclosed housing in which we 

adapt methods from stochastic processes, urban economics and facility location to 

rank candidates for acquisition by a community-based organization.

WA09 

■ WA09 

C - Room 11B, Level 4 

Empirical Studies in Operations Management 

Sponsor: Manufacturing and Service Operations Management 


Chair: Serguei Netessine, Professor, INSEAD, Boulevard de Constance, 

Fontainebleau, 77305, France, serguei.netessine@insead.edu 

1 - Global Sourcing and Operational Performance 

Karan Girotra, INSEAD, Boulevard De Constance, Fontainebleau, 

France, Karan.GIROTRA@insead.edu, Marcelo Olivares 

This study aims to provide the first large-scale empirical estimates on the 

consequences of global sourcing on firm-level inventory performance. We compile a 

novel data-set by merging data from US customs manifests and public data on firmlevel 

inventory performance. We estimate the impact of imports on inventory 

performance and the benefit of oft prescribed operational strategies: importing 

intermediate products (postponement) and importing low variability products. 

2 - The Inventory Billboard Effect 

Gerard Cachon, The Wharton School, 3730 Walnut St., JMHH Suite 

500, Philadelphia, PA, 19104, United States of America, 

cachon@wharton.upenn.edu, Santiago Gallino, Marcelo Olivares 

The challenges associated with identifying an inventory billboard effect are 

discussed. Then, using detailed data from car dealerships, we measure the extent 

that inventory drives sales. 

3 - Organizational Structure, Trust and Sourcing 

Anupam Agrawal, University of Illinois at Urbana-Champaign, 

Wohlers Hall, Champaign, IL, 61820, United States of America, 

anupam@illinois.edu 

This paper focuses on the linkages between the organizational structure of a buying 

firm, its relationships with its suppliers, and the resultant incoming quality of 

components - how do these change dynamically? The research is based on ongoing 

practices at a leading automobile manufacturer. The sourcing related organizational 

arrangements are different in the car and truck making units of this firm, and lead 

to different results on the above two dimensions (quality and relationships). 

4 - An Empirical Analysis of Service-based Strategies in the Automotive 

Industry: The Role of Warranties 

Serguei Netessine, Professor, INSEAD, Boulevard de Constance, 

Fontainebleau, 77305, France, serguei.netessine@insead.edu, 

Morris Cohen, Jose Guajardo 

We empirically analyze the role of warranties as part of the competitive strategy of 

car manufacturers, using data from the US automotive industry. Challenges in 

estimation, as well as implications for firms and consumers are discussed. 

■ WA10 


Operations Economics 



Chair: Fuqiang Zhang, Washington State University in St. Louis, One 

Brookings Drive, St. Louis, MO, 63130, United States of America, 

fzhang22@wustl.edu 

1 - Competition and the Value of Additional Replenishment Opportunity 

Yen-Ting Lin, University of North Carolina, Kenan-Flagler Business 

School, Chapel Hill, NC, United States of America, 

Yen-Ting_Lin@unc.edu, Ali Parlakturk 

We consider a manufacturer serving two competing retailers who sell their products 

during a selling season. The retailers place a regular order before the selling season 

begins. In addition, quick response allows a retailer to place a second order after 

better demand is obtained. We examine the value of this additional ordering 

opportunity for the retailers, manufacturer as well as the whole supply chain. 

2 - Dynamic Price and Lead Time Quotation for MTO Systems with 

Contract Customers and Spot Purchasers 

Baykal Hafizoglu, Arizona State University, Industrial Engineering, 

SCIDSE, 699 S. Mill Ave., #501, Tempe, AZ, 85281, United States of 

America, baykal@asu.edu, Esma Gel, Pinar Keskinocak 

We consider dynamic price and lead time quotation for a MTO company with 

demand from contract customers and spot purchasers. Contract customers are 

offered a uniform price and lead time, and prioritized service. Spot purchasers are 

subject to dynamically quoted price and lead times, which they accept or reject with 

known probability. We discuss the potential of dynamic quotation, various 

properties of optimal control policies and the optimal mix of contract customers and 

spot purchasers. 


364 

3 - Horizontal Alliances and Mergers in Multitier Supply Chains 

Soo-Haeng Cho, Assistant Professor, Tepper School of Business, 

Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, PA, 15215, 

United States of America, soohaeng@andrew.cmu.edu 

Supply chains often consist of multiple tiers in each of which one or more firms 

compete. Firms that belong to the same tier often form alliances or merge together. 

The primary objective of such alliances or mergers is to reduce marginal costs 

through economies of scale in R&D and production. In this paper, we examine the 

effect of the cost-reducing alliances and mergers in one tier on non-participating 

firms in the same tier and on firms in the upstream or downstream tiers. 

4 - Coordinating Capacity Investments in Joint Ventures 

Philippe Chevalier, Professor, CORE, Université catholique de 

Louvain, Voie du Roman Pays 34, Louvain-la-Neuve, 1348, Belgium, 

Philippe.Chevalier@uclouvain.be, Guillaume Roels, Ying Wei 

We model a strategic alliance between several manufacturing firms that decide to 

pool their resources so as to hedge their profits against demand variability. We 

propose a type of contract that coordinates capacity investments and compare this 

contract with several other contracts proposed in the literature. 

■ WA11 


Finance/Operations Link: Flow Models 



Chair: Nico Vandaele, Professor, Katholieke Universiteiet Leuven, 

Naamsestraat 69, Leuven, Belgium, Nico.Vandaele@econ.kuleuven.be 

1 - A Newsvendor Perspective on Value-based Performance and 

Risk Management 

Gerd Hahn, Catholic University of Eichstaett-Ingolstadt, Auf der 

Schanz 49, Ingolstadt, 85049, Germany, gerd.hahn@kuei.de, 

Heinrich Kuhn 

Economic Value Added (EVA) as a prevalent indicator of shareholder value creation 

is applied to the well-known newsvendor or cost-volume-profit model. An 

integrated approach to performance and risk management is developed exploiting 

properties of the EVA concept. We provide a managerial framework for decisionmaking 

considering the risk preference of the newsvendor. A numerical example is 

utilized to highlight implications of the presented approach. 

2 - Working Capital Decisions in Supply Chains Under Consideration of 

Cost of Capital Rates 

Erik Hofmann, Senior Lecturer, University of St. Gallen, 

Dufourstrasse 40a, LOG-HSG, St. Gallen, 9000, Switzerland, 

erik.hofmann@unisg.ch 

In this paper, working capital decisions are extended to supply chains, due to the 

deficiencies resulting from a single-company perspective. The weighted cash 

conversion cycle (WCCC) is combined with the weighted average cost of capital 

(WACC) model. The amount of funds in an inter-organizational setting is 

considered, transforming the WACC from an exogenous into an endogenous 

decision figure. A numerical study illustrates several performance impacts on supply 

chain companies. 

3 - Linking Operations and Finances: The Stochastic Lot 

Sizing Problem 

Lien Perdu, Katholieke Universiteit Leuven, Naamsestraat 69, 

Leuven, 3000, Belgium, lien.perdu@kuleuven-kortrijk.be, 

Nico Vandaele 

In contrast with a traditional cost model, we integrate a financial flow dimension in 

the stochastic lot sizing problem. The original model was built to optimize lead times 

and whereas the new objective function is based on the Economic Value Added 

concept, which allows a broader applicability. 

4 - Analysis of Card Based Flow Control in a Make-to-Order 

Production Shop 

Steven Harrod, Assistant Professor, University of Dayton, 1143 

Ashburton Dr, Dayton, OH, 45459, United States of America, 

steven.harrod@udayton.edu, John Kanet 

We examine the performance of a make-to-order production shop when a “pull” 

regimen (Kanban, Conwip, or POLCA) is enforced. After simulating random job 

routings, we conclude that flow control approaches do in fact reduce the number of 

jobs in process but total system inventory (including ready jobs) increases. Further 

we find that selection of priority rule has a greater influence on shop WIP than 

selection of a particular card based flow control system.

■ WA12 


Tactical and Operational Issues in Supply 

Chain Management 

Sponsor: Manufacturing and Service Operations Management/ 

Supply Chain 


Chair: Sila Cetinkaya, Professor, Texas A&M University, Industrial and 

Systems Engineering, College Station, TX, United States of America, 

sila@tamu.edu 

Co-Chair: James Lavin, NC State University, 1321 Crab Orchard Dr #002, 

Raleigh, 27606, United States of America, jalavin@ncsu.edu 

1 - Uniform vs Retailer-Specific Pricing in a Supply Chain 

Asoo Vakharia, Professor, University of Florida, Department of ISOM, 

Gainesville, FL, 32611-7169, United States of America, 

asoo.vakharia@warrington.ufl.edu, Lan Wang 

Should a supplier adopt a uniform or a retailer-specific price when selling a product 

to retailers with differing capabilities? Would the choice between these pricing 

strategies be moderated by the competitive market structure? Insights into this 

problem are provided for the cases of deterministic and stochastic end-product 

demand. 

2 - An Order-up-to-level (OUL) Inventory Model with Stochastic 

Demand and Lead Times 

Daniel Silva, OR Sr. Analyst, Kimberly Clark, Latin American 

Operations, KR 11A no 94 - 45, Piso 5, Bogota, Colombia, 

daniel.f.silvaizquierdo@kcc.com, Germàn Riaño 

We extend a periodic review, stochastic demand model to include stochastic lead 

times. We assume Normally distributed lead-time demand and solve for expected 

fill-rate, we achieve a better approximation than traditional methods. There is no 

closed form solution for the optimal OUL, but the fill rate function is convex and we 

use iterative methods to solve. Simulation results confirm fill rate goals are met by 

our model, while traditional models over-shoot. Real data results will be presented. 

3 - Stochastic Perturbed Demand Inventory Model 

James Lavin, NC State University, 1321 Crab Orchard Dr #002, 

Raleigh, 27606, United States of America, jalavin@ncsu.edu, 

Anita Vila-Parrish, Russell King 

Stockouts cause customers to lose faith in a retailer and potentially turn elsewhere 

to meet their future demands. Most inventory models use a penalty cost when a 

stockout occurs. An alternative first proposed by Schwartz (1966) instead discounts 

expected demand when stockouts occur through use of a “disappointment factor.” 

We extend Schwartz’s model for the case with stochastic demand. 

4 - A Supply-side Rationale for a Firm to Bundle 

Qingning Cao, PhD Candidate, University of Texas at Dallas, 800 

West Campbell Road, Richardson, TX, 75080-3021, United States of 

America, qxc071000@utdallas.edu, Jun Hang, Kathryn Stecke 

This paper examines a retailer’s two-product bundling decision when the supply of 

one product is limited. This paper derives the retailer’s optimal prices, stocking 

levels, and profits under unbundling and bundling. Demonstrating that limited 

supply can induce the retailer to bundle, this paper highlights a new supply-side 

rationale for bundling. 

■ WA13 


Incentives in Service Operations 


Service Management Special Interest Group 


Chair: Nitin Bakshi, Assistant Professor, London Business School, Sussex 

Place, London, NW1 4SA, United Kingdom, nbakshi@london.edu 

Co-Chair: Sang-Hyun Kim, Assistant Professor, Yale University, 135 

Prospect Street, New Haven, CT, 06520-8200, United States of America, 

sang.kim@yale.edu 

1 - Optimal Preventive Maintenance Under Contracting 

Sang-Hyun Kim, Assistant Professor, Yale University, 135 Prospect 

Street, New Haven, CT, 06520-8200, United States of America, 

sang.kim@yale.edu 

We investigate how various maintenance service contracts impact the optimal 

structure and performance of well-known preventive maintenance policies, such as 

age replacement and block replacement policies. We focus on comparing the results 

with those found in the classical reliability theory literature. 


365 

2 - Delaying the Delay Announcements 

Achal Bassamboo, Northwestern University, 

2001 Sheridan Road, Evanston, IL, United States of America, 

a-bassamboo@kellogg.northwestern.edu, Gad Allon 

This paper studies the impact of postponement of delay announcement on the 

ability of the firm to communicate non-verifiable congestion information to its 

customers as well as on the profits and utilities for the firm and the customers 

respectively. We show that this postponement can help the firm create credibility 

and augment the equilibrium language. However, in other settings this delay can 

also detract the equilibrium language. 

3 - An Auction Mechanism for Optimal Procurement From Multiple 

Suppliers with Asymmetric Information 

Yimin Yu, Assistant Professor, City University of Hong Kong, Hong 

Kong, yiminyu@cityu.edu.hk, Saif Benjaafar 

We study a retailer offering a procurement contract through a sealed auction to 

multiple suppliers. The production cost and the capacity of each supplier are private 

information. We design a modified VCG type auction to achieve the first best such 

that it is a dominant strategy for each supplier to reveal its true information. 

Furthermore, this auction has the following appealing properties: (1)Individual 

rational; (2)No free riding; (3)Budget balanced; (4)Coalition proof for the retailer. 

4 - Strategic Diagnosis and Pricing in Expert Services 

Mehmet Fazil Pac, PhD Candidate, Wharton School of Business, 

University of Pennsylvania, 3730 Walnut Street, Jon M. Huntsman 

Hall, Office 527.6, Philadelphia, PA, 19104, United States of America, 

mpac@wharton.upenn.edu, Senthil Veeraraghavan 

Customers often cannot identify the type of service they need, therefore they rely 

on experts, who also sell the service, for the diagnosis of their problem. The 

information asymmetry arising upon diagnosis leads to inefficiencies in the 

provision of the service. The expert has an incentive to over-provide or to ration 

services, based on the demand, capacity and the waiting cost. We investigate 

diagnosis, pricing and queue joining decisions in expert service markets using a 

queuing framework. 

■ WA14 

C - Room 14, Level 4 

Supply Chain Management VIII 


Chair: Wanxi Li, University of Wisconsin Milwaukee, 3202 N. Maryland 

Ave., Milwaukee, WI, 53201, United States of America, 

wanxili@uwm.edu 

1 - The Influence of Psychological Contract Violation on Supply Chain 

Decision-Making Behaviors 

Stephanie Eckerd, Ohio State University, 4943 Common Market 

Place, Dublin, United States of America, eckerd.2@osu.edu 

Conflict is an inevitable phenomenon in buyer-supplier relationships. In the face of 

varying types of conflict, individuals may respond differently depending on the level 

of psychological contract violation experienced. We report the results of a behavioral 

experiment that determines how those in boundary-spanning roles respond to 

conflict in the supply chain, specifically evaluating their economic decisions before 

and after an occurrence of conflict. 

2 - Procurement and Pricing in a Decentralized Multi-tier 

Assembly System 

Wanxi Li, University of Wisconsin Milwaukee, 3202 N. Maryland 

Ave., Milwaukee, WI, 53201, United States of America, 

wanxili@uwm.edu, Xiang Fang 

In a decentralized multi-tier assembly system, an assembler needs sets of modules 

produced by different module sub-assemblers, and each module needs multiple 

components purchased from different suppliers. The assembler faces stochastic 

demand. We characterize the equilibrium order quantity and pricing decisions, 

based on which we provide insights on the design of such multi-tier assembly 

systems.

WA15 

■ WA15 


Continuous Optimization 


Chair: John Carlsson, Assistant Professor, University of Minnesota, 111 

Church St SE, 130C, Minneapolis, MN, 55455, United States of America, 

jgc@me.umn.edu 

1 - The Scalar Equivalence of Optimization Criteria 

Surachai Charoensri, University of Texas at Arlington, Arlington, TX, 

76019, United States of America, surachai.charoensri@mavs.uta.edu, 

H. W. Corley 

We show that existing optimization criteria are equivalent to the maximization of a 

real-valued function in a one-dimensional Euclidean space. All and only solutions to 

an optimization problem in the original criterion can be obtained by scalarization 

without the typical convexity/concavity assumptions on the original objective 

functions. Examples include minimax, Pareto, and set-valued optimization, as well 

as cone-ordered optimization in abstract spaces. 

2 - Computational Studies of Randomized Multidimensional 

Assignment Problems 

Mohammad Mirghorbani, The University of Iowa, 208 Engineering 

Research Facility, 330 S. Madison Street, Iowa City, IA, 52240, 

United States of America, smirghor@engineering.uiowa.edu, 

Paul Krokhmal 

We propose a new heuristic approach for solving randomized multidimensional 

assignment problems (MAPs) with linear sum or bottleneck objectives that is based 

on recently obtained asymptotical properties of optimal value of random MAPs. The 

approach allows for drastic reduction of search space while guaranteeing high 

quality of the solution, and transforms the original problem into a maximum clique 

problem in multipartite graphs. 

3 - Convex Relaxations for Cubic Polynomial Problems 

Helder Inacio, Student, Georgia Institute of Technology, Georgia 

Institute of Technology, Atlanta, GA, 30332, United States of 

America, hinacio@isye.gatech.edu, Shabbir Ahmed, Matthew Realff 

We study convex relaxations for problems with polynomial constraints with degree 

less than or equal to 3. Specifically for terms of the form x^2 y we derive convex 

nonlinear underestimators in a similar fashion to McCormick underestimators for 

bilinear terms. We compare these estimators with other convex underestimators. 

4 - A Non-convex Geometric Partitioning Algorithm for 

Multi-vehicle Routing 

John Carlsson, Assistant Professor, University of Minnesota, 

111 Church St SE, 130C, Minneapolis, MN, 55455, 

United States of America, jgc@me.umn.edu 

We consider a stochastic vehicle routing problem in which vehicle depot locations 

are fixed and client locations in a service region are unknown, but are assumed to 

be i.i.d. samples from a given probability density function. We present an algorithm 

for partitioning the service region into sub-regions so as to minimize the maximum 

workload of any vehicle when the service region is simply connected and point-topoint 

distances follow some “natural” metric, such as any L^{p} norm. 

■ WA16 


Manufacturing II 


Chair: Paul E “Gene” Coffman, Jr, Technical Leader, Ford Motor 

Company, 6100 Mercury Drive, Dearborn, MI, 48126-2746, 

United States of America, gcoffman@ford.com 

1 - Optimization of Stochastic Flow Lines using Exact Linear 

Programming Formulations 

Raik Stolletz, Associate Professor, Technical University of Denmark, 

Department of Management Engineering, Lyngby, Denmark, 

raist@man.dtu.dk 

Several sampling approaches have been proposed to analyze stochastic flow lines 

with finite buffer capacities. If the number of buffers is given, the performance can 

be evaluated via a Linear Programming formulation. This presentation shows 

linearization approaches if the number of buffers is a decision variable. We develop 

a two-step optimization approach, where a discrete time approximation is used to 

get a first solution to speed up the solution of the exact linearization. 


366 

2 - Forecasting Stochastic Lead Times 

Jack Hayya, Professor Emeritus, Penn State University, School of 

Business, University Park, 16802, United States of America, 

jch@psu.edu, Uttarayan Bagchi 

Consider the case of iid lead times, which theoretically cannot be predicted. 

However, these lead times are subject to order crossover which transforms the iid 

lead times to an AR(1) process. But this AR(1) process contains outliers which make 

the residuals nonnormal. So we fit an ARCH(1) model with t-residuals. 

3 - Virtual Manufacturing at Ford Motor Company 

Paul E “Gene” Coffman, Jr, Technical Leader, Ford Motor Company, 

6100 Mercury Drive, Dearborn, MI, 48126-2746, United States of 

America, gcoffman@ford.com 

Eight years ago, Ford launched a Virtual Manufacturing Center and initiated a 

strategy to significantly reduce launch concerns by verifying new vehicles virtually 

before physical prototypes are built and by simulating manufacturing operations 

early in a new vehicle program. Ford’s recent success in customer satisfaction 

surveys is due in part to the 80% reduction in launch concerns achieved to date. 

We will describe the key elements of the strategy and the tools used to achieve 

these results. 

■ WA17 


OR For Infrastructure Development in India 

Cluster: OR/MS in India 


Chair: Ashok Mittal, Professor, IIT Kanpur, IME Dept., IIT Kanpur, 

Kanpur, UP, 208016, India, mittal@iitk.ac.in 

1 - Modeling Complex Aerospace Supply Chain With Delivery 

Guarantees 

Dinesh Kumar, Professor, Indian Institute of Management Bangalore, 

Bannerghatta Road, Bangalore, Ka, 560076, India, 

dineshk@iimb.ernet.in 

Aerospace has one of the complex supply chains that deals with assembly of 

millions of parts sourced from multiple vendors across the globe. Aircraft 

manufacturers expect their suppliers to deliver the parts just in time at their 

manufacturing facilities. In this paper we have used queueing models to analyse an 

assembly type manufacturing system with an objective to maximize the probability 

of on-time delivery of the parts. 

2 - Optimal Route Selection in a Computer Integrated Raw Material 

Handling Complex of an Integrated Steel Plant 

Salil K Dutta, SAIL- Durgapur Steel Plant India, TQM Deptt, 

Durgapur, 713203, India, salil_kumar_dutta@yahoo.co.in 

For efficient operation of Raw Materials Handling Complex of an Integrated Steel 

Plant, a PC based System is conceptualized. Two interactive modules: i) Optimal 

Blend-mix Module based on a mathematical programming model ii) Route Selection 

and Prioritization Module based on a integer programming model in conjunction 

with a Heuristic, has been proposed. 

3 - Understanding Passanger Switching Behavior and Yield 

Management for Indian Railways 

Ashok Mittal, Professor, IIT Kanpur, IME Deptt, IIT Kanpur, Kanpur, 

UP, 208016, India, mittal@iitk.ac.in, Rahul Sharma 

Indian rail provides different type of accomodation to rail passangers. For most of 

the trains seats in the choice class are not available until booked well in advance. 

We model the switching behavior of the passanges classified in different need 

categories. We use this behaviour to model yield management for Indian railways.

■ WA18 


OR in Practice I 

Sponsor: CPMS, The Practice Section 


Chair: Laura Galli, DEIS University of Bologna, Viale Risorgimento, 2, 

Bologna, 40136, Italy, l.galli@unibo.it 

Co-Chair: Bjarni Kristjansson, President, Maximal Software, Inc., 933 N. 

Kenmore St., Suite 218, Arlington, VA, 22201, United States of America, 

bjarni@maximalsoftware.com 

1 - OR at Ford 

Erica Klampfl, Technical Leader, Ford Research & Advanced 

Engineering, RIC Building, MD 2122, 2101 Village Rd, Dearborn, MI, 

48124, United States of America, eklampfl@ford.com 

I will provide a sampling of OR problems in areas such as Sustainability, 

Manufacturing, Purchasing, Product Development, Marketing, and Finance. We 

apply an analytical approach to understand the environmental implications of our 

products, enhance the sustainability of our business, and provide sound scientific 

input for corporate strategy and regulatory interactions. 

2 - Robust Planning and Online Re-scheduling for the Train 

Routing Problem 

Laura Galli, DEIS University of Bologna, Viale Risorgimento, 2, 

Bologna, 40136, Italy, l.galli@unibo.it, Alberto Caprara, Leo Kroon, 

Gabor Maroti, Paolo Toth 

Train Routing is a problem that arises in the early phase of the passenger railway 

planning process. However, train delays often disrupt the routing schedules thus 

railway nodes are responsible for a large part of the delay propagation. In this paper, 

we propose robust models and re-scheduling algorithms for train routing, and 

design a simulation framework to evaluate and compare their effectiveness. We 

present computational results based on real-world data from the Italian railways. 

3 - Combinatorial Model for Crew Scheduling in Train Transportation 

Hector Ramirez Cabrera, CMM, Universidad de Chile, Avda. Blanco 

Encalada 2120, Santiago, Chile, hramirez@dim.uchile.cl, Jorge 

Amaya, Paula Uribe 

This crew scheduling problem can be expressed as follows: given a set of crew teams 

(a pair composed by a driver and an assistant) and a set of trips (travel from one 

station to another one), the aim is to find an optimal allocation of these crews to 

the given trips satisfying operational constraints, such as labor laws, specific contract 

conditions, among others. The objective of our problem is to equilibrate the number 

of working hours realized by each crew in a given period of time. 

■ WA19 


Computational and Robust Approaches to Inventory 

and Revenue Management 

Sponsor: Revenue Management and Pricing Section 


Chair: Dan Iancu, IBM T.J. Watson Research Center, P.O. Box 218, 

Yorktown Heights, NY, 10598, United States of America, 

dan.iancu@us.ibm.com 

1 - A Dynamic Near-optimal Algorithm for Online Linear Program with 

Application to Revenue Management 

Zizhuo Wang, Stanford University, 14 Comstock Cir, Apt 106, 

Stanford, United States of America, zzwang@stanford.edu, Yinyu Ye, 

Shipra Agrawal 

We study a network revenue management problem where the customers come 

sequentially and decisions are made online. Our approach is distribution-free. We 

only assume that the customers come in a random order. By using a dynamic 

pricing algorithm where the prices come from the solution of a series of linear 

program, our algorithm is near-optimal given the initial inventory is large enough. 

This algorithm can be applied to a wide range of revenue management and resource 

allocation problems. 

2 - A Geometric Characterization of the Power of Finite Adaptability in 

Multi-stage Stochastic Optimization 

Andy Sun, Massachusetts Instititute of Technology, Operations 

Research Center, 50 Memorial Drive, Cambridge, MA, United States 

of America, sunx@mit.edu, Vineet Goyal, Dimitris Bertsimas 

We show a significant role that geometric properties of the uncertainty sets, such as 

symmetry, play in determining the power of robust and finitely adaptable solutions 

in multi-stage stochastic and adaptive optimization problems. We propose good 

approximation solution policies with performance guarantees that depend on the 

geometric properties of the uncertainty sets. To the best of our knowledge, these are 

the first approximation results for the multi-stage problems in such generality. 


367 

3 - Distributionally Robust Pricing and Replenishment Decisions for 

Multiple Products 

Dan Iancu, IBM T.J. Watson Research Center, P.O. Box 218, 

Yorktown Heights, NY, 10598, United States of America, 

dan.iancu@us.ibm.com 

In the current presentation, we discuss formulations and computational aspects 

related to determining pricing and replenishment policies for multiple products 

under uncertain customer demand. In particular, we look for policies directly 

parameterized in the model disturbances. Preliminary computational results, based 

on both synthetic, as well as real data from a large US retailer, are very promising, 

with adjustable policies considerably improving over open-loop decisions. 

■ WA20 


Challenges and Perspectives in Price 

Demand Relationship 



Chair: Aihong Wen, PROS Holdings, Inc., 3100 Main St Suite 900, 

Houston, TX, 77002, United States of America, awen@prospricing.com 

1 - Using Forecasts of Competitor Prices to Increase Sales 

Evan Brott, Scientist, PROS, 3100 Main Street #900, Houston, TX, 

77025, United States of America, EBROTT@prosrm.com 

Accurate predictions of competitor prices are invaluable for developing pricing 

strategies. By analyzing market conditions and recent pricing actions, we show a 

method of calculating an array of expected future competitor prices. Businesses may 

utilize this array to position themselves at a preferred position relative to a major 

competitor. Additionally, we show how the array can be used as input into rankbased 

optimization, to maximize margins subject to the expected competitive 

landscape. 

2 - A General Framework for using Customer Sensitivity to Execute 

Pricing Strategies 

Ed Gonzalez, Associate Scientist, PROS, 3100 Main St, Suite 900, 

Houston, TX, 77002, United States of America, 

EGonzalez@prosrm.com 

In this presentation, we demonstrate a general framework which allows for the 

automation of pricing strategies based on both business rules and fiscal goals. The 

main components of this framework are (1) developing price sensitivity (2) 

optimizing over given constraints and (3) measuring the effects of a pricing strategy 

at both the macro and micro level. 

3 - Data Mining Techniques in Modeling Price Elasticity 

Aihong Wen, PROS Holdings, Inc., 3100 Main St Suite 900, Houston, 

TX, 77002, United States of America, awen@prospricing.com 

Modeling price elasticity has always been a key step in pricing optimization. This 

presentation discusses the motivation behind seeking data mining tools for this 

challenging task, as well as our proposals and case study. 

■ WA21 


Carbon Reduction Policy Analysis 

Sponsor: Service Science 


Chair: Brian Jacobs, Assistant Professor, Michigan State University, 

Supply Chain Mgt Dept, N349 North Business Complex, East Lansing, 

MI, 48824-1122, United States of America, jacobsb@bus.msu.edu 

1 - Investment Planning for Electricity Generation Expansion Under 

CO2 Emission Reduction Policies 

Dong Gu Choi, Georgia Institute of Technology, 765 Ferst Drive, NW, 

Atlanta, GA, 30332, United States of America, 

doonggus@gatech.edu, Valerie Thomas 

As electricity demand increases and existing power plants age, electricity generators 

decide on supply technologies for new investment. This talk addresses the effects of 

CO2 emission reduction policies on the investment decision of an electricity 

generating firm. A dynamic programming model,incorporating policy uncertainty, is 

developed for technology investment choice.

WA23 

2 - China’s Regional CO2 Emissions: Characteristics and Emission 

Reduction Policies 

Lei Meng, Xi’an Jiaotong University, Box 1875, No.28 Xianning West 

Road, Xi’an, 710049, China, mleenig@gmail.com, Yong Xue, 

Ju’e Guo 

This paper analyzes the characteristics of regional CO2 emissions in China, using 

province level panel data from 1997 to 2007. The results show that there’re 

remarkable regional disparities among eastern coastal, midland and western areas. 

In view of uneven regional development and reverse distribution of energy 

resources and consumption, the CO2 emission reduction policies need customized 

combination of tax, price, investment and transfer payment to meet the actual 

situation in various areas. 

3 - Shareholder Value Effects of Voluntary Emissions Reductions 

Brian Jacobs, Assistant Professor, Michigan State University, Supply 

Chain Mgt Dept, N349 North Business Complex, East Lansing, MI, 

48824-1122, United States of America, jacobsb@bus.msu.edu 

Recent empirical evidence has demonstrated that the stock market reacts negatively 

to firm announcements of voluntary emissions reductions. In this work, we study 

how certain contextual factors influence the market reaction. Factors include the 

type of emission (regulated or unregulated), firm and industry characteristics, 

energy prices, and whether the firm’s announcement was standalone or part of a 

government or NGO initiative. 

4 - Optimal Fuel Conversion Strategy of Power Plant Under Different 

Carbon Policies 

Xiaohua Wu, Rensselaer Polytechnic Institute, 903 Peoples Ave 

Apt 3, Troy, NY, 12180, United States of America, wux4@rpi.edu, 

Aparna Gupta 

The carbon policy will accelerate the fuel type conversion process. In this paper, a 

general model of optimizing the long term fuel conversion strategy of a generator 

under different carbon policies is built and analyzed. The stochastic price evolutions 

of fuels, electricity and carbon emission are modeled to identify the impact of 

market fluctuations. Key decision factors are optimized to achieve the generator’s 

best economic performance and create a framework to assess policy impact. 

■ WA23 

C - Room 18D, Level 4 

Optimization in the Service Sector 



Chair: Ada Barlatt, Assistant Professor, University of Waterloo, 

Department of Management Sciences, 200 University Avenue West, 

Waterloo, ON, N2L3G1, Canada, abarlatt@uwaterloo.ca 

1 - Public School’s Meal Program: Finding the Best 

Cost-effective Menu 

Betzabe Rodriguez, Assistant Professor, University of Puerto Rico at 

Mayaguez, Call Box 9000, Mayaguez, PR, 00681, Puerto Rico, 

betzabe.rodriguez@upr.edu, Magaly Gonzalez 

Meal’s assortment highly affects the operational costs in the supply chain for the 

Puerto Rico School’s Meal Program (PRSMP). The government must comply with 

nutritional and service requirements while balancing delivery frequencies and 

transportation costs. We have developed a mathematical formulation for the 

operational costs of the food supply chain, with the objective to find a low cost 

meal’s assortment for the PRSMP. 

2 - Applying Value-at-Risk and Conditional Value-at-Risk to the 

Selective Newsvendor 

Arleigh Waring, University of Michigan, 1205 Beal Ave, Ann Arbor, 

MI, 48109, United States of America, awaring@umich.edu 

The selective newsvendor considers a single product firm that sells to several 

different markets in a single selling season. The firm decides which markets to serve 

and the total inventory to procure a priori. We evaluate the selective newsvendor 

using two common risk measures: Value-at-Risk and Conditional Value-at-Risk. We 

show the optimal order quantity and describe a selection criterion for the markets to 

serve and then compare the inherent tradeoffs between the two methods. 

3 - Evaluating Tradeoffs in Implementing Alternative 

Workweek Schedules 

Ada Barlatt, Assistant Professor, University of Waterloo, Department 

of Management Sciences, 200 University Avenue West, Waterloo, 

ON, N2L3G1, Canada, abarlatt@uwaterloo.ca, Juan Vera 

Climate change, 24/7 retail outlets, and technological advances have led to changes 

in the way people work. Around the globe, employees are switching to from the 

traditional 9AM to 5PM schedule to alternative workweek (AWW) schedules. In this 

presentation we will discuss the models developed to evaluate the tradeoffs between 

the benefits (e.g., increased operating hours) and the concerns (e.g., facilitating 

employee communication) in implementing AWW schedules. 


368 

■ WA24 


Network Models for Counterterrorism 

Sponsor: Public Programs, Service and Needs 


Chair: Susan Martonosi, Assistant Professor, Harvey Mudd College, 

Department of Mathematics, 301 Platt Boulevard, Claremont, CA, 91711, 

United States of America, martonosi@math.hmc.edu 

1 - A Network Flow Approach to Terrorist Network Disruption 

Susan Martonosi, Assistant Professor, Harvey Mudd College, 

Department of Mathematics, 301 Platt Boulevard, Claremont, CA, 

91711, United States of America, martonosi@math.hmc.edu, 

Doug Altner 

We present a new network disruption technique that tries to make otherwise 

secretive members of a terrorist group more visible. Through vertex deletion, this 

technique forces the secretive members to increase their participation in network 

communication. This talk will illustrate our disruption metric based on network 

flows, address graph-theoretic characteristics of promising vertices to target and 

discuss some computational challenges. 

2 - Diverting Communication Through a Clandestine Leader in a 

Social Network 

Doug Altner, Assistant Professor, United States Naval Academy, 

United States of America, altner@usna.edu, Susan Martonosi 

This talk investigates the following optimization problem: given a social network 

with a key vertex and a finite budget for deleting vertices, which vertices should be 

deleted to maximize the amount of communication that must be sent through the 

key vertex if the amount of communication between each pair of vertices equals the 

maximum flow between them? We present a meta-heuristic approach to this 

problem as well as computational results. 

3 - Tradeoffs in the Structure of Terrorist Networks 

Alexander Gutfraind, Postdoctoral Fellow, Los Alamos National 

Laboratory, Theoretical Division, Mail Stop B284, Los Alamos, NM, 

87545, United States of America, gfriend@lanl.gov 

Terrorist groups and other secret societies have a network structure reflecting their 

objectives of survival and attack. This talk will introduce a model that quantifies 

those using discrete optimization. Solving the model shows that the optimal 

structure of such networks is based on cells. Open non-violent activism pays under 

just two conditions: extreme tolerance and extreme repression. The model can also 

be used to design vital infrastructure networks. 

4 - Counter-Radicalization Influence Campaigns and Social Networks: 

What Does the Data Say? 

Richard Colbaugh, Sandia National Laboratories/New Mexico Tech, 

22 Camino Don Carlos South, Santa Fe, NM, 87506, 

United States of America, rcolbau@sandia.gov, Kristin Glass 

This talk presents results of a model-based, empirically-grounded study of social 

networks and influence campaigns and summarizes the practical implications of 

these findings. Interestingly our investigation reveals that some conventional 

wisdom regarding social networks and influence is either incomplete or incorrect. 

We consider two main topics: 1.) understanding influence generation/propagation/ - 

measurement as network dynamics phenomena, and 2.) roles for social media in 

influence campaigns. 

■ WA25 


Transportation, Intelligent Systems II 


Chair: Ali Guner, Research Assisstant, Wayne State University, 

4815 Fourth St., Detroit, MI, 48202, United States of America, 

arguner@wayne.edu 

1 - A Real Time Dynamic Rideshare System 

Ali Haghani, Professor, University of Maryland, College Park, 1173 

Glenn L. Martin Hall, College Park, MD, 20742, United States of 

America, haghani@umd.edu, Keivan Ghoseiri, Hadi Sadrsadat, 

Masoud Hamedi 

This paper presents an optimization model for a real-time dynamic rideshare system. 

The model maximizes the overall system performance in real time subject to ride 

availability, capacity, and passenger and driver time window constraints while 

considering users’ preferences. Model formulation and results are presented.

2 - Robust Multiple Priority Traffic Signal Control with Vehicle-to- 

Infrastructure Communication Systems 

Qing He, University of Arizona, 1127 E James E. Rogers Way, 

Tucson, AZ, 85721, United States of America, 

heqing@email.arizona.edu, Larry Head, Jun Ding 

This paper examines the multiple priority problems in traffic signal control under 

the condition that vehicle-to-infrastructure communication is available. Given the 

current multiple priority request information from on-board equipment (OBE), a 

robust MILP is developed with actuated control integrated to mitigate the delay for 

both vehicles with priority and passenger cars. A numerical experiment with 

VISSIM and GAMS shows the effectiveness of proposed approach. 

3 - Relationship of Pretrip Traveler Information System to 

Non-motorized and Public Traffic in China 

Yi Zhang, School of Transportation Engineering, Tongji University, 

No.4800 Cao’an Road, Shanghai, China, Shanghai, China, 

darrenzhy@gmail.com, Meiping Yun, Xiaoguang Yang 

Traveler information system is able to change travelers’ travel behavior and alleviate 

congestion. Based on a Travel Desire Survey in Zhongshan City, China, the 

relationship of pretrip traveler information system to non-motorized and public 

traffic was examined. The commuters’ propensity to change travel mode from 

private car to non-motorized and public traffic was obtained. It is also showed that 

the propensity would vary in the context of different trip distance and travelers’ 

characteristics. 

4 - Dynamic Routing in Stochastic Time-Dependent Networks for 

Milk-Run Tours with Time Windows 

Ali Guner, Research Assisstant, Wayne State University, 

4815 Fourth St., Detroit, MI, 48202, United States of America, 

arguner@wayne.edu, Ratna Babu Chinnam, Alper Murat 

JIT requires frequent and reliable pick-ups and deliveries within specified time 

windows. However, growing congestion on road networks is increasing variability in 

travel times, making it difficult to achieve efficient and reliable deliveries. We 

investigate the impact of utilizing real-time ITS information to route the vehicle. 

Our dynamic routing algorithm handles milk-run tours under a TSP framework, 

while modeling congestion on arcs as stochastic and time-dependent congestion 

states. 

■ WA26 


Data Mining and Knowledge Discovery in Health Care 

Sponsor: Data Mining 


Chair: Durai Sundaramoorthi, Assistant Professor, Missouri Western State 

University, 4525 Downs Drive, Saint Joseph, MO, 64507, United States of 

America, dsundaramoorthi@missouriwestern.edu 

1 - An Adaptive Pain Management Framework 

Ching-Feng Lin, Student, UTA, 212 S Cooper St. #220, Arlington, 

TX, 76013, United States of America, ching-feng.lin@mavs.uta.edu, 

Victoria Chen, Robert Gatchel 

The Eugene McDermott Center for Pain Management at the University of Texas 

Southwestern Medical Center at Dallas conducts a two-stage interdisciplinary pain 

management program that considers a wide variety of treatments. We structure this 

decision-making process using dynamic programming to generate adaptive 

treatment strategies for this two-stage program. State transition models were 

derived using data from the two-stage pain management program. 

2 - The Role of Insurance Claims Databases in Healthcare Research 

Yihan Guan, PhD Candidate, Stanford University, Huang Engineering 

Center 212F, Stanford, CA, 94305, United States of America, 

yihan@stanford.edu, Margret Bjarnadottir 

Health insurance claims data have been used in a wide spectrum of health care 

research during the past two decades, including studying drug therapy outcomes, 

assessing quality of care, estimating population disease burden, predicting health 

care cost, and detecting adverse drug effects. This talk discusses the evolution of 

applications of claims data, the current research frontier associated with claims data, 

and some future opportunities. 

3 - Analyzing Patient Discharge Data From an Urban Hospital using 

Data Mining Techniques 

Xiuli (Shelly) Qu, Assistant Professor, North Carolina A&T State 

University, 1601 E. Market Street, 424 McNair Hall, Greensboro, NC, 

27411, United States of America, xqu@ncat.edu, Xiaochun Jiang, 

Lauren Davis 

As aging is beginning to impact the baby boom generation, they begin to experience 

more chronic diseases and need more inpatient care. To deal with this trend, 

hospitals need to reduce the average length of stay (ALOS). In this study, data 


369 

mining techniques were used to analyze the inpatient discharge data from an urban 

hospital. Four models were built to predict ALOS. Our results indicated that the 

Ensemble model was the best fit and age and chronic disease were the important 

predictors. 

4 - Examining Relationships Between Medical Home and 

Patient Experience 

Sharon Johnson, Associate Professor, Worcester Polytechnic 

Institute, Department of Management, 100 Institute Road, 

Worcester, MA, 01566, United States of America, sharon@wpi.edu, 

Edward Westrick, Lori Pelletier 

The Patient-Centered Medical Home (PCMH) is a new model for comprehensive 

primary care that seeks to strengthen the physician-patient relationship. This 

exploratory study utilizes Pearson correlation coefficients to examine relationships 

between PPC-PCMH Survey results, which measure adoption of PCMH structures, 

and patient experience data. The results show an unexpected negative correlation 

between the PPC-PCMH structures of access and communication and the related 

patient experience measure. 

■ WA27 


Linear Programming 


Chair: Holly Floyd, Texas State University-San Marcos, 250 S. Stagecoach 

Trl. #136, San Marcos, United States of America, hf1046@txstate.edu 

1 - Explore the Higher-Order Rescaling Perceptron Algorithm 

Dan Li, Department of Industrial and Systems Engineering, 

Lehigh University, 200 West Packer Ave, Bethlehem, PA, 18015, 

United States of America, dal207@lehigh.edu, Tamàs Terlaky 

The rescaling perceptron algorithm solves LO problems with high probability in 

O(nln(1/r)) iterations, where r is the radius of the largest inscribed ball. It uses one 

vector to rescale the system at each iteration. We realize rescaling by using parallel 

processors and several vectors in one higher-order step. We explore how the 

number and quality of vectors affect the rescaling rate. With properly chosen 

vectors, we get better rescaling rates and improve the complexity. 

2 - A New Method to Solve Linear Programming Problems 

Oscar Buitrago, Professor, Universidad Libre de Colombia, 

oscary.buitragos@unilibrebog.edu.co, osyesu@gmail.com, Bogotà 

D.C, Colombia, oscary.buitragos@unilibrebog.edu.co, Andres Ramirez 

Many methods have been developed for solving LP problems, including the famous 

Simplex and interior point algorithms. In this study a new procedure for solving LP 

problems is described, based on orthogonal projections that move through the 

polyhedron frontier which defines the feasible region until it reaches the optimal 

point. 

3 - Optimal Deployment Plan of Emission Reduction Technologies 

Muhammad Bari, Student, Texas A&M University, University Dr., 

College Station, TX, 77843, United States of America, 

ehsanulbarihome@yahoo.com, Josias Zietsman, Luca Quadrifoglio, 

Mohamadreza Farzaneh 

The objective of this research was to develop methodologies for optimal deployment 

of emission reduction technologies for non-road equipment in a cost effective and 

optimal manner. The multi-objective problem consists of two weighted objectives, 

(i) maximizing NOx reduction and (ii) maximizing fuel savings. The models 

developed in this study serves as a tool to assist the decision makers to decide about 

the deployment preference of technologies. 

4 - A Multi-Period Energy and CO2 Emission Optimization Toward 

Sustainable Automotive Manufacturing 

Seog-Chan Oh, Senior Researcher, General Motors R&D, 30500 

Mound Road, Warren, MI, 48090, United States of America, 

seog-chan.oh@gm.com, Stephan Biller 

Fluctuating energy prices and enactment of climate change legislations add 

increasing uncertainty in costs associated with energy and CO2 emission for 

automotive companies. To combat the challenges, we propose a mixed-integer 

optimization model to maximize the reduction of energy and CO2 emission costs for 

the automotive manufacturing process. Given a multi-year budget, we analyze 

different scenarios, assuming different impacts of energy prices and environmental 

regulations.

WA28 

■ WA28 

C - Room 4C, Level 3 

Health Management of Complex Systems 

Sponsor: Quality, Statistics and Reliability 


Chair: Qingyu Yang, Assistant Professor, Wayne Sate University, 

4815 Fourth Street, Detroit, MI, 48202, United States of America, 

qyang@wayne.edu 

Co-Chair: Jian Liu, Assistant Professor, University of Arizona, 

Rm 268, 1127 E. James E. Rogers Way, Tucson, AZ, 85741, 

United States of America, jianliu@email.arizona.edu 

1 - Multi-level Multi-State Information Integration for System 

Performance Prediction 

Jian Liu, Assistant Professor, University of Arizona, Rm 268, 1127 E. 

James E. Rogers Way, Tucson, AZ, 85741, United States of America, 

jianliu@email.arizona.edu 

Heterogeneous data available at different levels of a complex system create great 

opportunity to more accurately predict the system level performance. This research 

provides a Bayesian approach that simultaneously combines multi-state event data 

of interdependent components and subsystems. A simulation example demonstrates 

the capability of the approach. 

2 - Sensor Recovery in Multivariate Condition Monitoring Systems 

Haitao Liao, Assistant Professor, The University of Tennessee, 211 

Pasqua Building, Knoxville, TN, 37996, United States of America, 

hliao4@utk.edu, Jian Sun 

Loss of sensor readings due to malfunction of connectors and/or sensors is crucial to 

fault diagnosis and prognosis in a multichannel condition monitoring system. To 

improve the operational reliability of the overall system, effective sensor recovery 

becomes an important, value added technique. This work addresses a statistical 

sensor recovery methodology to enhance multichannel condition monitoring. 

3 - Power Reliability Management in Smart Grids via Virtual 

Energy Provisioning 

Tongdan Jin, Texas State University, 601 University Drive, San 

Marcos, TX, 78666, United States of America, tj17@txstate.edu, 

Ying Yu, Mahmoud Mechehoul 

We propose a novel demand side management concept called Online Purchase 

Electricity Now (OPEN) to minimize load variations and generation uncertainties 

caused by renewable energies. The OPEN system allows customers to order and 

request advanced electricity via the Internet as if performing online purchasing. It 

aims for customers to achieve “order exactly what they need, and consume exactly 

what they ordered”. The new concept has great potentials and promises for smart 

grid technologies. 

4 - Failure Profile Analysis of Repairable Systems 

Qingyu Yang, Assistant Professor, Wayne Sate University, 4815 

Fourth Street, Detroit, MI, 48202, United States of America, 

qyang@wayne.edu, Yong Chen, Yili Hong, Jianjun Shi 

The relative failure frequency among major failure modes of a repairable system is 

referred to as failure profile. Identification of failure profile can provide valuable 

information for system design and maintenance management. In this research, a 

statistical model and two testing procedures are developed to study the statistical 

properties of the failure profile. The efficiency of the developed methods is verified 

by a case study of a high throughput screening (HTS) process. 

■ WA29 


Renewable Energy Integration Into Power Systems for 

Smart Operations 



Chair: Eunshin Byon, PhD, Postdoctoral Research Associate, Texas A&M 

University, 241 Zachry, 3131 TAMU, College Station, TX, 77840, 

United States of America, esbyun@neo.tamu.edu 

1 - Impact of Environmental Factors to the Degradation of Solar 

Photovoltaic Module 

Rong Pan, Assistant Professor, Arizona State University, Sch Compt 

Infor & Dec Sys Engr, Tempe, AZ, 85287, United States of America, 

Rong.Pan@asu.edu 

To make the solar energy economically competitive, PV manufacturers typically 

provide 20-30 year warranties to their customers. In this talk, we present a practical 

approach to weather modeling and its usage in PV module degradation analysis. We 

have analyzed the performance data of several PV modules collected over a long 

time of period (approximately 15 years). These data will be used to demonstrate the 

methodology to be developed in this study. 


370 

2 - Hierarchical Simulation Modeling Framework for Electrical Power 

Quality and Capacity 

Esfandyar Mazhari, University of Arizona, 1127 East James E. Rogers 

Way, Tucson, AZ, 85721-0020, United States of America, 

emazhari@email.arizona.edu, Young-Jun Son 

A two level hierarchical simulation modeling framework is proposed for an electric 

power network involving PV-based solar generators, various storage units, and grid, 

where the lower level model concerns power quality and the higher level model 

concerns capacity. The higher level is based on agent-based modeling while the 

lower level is based on circuit-level, continuous time modeling. An integration and 

coordination framework is developed, and it is demonstrated with a utility level 

scenario. 

3 - Reliability Issues in Power System Planning and Operation with 

Renewable Energy Sources 

Chanan Singh, Regents Professor, Texas A&M University, 

Department of Electrical & Computer Engi, College Station, TX, 

77843, United States of America, singh@ece.tamu.edu 

Renewable energy sources are fast penetrating the power grid. It is estimated that 

wind power alone may be close to 20 percent of the power scenario in America and 

many European nations have similar targets.Such significant penetrations of 

renewable energy pose considerable challenges for power system reliability in 

planning and operation. This presentation will review this problem and share some 

results to model and analyze the impact of their integration on reliability of the 

power grid. 

4 - Virtual Models of Wind Turbines 

Andrew Kusiak, Professor, The University of Iowa, Mechanical and 

Industrial Engineering, Iowa City, IA, 52242, United States of 

America, andrew-kusiak@uiowa.edu 

Complex nature of the wind, makes modeling wind turbines is a major challenge. 

Data mining offers algorithms for modeling wind turbines. A methodology for the 

development of virtual models of wind turbines is presented. The virtual models are 

developed and tested with data collected at a wind farm. Several data-mining 

algorithms for parameter selection and model extraction are analyzed. The research 

results are illustrated with industrial case studies. 

■ WA30 


Reliability III 


Chair: Elias Keedy, PhD Student, University of Houston, S350 

Engineering Bldg 1, Houston, TX, 77204, United States of America, 

eliekeedy@yahoo.com 

1 - Reliability Modeling in the Design of a New Energy Concept 

Sarah Riddell Powers, Lawrence Livermore Natl Lab, 7000 East 

Avenue, L-153, Livermore, CA, United States of America, 

powers22@llnl.gov 

Laser Inertial Fusion Energy (LIFE) is an advanced energy technology under 

development at LLNL. Achieving high system availability is a key project goal for 

the economic competitiveness of LIFE. A model is developed to simulate and study 

the reliability of various system designs, maintenance strategies and optimal spare 

component levels. Results show the advantage of design modularity in achieving 

high system availability without needing high component reliability thus expediting 

time to market. 

2 - Optimal Maintenance for Linear Consecutively Connected Systems 

Rui Peng, Department of Industrial and Systems Engineering, 

National University of Singapore, Singapore, ISE Department, BLK 

E1A, NUS, Singapore, Singapore, Singapore, g0700981@nus.edu.sg, 

Gregory Levitin, Szu Hui Ng, Min Xie 

This paper considers a linear multi-state consecutively connected system (LMCCS) 

consisting of N+1 linear ordered elements. Each element can provide a connection 

between the position in which it is allocated and the next few positions. The system 

fails if the first element is not connected with the (N+1)th element. A framework is 

proposed to solve the cost optimal maintenance strategy of the system subject to 

reliability requirement. 

3 - Comparative Study of Stochastic Models to Estimate Reliability of 

Infrastructure Systems 

Raha Akhavan-Tabatabaei, Assistant Professor, Universidad de los 

Andes, Cra 1 Este # 19A-40, Bogotà, Colombia, 

r.akhavan@uniandes.edu.co, Edgar Mauricio Sànchez Silva, 

Juan Sebastian Borrero 

We consider the problem of infrastructure reliability subject to progressive 

deterioration and shocks. We consider the deterioration mechanisms and compare 

the performance of various stochastic models in describing the remaining life of a 

component. Some conclusions are shown and stochastic models are classified 

according to their efficiency.

4 - Reliability and Maintenance of Stents Based on Probabilistic 

Analysis of Multiple Failure Processes 

Elias Keedy, PhD Student, University of Houston, S350 Engineering 

Bldg 1, Houston, TX, 77204, United States of America, 

eliekeedy@yahoo.com, Qianmei Feng 

The high demand in counteracting the effects of atherosclerosis and the ignorance of 

the probabilistic aspects in existing studies make the investigation of stents reliability 

a competitive concern. Based on fracture mechanisms, we analyze two processes: 

delayed and instantaneous failures. General reliability and maintenance models are 

developed to acquire an optimal replacement policy of stents. Our work provides 

new perspectives on approaching reliability concepts in medical devices evolution. 

■ WA31 


Forecasting I 


Chair: Ozden Gur Ali, Koc University, Sariyer, Istanbul, Turkey, 

oali@ku.edu.tr 

1 - Social Media Aided Event Forecasting 

Mohammad Ali Abbasi, PhD Student, Arizona State University, 699 

S. Mill Ave. #553, Attn: Mohammad Ali Abbasi, Tempe, AZ, 85281, 

United States of America, ali.abasi@asu.edu 

We introduce a novel method to forecast social events and behaviors of 

communities in the society. We’ve crawled tweets and blog posts of specific groups 

for a period of 5 months and classify them using opinion mining methods to extract 

members’ intention for specific events. Simultaneously we track the communities’ 

real (on the ground) activities. Experiments showed that with enough number of 

online activities we can forecast the events related to the communities with high 

confidence. 

2 - Time Series Forecasting in Oracle Crystal Ball 

Samik Raychaudhuri, Oracle Americas Inc., 2033 N Fork Dr, 

Lafayette, CO, 80026, United States of America, samikr@gmail.com, 

Eric Wainwright 

In this presentation we will have an overview of the functionality provided by 

Oracle Crystal Ball’s (CB) Predictor engine. CB Predictor has an intuitive interface 

for selecting, managing and cleaning data, outlier detection and filling in missing 

values, running multiple seasonal and nonseasonal forecasting algorithms and 

regression on large datasets, and a coherent way of presenting and extracting results 

or generating reports. We will also have a sneak preview of forthcoming features. 

3 - Using Conditional Kernel Density Estimation for Wind Power 

Density Forecasting 

Jooyoung Jeon, University of Oxford, 2 Alan Bullock Close, 

St Clements, Oxford, United Kingdom, 

joo.jeon@smithschool.ox.ac.uk, James Taylor 

Wind power is recognized as one of the most promising renewable energy resources. 

Despite great benefits from wind power density forecasting, most research has 

focused on point forecasting. We develop a novel approach to producing wind 

power density forecasts. The inherent uncertainty in wind speed and direction and 

the stochastic relationship of wind power to wind speed and direction are addressed 

using Monte Carlo simulation of a VARMA-GARCH model and conditional kernel 

density estimation. 

4 - Driver Moderator Model - Mining with Domain Knowledge 

Ozden Gur Ali, Koc University, Sariyer, Istanbul, Turkey, 

oali@ku.edu.tr 

We devise an interpretable method that predicts SKU sales as a function of the 

pricing, promotion and product availability decisions by the retailers consistently 

across similar situations. The method results in interpretable models, leverages 

domain knowledge, does concurrent model estimation and feature selection and 

relies on data pooling for generalization capability. We evaluate our method with 

two multi-store large scale grocery databases from Turkey and the USA. 


371 

■ WA32 


Applications and Heuristics Methods in 

Integer Programming 

Sponsor: Computing Society 


Chair: Haibo Wang, Assistant Professor, Texas A&M International 

University, 5201 University Blvd, Laredo, TX, 78041, United States of 

America, hwang@tamiu.edu 

1 - An Approach for Parallel Machine Job Scheduling Problem with 

Interrelated Processing Times 

Bahram Alidaee, The University of Mississippi, School of Business, 

University, United States of America, balidaee@bus.olemiss.edu, 

Haibo Wang 

This paper addresses the parallel machine job scheduling problem with interrelated 

processing times, which has applications in many areas such as storage allocation in 

computer design, organization restructuring, continuous project scheduling, and 

scheduling in political campaigns. we present a quadratic unconstrained binary 

optimization(QUBO) model and solve it with a neighborhood search heuristic. The 

computational time and the optimization gap reduction are reported. 

2 - Solving Large Max Cut Problems using Tabu Search 

Gary Kochenberger, Professor, University of Colorado at Denver, 

1250 14th Street, Denver, 80217, United States of America, 

Gary.Kochenberger@ucdenver.edu, Haibo Wang, Fred Glover, 

Zhipeng Lu, Jin-Kao Hao 

Large Max Cut Problems continue to pose a challenge for most approaches 

presented in the literature. In this paper we report our experience with a new Tabu 

Search approach for the general unconstrained binary quadratic program as it is 

applied to max cut test problems. Best known results on many instances with up to 

10,000 vertices are reported. 

3 - Warehouse Management - Cross Docking 

Jun Huang, Texas A&M International University, United States of 

America, huangjundragon@sina.com, Haibo Wang 

Warehouse management is a critical topic in logistic research field. The warehouse is 

divided into several functional areas such as reserve storage area, order picking area 

and cross docking. This study will combine the cross docking problem with storage 

area layout design and products allocation by using the graph partition method as 

finding a partition of the vertices of a given graph into subsets satisfying certain 

properties. Thus, the aforementioned problem can be solved more properly. 

■ WA33 


Integrating Constraint Programming and OR 



Chair: Tallys Yunes, Department of Management Science, University of 

Miami, Coral Gables, FL, 33124-8237, United States of America, 

tallys@miami.edu 

1 - Relaxation Based on Multivalued Decision Diagrams 

John Hooker, Carnegie Mellon University, Tepper School of Business, 

Pittsburgh, PA, United States of America, 

john@hooker.tepper.cmu.edu, Sam Hoda, Willem-Jan van Hoeve 

We solve scheduling problems with a branching method in which a multivalued 

decision diagram (MDD) plays the role of both the linear relaxation in mixed 

integer programming and the constraint store in constraint programming. We obtain 

order-of-magnitude speedups on problems with multiple “among” constraints, 

which are used in employee scheduling, assembly line sequencing, etc. 

2 - A Logic-Based Benders’ Decomposition Approach for Solving an 

Aircraft Maintenance Scheduling Problem 

J. Christopher Beck, University of Toronto, 5 King’s College Rd, 

Toronto, ON, M5S 3G8, Canada, jcb@mie.utoronto.ca, 

Maliheh Aramon Bajestani 

We address a flight maintenance problem where the goal is to schedule 

maintenance jobs to maximize the expectation that an existing flying program will 

have a full complement of aircraft. The schedule must consider the aircraft failure 

probabilities and maintenance capacities. We present a logic-based Benders’ 

decomposition that exhibits multiple orders-of-magnitude improvement over an 

existing mixed-integer programming model.

WA34 

3 - Improving the Held and Karp Approach with 

Constraint Programming 

Willem-Jan van Hoeve, Carnegie Mellon University, 5000 Forbes 

Avenue, Pittsburgh, PA, 15213, United States of America, 

vanhoeve@andrew.cmu.edu, Michel Rueher, Jean-Charles Régin, 

Louis-Martin Rousseau 

We show that domain filtering algorithms developed for the weighted spanning tree 

constraint can be adapted to the Held and Karp procedure to solve the TSP. In 

addition, we introduce a special-purpose filtering rule based on the underlying 

mechanisms used in Prim’s algorithm. Finally, we explore two different branching 

schemes to close the integrality gap. 

■ WA34 


Joint Session ICS/ Complex/ QSR: Sensing, Prediction 

and Prognostics in Complex Systems 

Sponsor: Computing Society/ Complex Systems/ Quality, 

Statistics and Reliability 


Chair: Satish Bukkapatnam, Professor, Oklahoma State University, 318, 

Enginnering North, School of Industrial Engineering, Stillwater, OK, 

74075, United States of America, satish.t.bukkapatnam@okstate.edu 

1 - Willingness-to-Pay Prediction using Empirical Mode Decomposition 

and Local Gaussian Process 

Satish Bukkapatnam, Professor, Oklahoma State University, 318, 

Enginnering North, School of Industrial Engineering, Stillwater, OK, 

74075, United States of America, satish.t.bukkapatnam@okstate.edu, 

Changqing Cheng, Akkarapol Sa-ngasoongsong 

Prediction of customer preferences over time is important for effective design of a 

product portfolio. However, the preferences evolution follows a nonlinear and 

nonstationary dynamics. We present two new approaches, based a local gaussian 

process (LGP) and empirical mode decomposition (EMD) for accurate prediciton of 

customer willingness-to-pay (WTP). 

2 - Identification of a Mixture of Hidden Markov Models using 

Metaheuristic Search 

Dragan Djurdjanovic, Assistant Professor, The University of Texas at 

Austin, 1 University Station, C2200, ETC 5.122, Austin, TX, 78712, 

United States of America, dragand@me.utexas.edu, Michael Cholette 

Mixture of Hidden Markov Models (HMMs) was recently proposed for modeling of 

degradation of systems working under variable operating conditions. We present a 

method for identification of a mixture of HMMs from a sequence of observations 

corresponding to a known sequence of operating conditions. The method utilizes a 

metaheuristic search to initialize the Baum-Welch algorithm and maximize the 

likelihood of observations. 

3 - Weather Forecasts and Power Grid Operations 

Victor Zavala, Argonne National Laboratory, Math and Comp. 

Science Div., Bdg 240, 9700 S Cass Ave, Argonne, IL, 60439, 

United States of America, vzavala@mcs.anl.gov, Mihai Anitescu, 

Emil Constantinescu 

We review motivations and challenges arising in the implementation of advanced 

numerical weather prediction models in power grid operations. In particular, we 

analyze trade-offs between computational bottlenecks, resolution, accuracy, and 

economic performance of power systems. 

4 - Binary Code Provenance and Heredity Detection Modeling 

and Error Analysis 

LiYing Cui, Reserach Assistant, Penn State University, 310 Leonhard 

Building, University Park, PA, 16802, United States of America, 

luc5@psu.edu, Soundar Kumara 

Information in this digital era exists as binary code. Due to the information 

explosion in recent years there is a need to detect the lineage among information. 

In the cases of malware, it will be necessary identify the variants and to predict the 

functionality of future viruses. In this work, we propose a binary code provenance 

and heredity detection and prediction method based on CyGene (Gene equivalent 

in Cyber engineering) detection. The error analysis of this method is studied 

thoroughly based on the ideas of false positives and false negatives. 


372 

5 - On the Optimal Control of Large Scale Systems with 

Stochastic Interactions 

Eugene Perevalov, Lehigh University, Department of Industrial & 

Systems Engineerin, Bethlehem, PA, 18017, eup2@lehigh.edu 

We study the problem of efficient control of large-scale systems that can be modeled 

by a collection of homogeneous elements with stochastic interactions. For the 

general case, we formulate the efficient control problem as an inverse inference 

problem in Bayesian networks. For the case of systems with a high degree of 

uniformity in the element interactions, we propose to use the renormalization 

group (RG) approach for the purpose of characterizing the phase structure of the 

system and feasibility of efficient control. We give several examples and perform 

numerical experiments to check the validity of the proposed approach. 

■ WA35 


Advances in Anomalous Diffusion I 

Sponsor: Applied Probability 


Chair: Iddo Eliazar, Professor, Holon Institute of Technology, 

P.O. Box 305, Holon, 58102, Israel, eliazar@post.tau.ac.il 

1 - Universal Generation of Fractal Statistics 

Iddo Eliazar, Professor, Holon Institute of Technology, P.O. Box 305, 

Holon, 58102, Israel, eliazar@post.tau.ac.il, Joseph Klafter 

We present a stochastic superposition model which is capable of generating - in a 

universal fashion - various “fractal statistics”. The stochastic superposition model is 

general and robust, and arises naturally in diverse fields of science and engineering. 

Universally-generated “fractal statistics” include: anomalous diffusion - power-law 

growth of temporal dispersion; Lévy flights - power-law amplitudinal fluctuations; 

1/f noises - power-law temporal correlations. 

2 - Detecting the Origins of Anomalous Siffusion: P-variation Test and 

its Applications 

Marcin Magdziarz, Dr, Wroclaw University of Technology, 

Wyspianskiego 27, Wroclaw, 50-370, Poland, 

marcin.magdziarz@pwr.wroc.pl 

Motivated by growing interest in single molecule spectroscopy, we propose a 

method to detect mechanisms leading to subdiffusion. We introduce the so-called pvariation 

test, which allows distinguishing between two models of subdiffusion on 

the basis of one realization of the unknown process. We apply our approach to 

experimental data (random motion of an individual molecule inside the E. coli cell). 

3 - Aging, Ergodcity Breaking and Universal Fluctuations in Continuous 

Time Random Walks 

Igor Sokolov, Prof., Humboldt University Berlin, Newtonstr. 15, 

Berlin, D-12489, Germany, igor.sokolov@physik.hu-berlin.de 

We consider subdiffusive transport within the continuous time random walk 

(CTRW) model. The anomalous diffusion under CTRW is a process with nonstationary 

increments and shows explicit dependence of observables on the time 

elapsed from preparing the system in its present state. This corresponds to aging of 

the process, leading to death of linear response to an external stimulus and intrinsic 

ergodicity breaking. Different manifestations of these properties will be discussed. 

■ WA36 


Panel Discussion: Research in Teaching Schools 

Sponsor: INFORM-ED 


Moderator: John Kros, Associate Professor of Marketing and Supply 

Chain Management, East Carolina University, 3121 Bate Building, 

Greenville, NC, 27858, United States of America, krosj@ecu.edu 

1 - Panel Discussion: Doing Research at Balanced Model Schools 

Panelists:John Kros, Associate Professor of Marketing and Supply 


Greenville, NC, 27858, United States of America, krosj@ecu.edu, 

Marvin Brown,Assistant Professor of CIS, Grambling University, 

403 Main Street, Grambling LA 71245, United States of America, 

brownm@gram.edu, Christopher Keller,Assistant Professor of 

Marketing and Supply Chain Management, East Carolina University, 

3136 Bate Building, School of Business, Greenville NC 27858, United 

States of America, kellerc@ecu.edu, Scott Nadler, Assistant Professor, 

University of Central Arkansas, COB 312, Conway AR 72035, 

United States of America, SNadler@uca.edu 

This is a session set up by the INFORM-ED forum it is a panel on the topic of 

“Doing Research at Balanced Model Schools.”

■ WA37 


Stochastic Control, Dynamic Games and Their 

Applications 



Chair: Hector Jasso-Fuentes, Mathematics Department, CINVESTAV-IPN, 

Apartado Postal 14-740, 07000, Mexico DF, Mexico, Mexico, 

hjasso@math.cinvestav.mx 

1 - Some New Results in Controlled Switching Diffusions 

Ari Arapostathis, Professor, University of Texas at Austin, 

1 University Station (C0803), Department Electrical and Computer 

Eng., Austin, TX, 78712, United States of America, 

ari@mail.utexas.edu 

We study the stability and ergodic control problems of controlled switching 

diffusions, modeled by a coupled system of Ito stochastic differential equations, 

under no assumption of irreducibility. Stability is defined as positive recurrence 

relative to an open ball in the space of the continuous component. We show among 

others that if the model is stable, then the upper envelope of the class of invariant 

probability measures is a finite measure, and we study the implications of this 

property. 

2 - Control of Inventories with Markov Demand 

Alain Bensoussan, Research Professor, University of Texas at Dallas, 

School of Management, Office 3.211, Dallas, 830688, 

United States of America, axb046100@utdallas.edu 

We consider inventory control problems in discrete time. The horizon is infinite, and 

we consider discounted payoffs as well non-discounted payoffs (ergodic control). We 

may have backlog or not. We may have set up costs or not. We show how the base 

stock policy and the s,S policy can be extended. 

3 - Overtaking Equilibria for Zero-sum Markov Games 

Onesimo Hernandez-Lerma, Prof., CINVESTAV-IPN, Math. 

Department, A.Postal 14-740, Mexico D.F. 07000, Mexico, 

ohernand@math.cinvestav.mx 

We study overtaking (a.k.a. catching-up) optimality for a class of zero-sum Markov 

games that includes stochastic differential games, and games with a countable state 

space. 

4 - Joint Optimization of Pricing Strategies and Inventory Control with 

Continuous Stochastic Demand 

Yongqiang Wang, University of Maryland, College Park, 3182 AVW, 

College Park, MD, United States of America, yqwang@umd.edu, 

Michael Fu, Steven Marcus 

We analyze the dynamic pricing problem for inventory systems with price-sensitive 

continuous stochastic demands. An analytical solution for a special demand is 

provided. For more general demand models, we propose a simulation-based method 

for solving the dynamic pricing problem, assuming a finite number of price changes 

over the time horizon of interest. In the framework of our simulation-based 

algorithm, we can also jointly optimize the price and the initial inventory level. 

■ WA38 


Complementarity Problems 

Sponsor: Optimization/Linear Programming and Complementarity 

(Joint Cluster ICS) 


Chair: Hande Benson, Associate Professor, Drexel University, Department 

of Decision Sciences, LeBow College of Business, Philadelphia, PA, 19104, 

United States of America, hvb22@drexel.edu 

1 - Single Timescale Regularization Schemes for Monotone 

Nash Games 

Uday Shanbhag, Asst. Professor, University of Illinois at Urbana 

Champaign, Urbana, Il, United States of America, 

udaybag@illinois.edu, Aswin Kannan 

We consider single-timescale schemes for the distributed computation of equilibria 

arising from montone Nash games, Specifically, we propose iterative regularization 

counterparts of Tikhonov and proximal-point schemes where 

regularization/centering parameters are updated after every projection step, rather 

than when an approximate solution of the regularized problem is available. 

Convergence theory, particularly in limited coordination settings, is presented along 

with numerical results. 


373 

2 - Controlling PHEV Recharging Through Effective Electricity 

Price Signals 

Lizhi Wang, Iowa State University, 3016 Black Engineering, Ames, 

IA, 50014, United States of America, lzwang@iastate.edu, Pan Xu 

We use a bilevel optimization model to design effective electricity price signals to 

control PHEV recharging profile. Various studies have shown that uncontrolled 

PHEV recharging could have a significant impact on the capacity adequacy, cost 

efficiency, and reliability of power systems. Price signals could serve as a demand 

management tool to control the recharging activities. Our models will be able to 

compare the effectiveness of different price signals. 

3 - A Branch-and-bound Algorithm for the Bilevel Mixed Integer Linear 

Programming Problem 

Pan Xu, Iowa State University, 3038 Black Engineering, Ames, 

United States of America, panxu@iastate.edu, Lizhi Wang, Shan Jin, 

Sarah Ryan 

We present a new algorithm for the bilevel mixed integer linear programming 

problem. This algorithm consists of two levels of branch-and-bound. At the upper 

level, we break the non-convex feasibility region into convex pieces. At the lower 

level, we solve a linear program with complementarity constraints, which is a 

simple case of mathematical program with equilibrium constraints. 

4 - Lateral Transhipment with Customer Switching 

Wenjing Shen, Drexel University, 101 N. 33rd Street, Philadelphia, 

United States of America, ws84@drexel.edu, Xinxin Hu, Yi Liao 

We consider a lateral transshipment problem between two retailers where an 

uncertain fraction of the unfulfilled demand may switch to another retailer with 

inventory. We show that the firm with surplus inventory may not transship all 

request inventory and identify conditions when full, partial or no transshipment 

takes place. We provide sufficient conditions for a unique Nash equilibrium and 

evaluate the impact of customer switching behavior on inventory decisions and 

equilibrium profits. 

■ WA39 


Joint Session OPTIM/ ICS: 

Sponsor: Optimization/ Computing Society 


Chair: Jeff Linderoth, University of Wisconsin-Madison, 1513 University 

Av., Madison, WI, United States of America, linderoth@wisc.edu 

1 - Disjunctive Cuts for Convex Mixed Integer Nonlinear 

Program (MINLP) 

Mustafa Kilinc, Graduate Student, University of Wisconsin-Madison, 

3226 Mechanical Engineering Building, 1513 University Avenue, 

Madison, WI, 53706, United States of America, kilinc@wisc.edu, 

Jeff Linderoth, James Luedtke 

Stubbs and Mehrotra [1999] generalized the disjunctive cutting plane method of 

Balas et al. into a branch-and-cut method for convex MINLPs which generates cuts 

by solving a convex projection problem in a higher dimensional space. Thus, it is 

computationally expensive to be included in a MINLP solver in practice. Our new 

method achieves this by solving a cut generating linear program iteratively. 

Computational results shows significant improvements in root node gap closure and 

solution times. 

2 - New Linear Relaxations for Quadratically Constrained Quadratic 

Programming Problems 

Mahdi Namazifar, University of Wisconsin-Madison, 1513 University 

Ave, Madison, WI, United States of America, namazifar@wisc.edu, 

Jeff Linderoth, James Luedtke 

We study a novel approach to build polyhedral relaxations for nonconvex 

quadratically constrained quadratic programming (QCQP) problems. This approach 

considers all of the constraints of the problem at once and tries to find tight 

relaxations for the problem which are reasonable in size. We present numerical 

comparisons in terms of lower bounds and relaxation size. 

3 - Inequalities for a Nonseparable Quadratic Set 

Hyemin Jeon, University of Wisconsin-Madison, Room 3227, 

Mechanical Engineering Bldg., 1513 University Avenue, Madison, 

United States of America, jeon5@wisc.edu, Jeff Linderoth 

We consider a nonseparable quadratic set which appears in applications such as 

portfolio optimization, and investigate ways to obtain a good approximation of its 

convex hull. Our work starts from transforming the set using Cholesky factorization, 

and studying its linear outerapproximation to obtain strong valid inequalities. Lifting 

plays a vital role in generating these inequalities.

WA40 

4 - Pooling Problems with Binary Variables 

Jeff Linderoth, University of Wisconsin-Madison, 1513 University 

Av., Madison, WI, United States of America, linderoth@wisc.edu, 

James Luedtke, Claudia D’Ambrosio, Andrea Lodi, Andrew Miller 

The pooling problem is a bilinear program that models linear blending in a network. 

Often, pooling problems contain binary variables that model network design issues. 

We study how to tighten relaxations of pooling problems with binary variables by 

studying the convex hull of simple sets associated with these problems. 

■ WA40 


Solver APIs II 

Cluster: John Forrest-fest | COIN-OR 10th (Joint Cluster Computing) 


Chair: Matthew Saltzman, Clemson University, Department of 

Mathematical Sciences, Martin Hall, Box 340975, Clemson, SC, 29631, 

United States of America, mjs@clemson.edu 

1 - The COIN-OR Open Solver Interface: A Status Report 

Matthew Saltzman, Clemson University, Department of 

Mathematical Sciences, Martin Hall, Box 340975, Clemson, SC, 

29631, United States of America, mjs@clemson.edu 

The Open Solver Interface (OSI) is the oldest multi-solver API in COIN-OR. We 

discuss its current status and plans for the future. 

2 - A COIN OSI Solver Plugin for Microsoft Solver Foundation 

Lou Hafer, Simon Fraser University, Burnaby, BC, V5A 1S6, Canada, 

lou@cs.sfu.ca 

MsfOsi implements a solver plugin that allows Microsoft Solver Foundation to use 

COIN solvers through the OSI API. There are interesting differences in design 

philosophy. This talk will examine what it takes to bridge the gap. 

3 - The Optimization Services Solver Interface 

Kipp Martin, University of Chicago, 5807 South Woodlawn, Chicago, 

IL, 60637, United States of America, kmartin@chicagobooth.edu 

In this talk we describe how to use the Optimization Services (OS) project to 

interface with COIN-OR solvers. The OS interface is quite flexible and allows the 

user to generate linear and nonlinear instances for solvers. In addition, there is an 

interface for solver options and solver results. 

■ WA41 


Vehicle Routing I 


Chair: Ahmed El-Nashar, Doctoral Student, University of Central Florida, 

4000 Central Florida Blvd, Orlando, FL, 32816, United States of America, 

aelnasha@mail.ucf.edu 

1 - Efficient School Bus Routing for Special Needs Students 

Behrooz Kamali, University of Arkansas, 4207 Bell Engineering 

Center, Fayetteville, AR, 72701, United States of America, 

bkamali@uark.edu, Ed Pohl, Scott J. Mason 

Special needs and medically fragile students ride specialized buses to and from 

school daily. Unfortunately, special needs service-to-school assignments are often 

made without any consideration of a student’s geographical location. We present 

optimization models that seek to improve administration-based performance metrics 

via smarter network assignments and effective bus routing decisions. 

2 - A New Capacitated Path Covering Problem 

Macarena Donoso, PhD Student, Diego Portales University, Ejercito 

441, Santiago, Chile, macarena.donosop@gmail.com, Ignacio Basulto 

We formulated and solve a particular capacitated vehicle routing problem. A path is 

built between two points of the network, for every vehicle of the fleet. Every route 

exceeds neither the vehicle capacity nor the maximum time of the trip. The 

objective is minimize the travelling cost. The nodes that to be not covered, must be 

inside a distance coverage to the the network. We propose an integer programming 

formulation and, an heuristic algorithm of resolution based on ants colonies was 

developed. 


374 

3 - A Near Optimal Algorithm for Multivehicle Dispatching and Routing 

with Time Windows 

Ahmed El-Nashar, Doctoral Student, University of Central Florida, 

4000 Central Florida Blvd, Orlando, FL, 32816, United States of 

America, aelnasha@mail.ucf.edu, Dima Nazzal 

We propose a metaheuristic for solving the VRPTW for a depot with limited number 

of docks. The metaheuristics clusters customers into groups based on their proximity 

to one another. A modified local improvement algorithm is applied to each cluster 

to find the best sequence for visiting customers with different dispatching times. 

Finally, an assignment problem formulation is used to determine the dispatching 

time and the visiting sequence for each vehicle to minimize the total traveled 

distance. 

■ WA42 


Computational Optimization and Applications I 

Sponsor: Optimization/Computational Optimization and Software 



Chair: Michele Samorani, PhD Candidate, Leeds School of Business, 

University of Colorado at Boulder, UCB 419, Boulder, CO, 80309-0419, 

United States of America, Michael.Samorani@Colorado.EDU 

1 - Solving Hard Combinatorial Optimization Problems as Implicit 

Hitting Set Problems 

Erick Moreno-Centeno, Assistant Professor, Texas A&M University, 

Industrial and Systems Engineering, College Station, 

United States of America, e.moreno@tamu.edu, Richard Karp 

The hitting set (HS) problem is: given a set U and a family S of subsets of U, find a 

minimum-cardinality set that intersects each set in S. In the implicit HS problem 

(IHS), S is given via an oracle which verifies that a given set is a HS or returns a 

not-intersected set from S. Many NP-hard problems can be solved as IHS. We solve 

IHS by combining efficient heuristics and exact methods. We present computational 

results for the minimum-feedback-vertex-set and the maximum-weight-trace 

problems. 

2 - Envy Quotes and the Iterated Core-Selecting Combinatorial Auction 

Abraham Othman, Graduate Student, Computer Science 

Department, Carnegie Mellon University, 5000 Forbes Ave, 

Pittsburgh, PA, 15213, United States of America, 

aothman@cs.cmu.edu, Tuomas Sandholm 

We describe an iterated combinatorial auction in which the allocation and prices 

converge to a solution in the core of the agents’ true valuations. In each round of 

the iterative mechanism, agents act on hints that suggest the prices of the bundles 

they are interested in. Prior work has required agents to have perfect information 

about every agent’s valuations to achieve a solution in the core. Here a core solution 

is reached even in the private value setting. 

3 - Data Mining Driven Neighborhood Search 

Michele Samorani, PhD Candidate, Leeds School of Business, 

University of Colorado at Boulder, UCB 419, Boulder, CO, 80309- 

0419, United States of America, Michael.Samorani@Colorado.edu, 

Manuel Laguna 

Metaheuristic approaches based on neighborhood search escape local optimality by 

applying predefined rules and constraints. Our general approach learns (offline) the 

guiding constraints that, when applied online, will result in effective escape 

directions from local optima. The user must define the neighborhood and provide 

an attribute representation of a solution and of a pair of solutions. We show our 

results on a set of task allocation and matrix bandwidth minimization problems.

■ WA44 

C - Room 2, Level 2- Mezzanine 

Resource Allocation in Healthcare 

Sponsor: Health Applications 


Chair: David Hutton, Stanford University, Palo Alto, CA, 

United States of America, billdave@stanford.edu 

1 - Cost-Effectiveness of Screening and Treating Acute HIV Infection in 

Men Who Have Sex with Men 

Jessie Juusola, PhD Candidate, Stanford University, Dept of 

Management Science & Engineering, 499 Terman Engineering 

Center, Stanford, CA, 94305, United States of America, 

jjuusola@stanford.edu, Eran Bendavid, Doug Owens, Margaret 

Brandeau, Elisa Long 

Given the highly infectious nature of acute HIV infection, identifying and treating 

acutely infected individuals could play a significant role in reducing HIV 

transmission. We develop a dynamic compartmental model of the HIV epidemic in 

the US to estimate the costs and health benefits of screening for acute infection and 

treating acutely infected men who have sex with men. We find such programs likely 

to be a cost-effectiveness method of reducing the burden of HIV in this high-risk 

group. 

2 - Initiatives Management in Public Healthcare Administration 

Zehra Bilginturk Yalcin, University of Texas, Austin, Operations 

Research and Industrial Eng, Graduate Program, Austin, TX, 78712, 

United States of America, zehra.yalcin@gravitant.com, Ilyas M. Iyoob 

A Public Healthcare Administration agency in a large state in the US is constantly 

running multiple initiatives within the organization. The agency faces the issue of 

selecting, prioritizing and scheduling initiatives based on the objectives of interest at 

the time, while maintaining dependencies between initiatives as well as limited 

resources and Federal budget constraints. The problem is modeled as an MIP and 

serves as a useful decision support tool in Public Healthcare Administration. 

3 - Modeling Cost-Effectiveness Data for Medical Decision Making: 

A Statistical Framework 

Megan DeFauw, University of Michigan, mcdefauw@umich.edu, 

Vijayan Nair, Joseph Norman, Allison Rosen 

The effect of treatment on cost and QALYs in CEA is determined through simulation 

or a clinical study. In either case, the mean cost-effectiveness ratio is a poor 

characterization of treatment effect. We develop a statistical framework exploiting 

the stochastic evolution of disease processes over time to characterize the effect of 

treatment and also exploit the concept of stochastic dominance to aid in decisionmaking 

when the utility function is unknown. Finally, we illustrate with an 

example. 

■ WA45 


Pierskalla Finalists V 



Chair: Mariel Lavieri, Assistant Professor, University of Michigan, 1205 

Beal Avenue, Ann Arbor, MI, 48109-2117, United States of America, 

lavieri@umich.edu 

1 - Real-time Differentiation of Nonconvulsive Status Epilepticus From 

Other Encephalopathies using Quantitative EEG Analysis: 

A Pilot Study 

Jicong Zhang, PhD Candidate, University of Florida, 303 Weil Hall, 

ISE Department, Gainesville, FL, 32611, United States of America, 

jicong@ufl.edu, Panos Pardalos, Chang-Chia Liu, Petros 

Xanthopoulos, Scott Bearden, Basim M. Uthman 

Generalized NonConvulsive Status Epilepticus (NCSE) and some non-epileptic 

encephalopathies have similar clinical symptoms and exhibit similar EEG 

waveforms. To distinguish NCSE from some non-epileptic encephalopathies is 

difficult and significant clinically. Nonlinear dynamics are extracted from EEG and 

classifiers are designed to differentiate NCSE and toxic/metabolic encephalopathy. 

The results showed strong evidence that nonlinear dynamic measures can be useful 

in clinical diagnosis of NCSE. 


375 

2 - A Bilevel Model for Designing Preventive Healthcare 

Facility Networks 

Yue Zhang, Assistant Professor, The University of Toledo, 2801 West 

Bancroft Street, Toledo, OH, 43606, United States of America, 

Yue.Zhang@sauder.ubc.ca, Oded Berman, Vedat Verter, 

Patrice Marcotte 

This paper presents a methodology for designing a network of preventive healthcare 

facilities to improve its accessibility to potential clients, so as to maximize 

participation to preventive healthcare programs. We formulate the problem as a 

mathematical program with equilibrium constraints. We use the model to analyze 

an illustrative case, the network of mammography centers in Montreal. A number 

of interesting results and managerial insights are discussed. 

3 - Inferring Model Parameters in Network-based Disease Simulation 

Eva Enns, Stanford University, 117 Encina Commons, Stanford, CA, 

94035, United States of America, evaenns@stanford.edu, 

Margaret Brandeau 

Many models of infectious disease ignore the underlying contact structure through 

which the disease spreads. However, in order to evaluate the efficacy of certain 

disease control interventions, it may be important to include this network structure. 

We present a network modeling framework of the spread of disease and a 

methodology for inferring important model parameters, such as those governing 

network structure and network dynamics, from readily available data sources. 

4 - The Mayo Clinic Optimizes Patient Transport Staffing 

Dustin Kuchera, MS, Business Analyst, Mayo Clinic, 626 8th St SW, 

Rochester, MN, 55902, United States of America, 

kuchera.dustin@mayo.edu, Thomas R. Rohleder, PhD 

In this paper, we report on the implementation of simple integrated queuing and 

mathematical programming methods to optimize staffing for patient transport at the 

Mayo Clinic. A tool was developed and implemented in Microsoft Excel and Visual 

Basic for Applications and includes an easy-to-use interface. Results of the 

implementation include significant staff savings via more efficient scheduling. 

5 - Prioritization of Medical Equipment for Maintenance Decisions 

Sharareh Taghipour, PhD Candidate, University of Toronto, 

5 King’s College Road, Toronto, ON, M5S 3G8, Canada, 

sharareh@mie.utoronto.ca, Dragan Banjevic, Andrew K.S. Jardine 

Clinical engineering departments in hospitals are responsible for establishing and 

regulating a Medical Equipment Management Program to ensure that medical 

devices are safe and reliable. To mitigate functional failures, significant and critical 

devices should be identified and prioritized. We present a multi-criteria decisionmaking 

model to prioritize medical devices according to their criticality. 

■ WA47 


New Directions in Project Management 

Cluster: Topics in Project Management 


Chair: Willy Herroelen, Emeritus Professor, K.U.Leuven, Research Center 

for Operations Management, Department of Decision Sciences & 

Information Management, Naamsestraat 69, Leuven, B-3000, Belgium, 

willy.herroelen@econ.kuleuven.be 

1 - A New Approach for Project Risk Analysis 

Stefan Creemers, K.U.Leuven, Research Center for Operations 

Management, Department of Decision Sciences & Information 

Management, Naamsestraat 69, Leuven, B-3000, Belgium, 

stefan.creemers@econ.kuleuven.be, Erik Demeulemeester, 

Stijn Van de Vonder 

Most quantitative project risk analysis techniques provide insight in the risk profile 

of the project and in the feasibility of certain project completion dates. Far fewer 

efforts exist that aim at identifying the underlying risk factors that cause the project 

schedule to slip. We introduce an approach to calculate the impact that each risk 

has on the project completion date. Such an approach allows to focus mitigation 

efforts on those risks whose mitigation would be most effective. 

2 - On the Interaction Between Railway Scheduling and Resource 

Flow Networks 

Erik Demeulemeester, Professor, KU Leuven, Naamsestraat 69, 

Leuven, B-3000, Belgium, erik.demeulemeester@econ.kuleuven.be, 

Wendi Tian 

In previous research, we have shown that in realistic situations railway scheduling 

improves both the stability and the expected project length over roadrunner 

scheduling. In this research, we introduce the concept of resource flow networks 

and analyze what the impact is of the resulting combinations on average project 

length, its standard deviation, the timely project completion probability and the 

stability cost. Extensive computational results will be presented on small and larger 

projects.

WA48 

3 - Proactive Execution Policies for the Stochastic RCPSP 

Filip Deblaere, KU Leuven, Naamsestraat 69, Leuven, B-3000, 

Belgium, filip.deblaere@econ.kuleuven.be, Erik Demeulemeester, 

Willy Herroelen 

We propose a methodology for the determination of a project execution policy for 

the stochastic RCPSP that attempts to minimize the project execution costs, defined 

as the sum of the expected costs due to activity starting time deviations and the 

expected penalties or bonuses associated with late or early project completion. We 

show that our approach significantly outperforms existing proactive scheduling 

procedures for resource-constrained projects with uncertain activity durations. 

■ WA48 


Software Demonstrations 

Cluster: Software Demonstrations 


1 - FICO - Building Optimization Applications in Xpress 

Oliver Bastert, Product Management, FICO, 901 Marquette Avenue, 

Suite 3200, Minneapolis, MN, 55402, United States of America 

This tutorial will focus on developing and deploying complete optimization 

applications using FICO’s array of mathematical modeling and optimization tools. 

These tools can be used for modeling, solving, analyzing and visualizing 

optimization problems, and integrating them seamlessly in business applications. 

During this tutorial, Bastert will explain how Xpress-Mosel, Xpress-IVE and Xpress- 

Application Developer can decrease development time for new optimization 

applications and enable you and your customers to make smarter decisions. The 

proven technologies offered by FICO can be used in range of applications such as 

supply chain management, transportation, finance, energy, manufacturing, retail, 

insurance and manufacturing industries, to name a few. 

2 - Palisade Corporation - DecisionTools Suite Software Introduction 

Thompson Terry, Senior Risk Analyst, Trainer, Consultant, Palisade 

Corporation, 798 Cascadilla Street, Ithaca, NY, 14850, United States 

of America, tterry@palisade.com 

Palisade’s DecisionTools Suite includes 7 software packages that can be used by a 

wide variety of departments and individuals within any organization to better assess 

risk and make well-informed decisions. This comprehensive example demonstrates 

how all the components of the Suite can be used together to assess the likelihood of 

success of a new product launch, determine critical variables to limit risk exposure, 

optimize the inclusion of the new product into existing manufacturing facilities, and 

plan for the expansion of both production and distribution networks. 

■ WA49 

C -Room 10, Level 2- Mezzanine 

Machine Learning and Business Intelligence 

Sponsor: Information Systems 


Chair: Maytal Saar-Tsechansky, Assistant Professor, University of Texas at 

Austin, 1 University Station, Austin, 78712, United States of America 

1 - Active Inference and Active Learning for Data Streams 

Foster Provost, Professor, New York University, 44 West 4th Street 

#8-86, New York, NY, 10012, United States of America, 

fprovost@stern.nyu.edu, Josh Attenberg 

We consider applications where predictive models are applied to a stream of 

instances that can repeat, such as web pages for ad impressions, and where there is 

a budget for applying human resources to acquire ground truth labels for carefully 

chosen instances—both for learning and for direct inference (in lieu of using the 

predictive model). We introduce strategies for allocating human resources, which 

consider: p(x), which may be highly skewed; error cost; and the value for (active) 

learning. 

2 - Bias in Cross Validation 

Claudia Perlich, Chief Scientist, Media 6 Degrees, 16 Oakrdige Rd, 

Mt Kisco, United States of America, claudia@media6degrees.com, 

Grzegorz Swirszcz 

Evaluation of model performance has a long tradition in statistics and machine 

learning. Non-parametric estimations of model performance include bootstrapping, 

jackknife, random sub-sampling and cross-validation. We show that cross-validation 

to data with low signal can lead to `holdout’ predictions with perfectly opposite 

ranking. While such a `model’ would raise suspicion, great harm can be done if it is 

integrated in an automated process that includes stacking or ensemble selection 

methods. 


376 

3 - Collaborative Information Acquisition (Talk) 

Danxia Kong, The University of Texas at Austin, 

1 University Station, Austin, TX, 78712, United States of America, 

Danxia.Kong@PhD.mccombs.utexas.edu, Maytal Saar-Tsechansky 

Most information acquisition policies aim to improve the predictive accuracy of a 

model. However, in practice, a predictive model is used with other models to inform 

arbitrarily complex decisions. This paper discusses a new kind of collaborative 

information acquisition (CIA) policies, where multiple predictive models 

collaboratively prioritize information acquisitions to promote the decisions they 

inform. We present a framework and a specific CIA policy that yields superior 

decision performance. 

■ WA50 


Managing Workload Dependencies in the Cloud 

Cluster: Cloud Computing 


Chair: Hani Jamjoom, IBM Research, 19 Skyline Dr., Hawthorne, 

United States of America, jamjoom@us.ibm.com 

1 - Uncovering Causal Factors in HPC Workloads 

Anshul Sheopuri, Research Staff Member, IBM TJ Watson Research 

Center, 19 Skyline Drive, Hawthorne, United States of America, 

sheopuri@us.ibm.com, Zon-yin Shae, Eric Shiu, Hani Jamjoom 

Scheduling and Pricing engines are critical components of a High Performance 

Computing (HPC) system. To deploy these engines in a HPC system, it is important 

to understand the causal factors of demand fluctuations - number of users, time of 

day, time of month, etc. The objective of our work is to bring scheduling and pricing 

models in HPC to market by validating or invalidating the assumptions used for 

developing these models by uncovering the causal factors of demand fluctuations. 

2 - Using Application Dependencies for Workload Migration Decisions 

in Cloud Datacenter Environments 

Petros Zerfos, Research Staff Member, IBM T.J. Watson Research 

Center, 19 Skyline Drive, Hawthorne, NY, 10532, United States of 

America, pzerfos@us.ibm.com, Hani Jamjoom, Yew Huey Liu, 

Kang-Won Lee, Vivek Shrivastava 

Virtual machine (VM) migration can optimize the use of physical servers in cloud 

datacenters. Enterprise applications consist of multiple interdependent VMs. Current 

research optimizes for intra-server constraints (e.g., CPU); it ignores inter-server 

dependencies (e.g., network communication). Our work formulates the VM 

migration problem to also account for inter-VM dependencies, and proposes an 

online approximation. Using workloads from a datacenter, we explore the efficacy 

of our solution. 

3 - Overdriver: Enabling High Data Center Utilization Through 

Aggressive Memory Oversubscription 

Dan Williams, Cornell University, 4104 Upson Hall, Ithaca, NY, 

14853, United States of America, djwill@cs.cornell.edu, 

Hani Jamjoom, Yew Huey Liu, Hakim Weatherspoon 

With the intense competition between cloud providers, resource oversubscription is 

essential for achieving higher utilization and profits from the underlying 

infrastructure. Resource oversubscription, however, comes at a price: it increases the 

likelihood of overload due to insufficient physical resources. We present Overdriver, 

a system that aggressively oversubscribes memory and immediately reacts to 

mitigate the effects of all types of overload, including transient overload.

■ WA51 


Optimal Routing Through Military Networks 

Sponsor: Military Applications 


Chair: Chris Odom, ENS, US Navy, Colorado School of Mines, 

Engineering Hall 816 15th Street, Golden, CO, 80401, United States of 

America, codom@mines.edu 

1 - Routing and Scheduling Supply Ships for the Combat 

Logistics Force 

Gerald Brown, Operations Research Department, Naval Postgraduate 

School, Monterey, CA, 93943, United States of America, 

ggbrown@nps.navy.mil, Matthew Carlyle, Patrick Burson, Jeff Kline, 

Anton Rowe 

We synchronously route a fleet of supply ships between ports and combatant 

customer ships throughout an area of operations, or worldwide. Key concerns 

include keeping combatant inventories above required operational safety stock 

levels, minimizing port visit and other supply ship costs, sourcing costs, and the cost 

of fuel consumed. We show how the system works and what we have learned in 

practice along with our Military Sealift Command sponsor. 

2 - Interdicting Networks to Competitively Minimize Evasion with 

Synergy Between Applied Resources 

Brian Lunday, Assistant Professor, Department of Mathematical 

Sciences, U.S. Military Academy, Department of Math. Sci. 

(Building 601), United States Military Academy, West Point, NY, 

10996, United States of America, lunday@vt.edu, Hanif Sherali 

We examine the problem of minimizing the maximum probability of evasion by an 

entity traversing a network from a given source-and-terminus, incorporating novel 

forms of superadditive synergy between resources applied to arcs in the network. 

We propose an alternative model for sequential overt and covert deployment of 

subsets of interdiction resources, and conduct comparative analyses between models 

for purely overt (with or without synergy) and composite overt-covert strategies. 

3 - New Results on the Network Diversion Problem 

Christopher Cullenbine, PhD Candidate, Colorado School of Mines, 

1500 Illinois Street, Golden, CO, 80401, United States of America, 

ccullenb@mines.edu, Kevin Wood, Alexandra Newman 

The network-diversion problem seeks a minimum-weight, minimal s-t cut in a 

graph that contains a pre-specified edge. The problem arises in intelligencegathering 

and war-fighting scenarios. We use Lagrangian relaxation with a stronger 

integer linear-programming formulation to modify edge weights for near-minimumweight 

cut enumeration. We also describe new NP-completeness results, 

polynomially solvable special cases, and provide computational results that show 

improvements over the original. 

4 - Routing Military Vehicles in a Threat Environment Accounting for 

Arc-Dependent Risk Costs 

Chris Odom, ENS, US Navy, Colorado School of Mines, Engineering 

Hall 816 15th Street, Golden, CO, 80401, United States of America, 

codom@mines.edu, Alexandra Newman, Kevin Wood 

A directed graph with nodes representing waypoints and sets of arcs denoting paths 

for vehicle transit models the area of operations. We develop an integer program 

and modify an enumeration algorithm to determine a path that maximizes the 

probability of mission success subject to flow balance and side constraints. We 

account for arc-dependent risk costs for the case in which a threat contributes risk 

to two or more arcs, and we provide preliminary numerical results. 

■ WA52 


Decision Support for Military Operations 

Cluster: Homeland Security and Defense 


Chair: Janet Spoonamore, Army Research Office, P.O. Box 1221, 

Research Triangle Park, NC, 27709, United States of America, 

janet.spoonamore@us.army.mil 

1 - Solution Method for Large Scale Chance-constrained Problems - 

Application Potential to Defense and Homeland Security Problems 

Miguel Lejeune, George Washington University, 2201 G Street, NW 

Funger 415, Washington, DC, 20052, United States of America, 

mlejeune@gwu.edu 

We introduce a new method for the solution of chance-constrained problems. It is 

based on the extraction of patterns which define sufficient/minimal conditions for 

the chance constraint to hold and allows the exact reformulation of the stochastic 

problem as a linear problem. The method is computationally very efficient and 


377 

allows the very fast solution of problems in which up to 50000 scenarios are used to 

characterize uncertainty. Applications to defense and homeland security problems 

will be discussed. 

2 - Army Research Office Program - Decision and Neural Sciences 

Janet Spoonamore, Army Research Office, P.O. Box 1221, 

Research Triangle Park, NC, 27709, United States of America, 

janet.spoonamore@us.army.mil 

The Decision and Neuro Sciences program addresses development of new advanced 

modeling, simulation, optimization and other analysis methodologies to support 

command-level decision-making at the operational level. The program includes two 

thrusts - one addressing advanced numerical methods - especially addressing 

stochastic behaviors and second addressing modeling of cognitive and neural 

phenomenology of decision making. 

3 - Tactical Behavior Composition 

Evan Clark, Intelligent Automation, Inc., 15400 Calhoun Drive, 

Suite 400, Rockville, MD, 20855, United States of America, 

eclark@i-a-i.com, Jeffrey Smith 

Behavior composition for computer generated forces is a technique that facilitates 

the creation and validation of agent behavior. It refers to the practice of creating 

reusable primitives that can be combined to construct new complex agent 

behaviors. Research in behavior composition has often focused on the use of 

procedural primitives. This paper discusses a framework for commander agent 

behavior composition that includes not only procedural primitives, but also those 

representing tactical concepts such as spatial relationships, subordinate coordination, 

terrain analysis, firepower and mobility. These primitives give the domain expert the 

ability to influence the manner in which tactical decisions are made. These 

primitives are elements of a tactics description language called Tesla Using the Tesla 

language, a tactical behavior expert composes tactic templates which can later be 

used by commander agents in course of action development and to solve tactical 

problems. 

4 - Resource-constrained Project Scheduling Under Uncertainty: 

Models, Algorithms and Applications 

Haitao Li, University of Missouri - St. Louis, 229 CCB, One 

University Blvd, St. Louis, MO, 63121, United States of America, 

lihait@umsl.edu 

This research studies the stochastic resource-constrained project scheduling problem 

(SRCPSP), which has a wide range of applications in machine scheduling, supply 

chain design, project portfolio optimization and personnel/manpower optimization. 

The SRCPSP is modeled as a Markov decision process (MDP) to cope with both 

structural and non-structural randomness. Techniques from optimization, artificial 

intelligence, simulation and statistics are built into an approximate dynamic 

programming (ADP) framework to tackle the “curses of dimensionality”. 

Preliminary computational results on the deterministic RCPSP are promising. 

■ WA53 


Marketing 


Chair: Nancy Ryan, Associate Professor of Marketing, St. Edward’s 

University, 3001 South Congress Avenue, Austin, TX, 78701, 

United States of America, nancym@stedwards.edu 

1 - Differences Between Imagers and Verbalizers in Users’ 

Preference-making 

Sangwon Lee, The Pennsylvania State University, 343 Leonhard 

Building, University Park, PA, 16802, United States of America, 

sangwon.advance@gmail.com, Richard Koubek 

Understanding target users is a crucial issue in establishing design and marketing 

strategies for computer-based applications. To contribute to the comprehension of 

target users, this study examines the effects of cognitive style (imagers vs. 

verbalizers) on user preference based on usability and aesthetics through an 

experiment using four simulated systems with different levels of usability and 

aesthetics. 

2 - The Market Entry Decision Based on Time-series Country Efficiency 

Gary Chao, Kutztown University, P.O. Box 730, Kutztown, PA, 

19530, United States of America, chao@kutztown.edu, Maxwell Hsu 

Companies may examine country ranking reports before they expand business 

operations abroad. Most reports focus on some socio-economic variables and 

compute the country ranking results using a weighted averaging method. This 

research picks up an attractive foreign market by taking advantage of the data 

envelopment window analysis (DEWA) in finding a relatively more efficient 

marketplace to invest. We apply to evaluate the efficiency of 22 countries based on 

the globalEDGE data from 2002 to 2008.

WA54 

3 - Let Me Stack Them Up: An Analysis of Internet Marketing 

Service Contract 

Ryan Choi, PhD Student, University of California, Irvine, 6454 

Adobe Circle, Irvine, CA, United States of America, jihungc@uci.edu 

We examine two service contracts which Amazon.com offers to its competing 

sellers; i.e., Selling on Amazon (SOA) and Fulfillment by Amazon (FBA). We 

analyze how Amazon is incentivized to offer one or both of these contracts. 

Compared to pure competition, either contract is more likely offered when 

consumers’ valuation gap gets small. Assuming that this valuation gap disappears 

once a contract is signed up, Amazon’s contracting decision varies depending on 

which seller has a cost advantage. 

4 - Attracting Customers with Attractive Signage and Architecture 

Nancy Ryan, Associate Professor of Marketing, St. Edward’s 


United States of America, nancym@stedwards.edu, Helene Caudill 

Retail establishments should not overlook the importance of signage and 

architecture. We conducted a laboratory study using over 125 students who 

reviewed the exterior signs and architecture of five different restaurants. Results 

reveal how important these two features are in terms of customers’ intentions to eat 

there. 

■ WA54 


Technology Assessment and Forecasting II 

Sponsor: Technology Management/New Product Development 


Chair: Fred Phillips, Professor, Alliant International University, Avenue of 

Nations, San Diego, CA, 92131, United States of America, 

fphillips@alliant.edu 

1 - How do Small Biotechnology Firms Innovate? The Cases 

From Taiwan 

Yu-Shan Su, Associate Professor, National Taiwan Normal University, 

162, HePing East Road, Section 1, Taipei, 106, Taiwan - ROC, 

bellesu222@yahoo.com.tw 

How do small firms innovate? How do a small firm’s internal capabilities and 

external alliances contribute to its innovativeness? The main purpose of this study is 

to adopt theoretical angle of open innovation to discuss small biotechnology firms in 

Taiwan. This study offers an integrated perspective of open innovation in the 

literature. 

2 - Forecasting Timely Revolutions in Organizational Performance 

Charles Weber, Associate Professor, Portland State University, P.O. 

Box 751, Engineering and Technology Management, Portland, OR, 

97207, United States of America, charles.weber@etm.pdx.edu, 

Nitin Mayande 

It has been shown that subsystem-level learning activities and firm-exogenous 

learning activities can induce delayed, timely revolutions in organizational 

performance. This paper develops a method for forecasting revolutions in 

organizational performance in an open innovation system. Firm-internal and firmexternal 

factors are taken into consideration. 

3 - Forecasting the Development of Clean Coal and Natural 

Gas Technologies 

Christopher Ordowich, SRI International, 1100 Wilson Blvd, 

Arlington, VA, United States of America, 

christopher.ordowich@sri.com, John Chase 

This study estimates the costs and efficiencies of several types of coal and natural gas 

power plants with and without carbon capture technologies through 2050. 

Improvements in plant efficiency and reductions in capital and O&M costs are 

modeled using technology learning curves established by a detailed analysis of 

historic performance data. Combined with demand and input cost forecasts, the 

learning curves were used to project the cost and adoption of each plant type over 

time. 

4 - Exploring the Societal Dimensions of IT: A Look at the Future of 

Sustainable IT Services 

Robert Harmon, Professor of Marketing & Technology Management, 

Portland State University, P.O. Box 751, Portland, OR, 97207, 

United States of America, harmonr@pdx.edu, Haluk Demirkan 

Green IT, the first wave of sustainable IT, developed strategies for reducing energy 

costs and carbon footprints, primarily in data centers. Green IT has been productoriented 

and internally focused within the IT function. The second wave of 

sustainable IT will be service-oriented and focused beyond the IT function to serve 

the organization’s business ecosystem and society at large. This work explores the 

key dimensions driving the development and applications of sustainable IT services. 


378 

5 - Inter-institutional Relationships and Emergency Management 

Fred Phillips, Professor, Alliant International University, and General 

Informatics LLC, 10622 Sunset Ridge Drive, San Diego, CA, 92131, 

United States of America, fp@generalinformatics.com 

Public disasters from the Exxon Valdez spill to the US mortgage meltdown involve 

many agencies. In improved networks of institutions, the fox will not guard the 

henhouse, accountability is enhanced rather than clouded, and remediation is 

quick, with blame assigned later. This paper advocates a new field of High- 

Performance Inter-Organizational Interaction (HPII). It maps dimensions of HPII 

against an extended Multiple Perspectives schema. Recent disasters and research 

directions are discussed. 

■ WA55 


New Product Innovation Strategy 



Chair: Sreekumar Bhaskaran, Assistant Professor, 

Southern Methodist University, 6212 Bishop Blvd, Dallas, TX, 75205, 

United States of America, sbhaskar@mail.cox.smu.edu 

1 - Managing Delegated Search Over Design Spaces 

Sanjiv Erat, University of California San Diego, San Diego, CA, 

United States of America, serat@ucsd.edu, Vish Krishnan 

Organizations increasingly seek solutions to open-ended design problems by 

employing an approach where the search over a solution space is delegated to 

outside agents. We study this new class of problems, and through an analytical 

model, we examine the relationship between problem specification, award 

structure, and breadth of solution space searched by outside agents towards 

characterizing how a firm should effectively manage such open-ended design 

contests. 

2 - Drivers of Value and Growth: An Examination of Innovation in the 

Solar Energy Supply Chain 

Jane Davies, University of Cambridge, Judge Business School, 

Trumpington Street, Cambridge, CB2 1AG, United Kingdom, 

j.davies@jbs.cam.ac.uk, David Kirkwood 

The introduction of government incentives has seen a plethora of firms enter the 

solar sector. These include both start-ups and firms diversifying from other 

industries. Along with incumbents, these firms face the dual pressures of reducing 

production costs while increasing the technology efficiency of solar power. We 

combine secondary data and case study analysis of 70 solar firms to show that 

process and technological innovation have differential effects on revenue growth 

and market value. 

3 - Design and Introduction of Conspicuous Durable Products 

Vishal Agrawal, Georgia Institute of Technology, 800 W Peachtree St 

NW, Atlanta, GA, United States of America, 

Vishal.Agrawal@mgt.gatech.edu, Stylianos Kavadias, Beril Toktay 

We study the implications of exclusivity-seeking consumer behavior on the design 

and introduction decisions for a durable product, namely the durability and pricing 

choices of the firm. We draw upon the traditional market models of vertically 

differentiated durable products to incorporate exclusivity-seeking behavior, and 

show that firms should instead consider designing products that undergo slow value 

erosion in conjunction with a high-price, low-volume product introduction strategy. 

4 - Product Introduction Timing for Start-ups 

Sinan Erzurumlu, Assistant Professor, Babson College, Tomasso 123, 

Babson Park, MA, 02457, United States of America, 

serzurumlu@babson.edu, Sreekumar Bhaskaran, Karthik 

Ramachandran 

We study how a start-up organization should structure its development process. 

While cash constraints pressure the firm to launch a product as soon as possible (to 

avoid or delay bankruptcy), it could affect the future products. We develop optimal 

policies for the start-up firm to determine whether and when to launch an existing 

product under technological uncertainty about future development of products.

■ WA56 


Simulation Optimization and Global Optimization of 

Expensive Functions 

Sponsor: Simulation Society 


Chair: Peter Frazier, Assistant Professor, Cornell University, 232 Rhodes 

Hall, Ithaca, NY, 14853, United States of America, pf98@cornell.edu 

1 - Simulation-Optimization Methods for Resource Allocation to 

Achieve Equity 

Muer Yang, University of Cincinnati, Dept of QAOM, College of 

Business, Cincinnati, OH, 45221, United States of America, 

yangmr@mail.uc.edu, Theodore Allen, Michael Fry, David Kelton 

This paper uses simulation-optimization models to allocate limited resources under 

uncertainty to achieve equity. By equity, we mean that all jobs, customers, etc. 

should have approximately equivalent performance measures (e.g., time in queue). 

We can provide probabilistic guarantees of global optimality for these allocations. 

We present one possible application of this work in allocating voting machines to 

precincts for local and national elections. 

2 - Global Optimization with Reponse Surfaces for Expensive 

Simulation Models 

Christine Shoemaker, Professor, Cornell University, 210 Hollister 

Hall, Cornell University, Ithaca, NY, 14853, United States of America, 

cas12@cornell.edu 

We will describe an effective algorithm for global optimization of simulation models 

that are computationally expensive and apply it to several problems, including an 

environmental simulation model involving systems of nonlinear partial differential 

equations. We will demonstrate applications of up to 30 decision variables. 

3 - Noise-Tolerant Bayesian Bisection 

Peter Frazier, Assistant Professor, Cornell University, 232 Rhodes 

Hall, Ithaca, NY, 14853, United States of America, pf98@cornell.edu, 

German Gutierrez, Shane Henderson 

We consider the stochastic optimization problem where we observe noisy 

derivatives of the objective function. These derivatives observations are expensive to 

obtain, and so our goal is to optimize the function as accurately as possible with a 

bounded number of observations. We derive the optimal sequential sampling policy 

using dynamic programming for a simplified version of this problem, and then 

discuss the use of this policy in the original problem. 

■ WA57 


Recent Advances in Portfolio Optimization 

Sponsor: Financial Services Section 


Chair: Chaithanya Bandi, Massachusetts Institute of Technology, 77 Mass 

Ave, Cambridge, MA, 02139, United States of America, cbandi@mit.edu 

1 - Fairness in Multi-account Optimization with Transaction Costs 

Dan Iancu, Stanford University, Stanford GSB, Stanford, CA, United 

States of America, Iancu_dan@gsb.stanford.edu, Dimitris Bertsimas, 

Nikolaos Trichakis 

In this work, we focus on the problem faced by fund managers when executing 

rebalancing trades for multiple accounts simultaneously. We formulate a model that 

addresses two key issues, namely how to split the trading costs across the accounts, 

and how to incorporate all the information in a scheme for rebalancing the accounts 

in a fair way. 

2 - Estimating the NIH Efficient Frontier 

Dimitrios Bisias, PhD Student, MIT, 70 Pacific Street Apt.327, 

Cambridge, MA, 02139, United States of America, dbisias@mit.edu, 

Andrew Lo, James Watkins 

The National Institutes of Health (NIH) is the preeminent source of funding for 

biomedical research, hence its funding allocation decisions have enormous impact 

on public health and social welfare. Modern financial portfolio theory is used to 

estimate the risk/reward trade-offs of NIH allocation decisions by treating 

appropriation as an investment and the decrease in years of life lost as the 

investment return. 


379 

3 - Optimal Portfolio Strategy with Liquidity Capacity 

Wei Chen, Analytical Solutions Manager, SAS Institute Inc., SAS 

Campus Dr., Cary, NC, 27513, United States of America, 

Wei.Chen@sas.com 

Liquidity is essential for banks to maintain long-term viability. An optimal portfolio 

strategy should not only manage the cash flow gaps but also provide adequate low 

cost contingent funding sources in distressed events. This paper proposes a model of 

finding minimal cost portfolio with a chance constrained level of positive excess 

cash flow and self-sustaining liquidity capacity. The model results in a tractable 

linear programming problem using the scenario based approach. 

■ WA58 


Finance-Risk Management 


Chair: D.J. Alexander-Houle, Adjunct; Program Manager, University of 

Phoenix, HP, 14207 Torrey Vista Dr, Houston, TX, 77014, United States of 

America, dahoule@sbcglobal.net 

1 - A Coherent Valuation Approach for Valuing Risky Projects 

S. Reza Seyedshohadaie, Texas A&M University, 3131 TAMU, 

College Station, United States of America, sreza@tamu.edu, 

Sergiy Butenko, Ivan Damnjanovic 

We present a coherent valuation approch for valuing risky projects in partially 

complete markets. The model is based on the exposure to risk and the trade-off 

between risk and reward. We demonstrate the application of the model on a largscale 

engineering project. 

2 - Estimation Error Reduction in Portfolio Optimization with 

Conditional Value-at-Risk 

Gah-Yi Vahn, PhD Student, UC Berkeley, 4141 Etcheverry Hall, 

Mail Code 1777, Berkeley, CA, 94704, United States of America, 

gyvahn@berkeley.edu, Andrew Lim 

We introduce a novel method of obtaining robust solutions to data-driven portfolio 

optimization with Conditional Value-at-Risk (Expected Shortfall). This method can 

be interpreted as penalizing model uncertainty in the optimization problem. We 

present some analysis of the method and empirical results that show the superior 

performance of this method when the underlying log-return data is both wellbehaved 

(multivariate Gaussian) and heavy-tailed (multivariate t). 

3 - Dynamic Models for Consumer Default Risk 

Jonathan Crook, Professor, University of Edinburgh, Credit Research 

Centre, Business School, 50 George Square, Edinburgh, EH10 4SW, 

United Kingdom, j.crook@ed.ac.uk, Tony Bellotti 

This talk explains the use of survival analysis for the prediction of default for 

consumer loans. The talk will explain how macroeconomic variables can be 

incorporated into the model and show results from the parameterisation of such a 

model and the predictive performance of it using a data set relating to credit cards. 

The model is used to stress test the portfolio of loans using Monte Carlo simulation 

to estimate the Value at Risk and Expected Shortfall of the portfolio. 

4 - Risk Tolerance Tempered with Optimal Time Increments 

D.J. Alexander-Houle, Adjunct; Program Manager, University of 

Phoenix, HP, 14207 Torrey Vista Dr, Houston, TX, 77014, 

United States of America, dahoule@sbcglobal.net, G. R. Houle 

How people reason about economic choices is risk tolerance.Of social concern is the 

risk shift from pooled to personal responsibility. Using the fractal nature of people 

creates the framework for risk complexities Grable indicates as, “underlying factor 

within Ö decision frameworks” and the aggressiveness of decisions will reflect risk 

tolerance levels (2008, p. 4). Research results identifying worst case performance 

illustrates a foundation for creating systemic risk tolerance.

WA61 

■ WA61 

H - Room 400, 4th Floor 

Operations Management V 


Chair: Heping Liu, North Dakota State University, Department of 

Industrial & Manufacturing, Fargo, ND, 58108, United States of America, 

hepingliu@yahoo.com 

1 - Constraint and Flight Rule Management for Space 

Mission Operations 

John Chachere, Senior Computer Scientist, Stinger Ghaffarian 

Technologies, 1060 Arbor Road, Menlo Park, CA, 94025, United 

States of America, john.m.chachere@nasa.gov, Jeremy Frank, 

Javier Barreiro 

NASA’s Mission Operations Directorate (MOD) has formalized thousands of 

operational constraints to help govern human spaceflight missions. MOD collects, 

develops, documents and applies these constraints to ensure the safety of the crew, 

as well as proper operation of the spacecraft systems and payloads. These constraints 

are stored in human-readable documents and also used to configure tools used by 

the flight controllers who plan and fly missions. We have begun developing a novel 

capability for authoring and maintaining such constraints called the Constraint and 

Flight Rule Management system (ConFRM). ConFRM provides history tracking and 

commenting features that allow authors to trace the history of constraints 

throughout their lifecycle. ConFRM maintains links between related constraints, 

and between constraints and source information used to create the constraints. 

ConFRM uses these links to ensure consistency between constraints throughout 

their lifecycle, and provides authors and reviewers the ability to navigate between 

constraints and related data. Finally, ConFRM enables export of constraints into 

many different forms, including human readable documents and tool 

configurations, thereby eliminating manual labor and reducing transcription errors. 

2 - Reverse Ranking and Overshooting in Supply Chains with Private 

Inventory Information 

Alexandre Belloni, Assistant Professor, Duke University, 

1 Towerview Drive, Durham, NC, 27708, United States of America, 

abn5@duke.edu, Giuseppe Lopomo, Shouqiang Wang 

We study contracts for a single supplier with fixed capacity selling to multiple 

retailers who are privately informed on their inventory levels. In symmetric 

environments, instead of a balancing policy, it is optimal for the supplier to make 

retailers with lower initial inventory levels end up with a larger final positions 

(reverse ranking). We characterize when it is optimal to ``overshoot” to provide a 

larger quantity to a retailer than it would get in the corresponding centralized 

system. 

3 - Backup Agreements in Multiple Component Procurements with 

Demand Updates 

Mingchang Wu, PhD Student, University of Connecticut, 2100 

Hillside Road, Storrs Mansfield, Storrs Mansfield, CT, 06269, United 

States of America, mingchang.wu@business.uconn.edu, Suresh Nair, 

Lakshman Thakur 

We study a supply chain where a manufacturer buys two different components 

from two suppliers to produce one product for the selling season.This is an 

extension of Eppen and Iyer’s work,where we now look at components rather than 

end product.With each supplier,the manufacturer has a particular backup 

agreement.Our objective is to derive an optimal purchase policy which describes the 

optimal commitments to each supplier and a reaction plan to the demand updates at 

the later stage. 

4 - The Optimum Policies to the Distribution/Inventory System Problem 

in the Medical Center 

Jimmy Alexander Carvajal Beltran, Student, Universidad de los 

Andes, Carrera 1 N° 18A - 12, Bogotà, Colombia, 

ja.carvajal911@uniandes.edu.co, Ciro Alberto Amaya Guio, 

Fabiàn Andrés Castaño Giraldo, Nubia Milena Velasco Rodriguez 

This paper describes a new approach for solving the budget constrained distribution 

and inventory problem for one warehouse and n retailers. This is a typical situation 

in medical supply networks where a single warehouse supplies a set of pharmacies 

with the main objective of minimizing total costs. To solve the problem, we propose 

an iterative algorithm which uses mathematical programming methods. We present 

the principal ideas of this approach and results using random instances. 

5 - Simulation Modeling for Clinic Telephone Systems 

Heping Liu, North Dakota State University, Department of Industrial 

& Manufacturing, Fargo, ND, 58108, United States of America, 

hepingliu@yahoo.com, Jing Shi 

Clinic telephone systems can guide patients to obtain optimal healthcare delivery 

services. In our research, simulation modeling is used to optimize clinic telephone 

systems with the functions of leaving messages and call-back services. Centralization 

and decentralization are two considered telephone system frames. According to 

operator utilization and patient satisfaction level, various scenarios are designed and 

simulated. The obtained results can benefit the telephone system configuration. 


380 

■ WA62 


Aviation Applications III 


Chair: Akira Kondo, Federal Aviation Administration, 800 Independence 

Ave., SW, Washington, DC, 20591, United States of America, 

akira.kondo@faa.gov 

1 - Customer Profitability Analysis for Marketing Initiatives in a 

Fractional Airline 

Jintao Ouyang, CitationAir, 5 American Lane, Greenwich, CT, 

06831, United States of America, jouyang@citationair.com, 

Aang Daniel, Roger Zhan, Haiyuan Wang 

Due to the operation nature of fractional airlines, operation costs such as charter 

premium and position do not directly link to specific customers. A regression based 

statistical method is presented to associate these costs to each customer. We also 

identify the factors that affect the profitability of customers. These factors can be 

used for marketing initiatives such as contract renewal, customer retention, and 

devising new pricing schemes. 

2 - Optimized Airport Security Study 

Shannon Harris, Technomics, Inc, 201 12th Street South, Suite 612, 

Arlington, VA, 22202, United States of America, 

sharris@technomics.net, Joon Kim, Jaime Gonzalez, 

Elizabeth Wilson 

The current security measures implemented at airports produce “soft targets” in the 

form of lengthy queues that heighten the risk of a terrorist attack. This study 

examines Washington Dulles and designs a model to improve the allocation and 

usage of security resources. Two alternative designs make use of layered, defense-indepth 

security measures and the concept of unpredictability. The best alternative is 

selected by analyzing simulation outputs based on a value function. 

3 - A Value-based Time-Phased Bayesian Network (VTBN) for 

Augmented Safety Risk Assessment 

James Luxhoj, Professor, Rutgers University, 96 Frelinghuysen Road, 

Dept. of ISE, Piscataway NJ 08854, United States of America, jluxhoj@rci.rutgers.edu, 

Michael Morton 

The use of Unmanned Aircraft Systems (UAS) for civil applications is increasing in 

the United States. There is a need to develop advanced risk models that capture the 

complexities of integrating aviation regulations, functions, hazards and causal factors 

into a probabilistic model for augmented safety assurance. This presentation 

introduces a Bayesian Network that integrates the use of value functions and time 

phasing for modeling safety risk analysis of small UAS applications. 

4 - Departure Demand, Efficiency, Taxi-out Delay, and Unimpeded 

Taxi-out Time 

Akira Kondo, Federal Aviation Administration, 800 Independence 

Ave., SW, Washington, DC, 20591, United States of America, 

akira.kondo@faa.gov 

Taxi-out delays play an important role in the computation of key airport 

performance metrics. In this system, departure demand and resulting airport 

departure efficiency significantly impact crucial parts in System Airport Efficiency 

Rate (SAER). This study provides a new definition of taxi-out times and how 

estimation of unimpeded taxi-out time affects computation of taxi-out delays. 

■ WA63 


Decision Analysis For Public Health 

Sponsor: Decision Analysis 


Chair: Ozgur Araz, Postdoctoral Fellow, University of Texas at Austin, 

Patterson Lab 628, Austin, TX, 78712, United States of America, 

oaraz@mail.utexas.edu 

1 - Comparing Cost-effectiveness of Physical Activity Interventions 

Shinyi Wu, Assistant Professor, University of Southern California, 

3715 McClintock Ave., GER 240C, Los Angeles, CA, 90089, 

United States of America, shinyiwu@usc.edu 

We conducted a systematic review and developed a method to compare costeffectiveness 

(CE) across a wide variety of physical activity (PA) intervention 

strategies. Study quality was variable. Decision prompts were most cost-effective but 

had tiny effects. School-based and other wide-reach interventions ranked well. High 

intensity programs were least cost-effective but had the largest effect sizes. The CE, 

effect size, and study quality should all be considered when choosing PA 

interventions.

2 - A Decision Analytic Approach for Modeling School Closures During 

an Influenza Pandemic 

Ozgur Araz, Postdoctoral Fellow, University of Texas at Austin, 

Patterson Lab 628, Austin, TX, 78712, United States of America, 

oaraz@mail.utexas.edu, Sean Burke, Paul Damien, Lauren Meyers, 

Bryce van de Geijn, Alison Galvani 

In this presentation, we present a decision analytic approach for making school 

closure decisions during an influenza pandemic. We build a mass action model and 

assume the severity of the disease is uncertain for the decision makers. A decision 

tree is used to evaluate several closure and reopening decisions simultaneously 

based on their cost effectiveness. Lastly, we perform several sensitivity analyses on 

decision making parameters and present our results. 

3 - Dynamics of a Pharmaceutical Risk Sharing Agreement 

Reza Mahjoub, PhD Student, Ivey Business School, 1151 Richmond 

St., London, ON, Canada, rmahjoub@ivey.uwo.ca, Fredrik Odegaard, 

Greg Zaric 

Some new drugs such as cancer drugs could be very costly to develop and 

manufacture while their effectiveness and efficiency in real life may be unproven. A 

risk sharing agreement is a contract between the drug manufacturer and a 

healthcare payer to manage uncertainties regarding the cost and effectiveness of 

those drugs. We develop a model to examine the dynamics of a risk sharing contract 

for a drug manufacturer. 

■ WA64 


Health Care, Modeling and Optimization V 


Chair: Ilgin Acar, Anadolu University, Department Of Industrial 

Engineering, Iki Eylul Campus, Eskisehir, 26555, Turkey, 

ipoyraz@anadolu.edu.tr 

1 - Consensus by Averaging Phylogenetic Trees 

Scott Provan, University North Carolina, Department Statistics and 

Operations Research, Chapel Hill, NC, 27599, United States of 

America, Scott_Provan@UNC.edu, Ezra Miller, Megan Owen 

We introduce a new notion of the consensus tree for a set of phylogenetic trees, by 

representing them as points in the phylogenetic tree space of Billera, Holmes, and 

Vogtmann. The property of non-positive curvature of this space ensures that the 

consensus tree captures the correct notion of an “average” tree, and that there is an 

algorithm for computing this tree. We give as applications reconstructing species 

trees from gene trees and comparing the structure of blood vessels in the brain. 

2 - A Probabilistic Optimization Approach to Analyze Large Scale 

Emergency Medical System on Highways 

Ana Iannoni, Ecole Centrale Paris - Laboratoire Genie Industriel, 

Grande Voie des Vignes, Chatenay Malabry, 92295, France, 

iannoni93@hotmail.com, Reinaldo Morabito, Cem Saydam 

In this study we present optimization methods based on greedy heuristics 

embedding an approximate hypercube queuing model. The proposed approach can 

support two decisions in the operation of emergency systems on highways: the 

location and districting of the ambulances. We applied these methods to a case study 

and to instances of up to 100 ambulances. The present approach is an alternative for 

the analysis of large scale systems, which provides reasonable accuracy and 

affordable running times. 

3 - A Multi-class Open Queuing Network with Priority Discipline 

Applied to the Emergency Department 

Sumi Kim, Yonsei University, Shinchon-dong, Seodaemun-gu, Seoul, 

Korea, Republic of, sumi_kim@yonsei.ac.kr, Seongmoon Kim 

We formulate the patient flows in the emergency department using a multi-class 

open queueing network. The unique aspect of this paper is that it incorporates the 

priority discipline with the queueing network, in order to control the waiting time 

of more urgent patients. A case study based on actual data from an emergency 

department demonstrates how effective the introduced model and the policy with 

the priority discipline are in controlling the waiting time and improving the quality 

of service. 

4 - Optimal Clinical Scheduling with No-shows 

Ji Lin, PhD Candidate, Purdue University, Weldon School of 

Biomedical Engineering, 206 S. Martin Jischke Drive, West Lafayette, 

IN, 47907, United States of America, lin35@purdue.edu, 

Mark Lawley, Kumar Muthuraman 

The accessibility and efficiency of outpatient care are largely affected by the 

appointment schedules. Patient no-show causes waste of physician time and 

revenue loss. Random booking results in long waiting time and overtime. Patients 

usually have different no-show rates. Thus, we propose MDP model and optimal 

scheduling policies for heterogeneous no-show patients. Approximate Dynamic 

Programming methods are developed to solve the curse of dimensionality. 


381 

5 - Workload Scoring of Nurse to Patient Assignments 

Ilgin Acar, Anadolu University, Department Of Industrial 

Engineering, Iki Eylul Campus, Eskisehir, 26555, Turkey, 

ipoyraz@anadolu.edu.tr, Steven Butt 

This work is the first to explicitly consider incorporating travel distances into the 

construction of a nurse’s patient assignments through the construction of a 

workload score based on nurse travel distances and patient acuity measures. The 

workload scores were developed through consultation with charge nurses using 

AHP. The resulting tool displays a nurse’s assigned rooms and associated workload 

score in an accessible spreadsheet format. 

■ WA65 


Inventory Management V 


Chair: Ayse Gönül Karaarslan, PhD Candidate, Eindhoven University of 

Technology, P.O. Box 513, Eindhoven, 5600 MB, Netherlands, 

a.g.karaarslan@tue.nl 

1 - A Simulation-optimization Approach to the Inventory Mangement of 

Perishable and Substitutable Items 

Bret Myers, Villanova University, 442 Hartford Square, West Chester, 

PA, 19380, United States of America, bret.myers@villanova.edu 

A simulation-optimization approach is used to analyze inventory policy for two 

perishable and substitutable items. A set of perishable items is considered under a 

periodic system of inventory control where demand for a preferred item can be 

satisfied by a substitute item with in the event of a stockout of the preferred item. 

The retailer is faced with the decision of determining the order-up-to levels which 

maximize expected total profit. 

2 - Policy Parameter Adjustments to Meet Desired Service Level 

Requirements for (r, NQ) Inventory Models 

Yasin Unlu, University of Arkansas, 4207 Bell Engineering Center, 

Fayetteville, AR, 72701, United States of America, yunlu@uark.edu, 

Manuel D Rossetti 

This study introduces a simulation optimization based procedure for setting policy 

parameters of continuous review (r, NQ) inventory models in the face of complex 

demand cases. The experiments done with various demand scenarios show that the 

procedure yields promising results in terms of attaining specified target service 

levels. 

3 - A Multiple-level Supply Chain Coordination Model by (Q, r) Policies 

in a Fuzzy Environment 

Xinmin Wu, OR Specialist, SAS, Sas Campus Drive, Cary, 27513, 

United States of America, xinmin.wu@sas.com, Don Warsing 

we present a serial multiple-level supply chain coordination model by (Q,r) policies 

in a fuzzy risk environment. Sources of risk and uncertainty in our model include 

demand, lead time, supplier yield, which are modeled by fuzzy sets. Heuristics are 

presented to determine local optimal policies on the basis of techniques identified in 

the literature on fuzzy sets. A coordination process with an external coordinator is 

implemented to improve supply chain performance based on global view. 

4 - A Modified Base Stock Policy for an Assemble to Order System 

with Different Review Periods 

Ayse Gönül Karaarslan, PhD Candidate, Eindhoven University of 


a.g.karaarslan@tue.nl, Ton de Kok, Gudrun P. Kiesmüller 

We have a single item assembled from two components. The inventory levels are 

reviewed periodically and customer demand is stochastic. One of the components 

has a longer lead time, higher holding cost and shorter review period compared to 

the other one. We analyze a modified base stock policy such that one stockpoint is 

controlled by an order-up to policy. The orders for the other stockpoint are 

synchronized depending on its order-up-to level and the inventory level of the 

other component.

WA66 

■ WA66 


Transportation, Other 


Chair: Gokhan Memisoglu, PhD Student, Texas A&M University, 

Department of Industrial and Systems Eng, Texas A&M University, 

College Station, TX, 77843-3131, United States of America, 

gmemis@tamu.edu 

1 - A Net Present Value Model for Investment in Airport Assets 

Alexandre de Carvalho, Instituto de Pesquisa EconÙmica Aplicada - 

IPEA, Diretoria de Estudos Regionais e Urbanos, SBS Quadra 1, 

Edificio BNDES, sala 718, Brasilia, DF, 70076-900, Brazil, 

alexandre.ywata@ipea.gov.br, Alessandro VM de Oliveira, 

Reinaldo C. Garcia 

Improvement and expansion in infrastructure transport assets are among the main 

concerns facing the transport specialists due to the high investments involved. The 

timing and the combination of new transport investments is key to analyze their 

long-term effects. This paper proposes an investment model based on Net Present 

Value (NPV) analysis. In particular, the model is applied to: (i) the building of a new 

airport; and (ii) the expansion of already existing “major” and “minor” airports. 

2 - Network Connectivity Analysis for Port Competitiveness Study 

Ek Peng Chew, Associate Professor, National University of Singapore, 

10 Kent Ridge Crescent, Singapore, 669606, Singapore, 

isecep@nus.edu.sg, Loo Hay Lee, Jianlin Jiang, Chee Chun Gan 

Port connectivity is an important factor that deserves much attention in the study of 

port competitiveness. Determining how well one port connects to others is hardly 

easy, as it being an abstract concept might bring about different interpretations and 

definitions in different cases. In this paper, we will propose a measure using 

network anaylsis to measure port connectivity. 

3 - Student Assignment Models for a School System 

Trivikram Rao, PhD Student, University of Louisville, Department of 

Industrial Engineering, JB Speed School of Engineering, Louisville, 

KY, 40292, United States of America, trivikram.rao@louisville.edu, 

Arsalan Paleshi, Bulent Erenay 

This research proposes a mathematical programming approach for student 

assignment to schools while incorporating quantitative, qualitative constraints such 

as travel cost, parental preferences, and socio-economic constraints. The model’s 

sensitivity in cost, parental satisfaction terms to certain constraints is tested. Also, 

extension of this approach to other assignment problems is discussed. 

4 - Optimal Deployment of Emissions Reduction Technologies for 

Large Fleets 

Gokhan Memisoglu, PhD Student, Texas A&M University, 

Department of Industrial and Systems Eng, Texas A&M University, 

College Station, TX, 77843-3131, United States of America, 

gmemis@tamu.edu, Mohamadreza Farzaneh, Kiavash Kianfar 

In states that have serious air quality problems, such as Texas and California, public 

fleet managers are under pressure to reduce emissions from their fleet. In order to 

help them with this problem, this research study will create an optimization model 

capable of determining the most efficient assignment of emission reduction 

strategies among vehicles and equipment in a large fleet. To achieve the goal, this 

study will focus on Texas Department of Transportation’s (TxDOT’s) fleet. 

■ WA67 


Simulation I 


Chair: Ahmet Ozkul, Assistant Professor of Management, University of 

New Haven, 300 Boston Post Rd., Maxcy Hall, West Haven, CT, 06516, 

United States of America, aozkul@newhaven.edu 

1 - Incorporating Analytic Hierarchy Process with Simulation to Assess 

Health Service Quality 

Yan Li, The University of Texas-Pan American, 508 E Redbud Ave, 

Mcallen, TX, 78504, United States of America, 

tongjijacky@hotmail.com, Jianzhi Li 

While improving quality in health care is currently at the forefront of professional, 

political, and managerial attention, it has yet to fully understand the key dimensions 

constituting health-care quality and develop valid approaches to their measurement. 

The methodology proposed in this paper, AHP based simulation, will provide great 

insights in tackling this problem and open up a world of possibilities for future 

research. 


382 

2 - Modeling Parameter Uncertainty in Stochastic Simulations 

Canan Gunes, PhD Student, Carnegie Mellon University, 5000 

Forbes Avenue, Pittsburgh, PA, 15213, United States of America, 

cgunes@andrew.cmu.edu, Bahar Biller 

We consider a stochastic simulation and demonstrate how to account for stochastic 

uncertainty and parameter uncertainty in the output analysis. We further 

decompose the variance of the output data into terms associated with each type of 

uncertainty and use this decomposition to develop a data-collection schema for 

reducing parameter uncertainty. We illustrate our approach with inventory system 

simulations. 

3 - A Discrete Simulation Approach to the Design of a Warehouse for 

Air Cargo Operations 

Carlos Osorio, Politecnico Grancolombiano, Calle 57 3-00 Este Fac 

Ingenieria, Bogota, Colombia, caosorio@poli.edu.co, 

Oscar Javier Parra Ortega 

This paper shows the application of discrete simulation for designing the main air 

cargo warehouse of one of the biggest airlines in Colombia and South America. The 

research begins with the evaluation of the operations carried out in the warehouse; 

the current performance is analyzed using key logistics indicators, and prospective 

scenarios are then generated and evaluated. Preeliminary results are also reported. 

4 - Simulation to Identify Errors in Generalized Tournaments 

Christopher Keller, Assistant Professor of Marketing and Supply 


School of Business, Greenville, NC, 27858, United States of America, 

kellerc@ecu.edu 

This paper presents a Monte Carlo simulation of ranking participants in a 

generalized tournament, including variations in: number of participants; 

completeness of the tournament; and paired comparison errors. Some paired 

comparison errors are not reliably identifiable and a complete and accurate ordering 

of participants is unlikely. Using a partial ordering topological rank, jackknife 

estimates of the average rank do admit the identification of some of the paired 

comparison errors. 

5 - Simulating Social Networks in Understanding Dynamics of 

Customer Purchase Behavior: Case of Cult Brands 

Ahmet Ozkul, Assistant Professor of Management, University of New 

Haven, 300 Boston Post Rd., Maxcy Hall, West Haven, CT, 06516, 

United States of America, aozkul@newhaven.edu, Aqin Hu 

When we consider customers as actors in a purchasing relationship of products, 

Social Network Analysis may be used to reveal unknown patterns and explain 

customer behavior. Using a simulation analysis, we create a market of M customers 

and N products. Customers create a relationship when they buy a given product. 

The number of links and frequency of the purchase determine the nature of the 

relationship. We calculate social network measures in each cycle to observe the 

dynamics. 

■ WA68 


Innovation/Entrepreneurship I 


Chair: Arash Dadvand, CGN and Associates, 415 SW Washington St, 

Peoria, IL, 61602, United States of America, adadvand@cgn.net 

1 - Diffusion of Innovation Products: Network Effects and 

Bandwagon Effects 

Shuzhen Sun, School of Industrial Engineering & Management, 

Oklahoma State University, School of Industrial Engineering & Mgt., 

Oklahoma State University, Stillwater, Ok, 74078, 

United States of America, zhener18@gmail.com, Di Xu 

Bandwagon effects and network effects play important roles in consumers’ purchase 

decisions. This paper uses the small-world network model to study the two effects 

on the diffusion of innovation products. Simulations are conducted to examine the 

diffusion process and adoption rate under different levels of bandwagon pressures, 

different strength of network effects,and different network configurations, on the 

assumption that bandwagon effects exist in all conditions. 

2 - A System Dynamics Model to Understand Innovation in the 

Design Process 

Nur Ozge Ozaltin, University of Pittsburgh, 1048 Benedum Hall, 

Pittsburgh, PA, 15261, United States of America, noo7@pitt.edu, 

Mary Besterfield-Sacre, Larry Shuman 

To improve innovation, it should be better understood how the teams navigate the 

design process from initial conception to prototype. We examine the bioengineering 

design process to investigate associations between design patterns and the quality of 

the resulting artifact. We identify the critical patterns and factors that lead to 

innovation.

3 - Entrepreneurship and Community Development: A Fallacy From 

Central Mexico 

Eliseo Vilalta-Perdomo, Head Dept Industrial and Systems 

Engineering, Tecnológico de Monterrey, Av Gral Ramón Corona 

2514, Col. Nuevo México, Zapopan, 45201, Mexico, 

eliseo.vilalta@itesm.mx, Cynthia Montaudon-Tomas 

To consider entrepreneurship as a dynamo for economic development is challenged 

throughout this work done in central Mexico. Two programs from different rural 

communities in Guanajuato are studied. Probably the most promising finding is to 

recognize that individual economic development and community social 

development are not linked. A proposal to increase quality of life in rural 

communities is presented. It is centered on creating and maintaining self-organized 

web-based networks. 

4 - Efficient Targeted Concept Development Method 

Arash Dadvand, CGN and Associates, 415 SW Washington St, Peoria, 

IL, 61602, United States of America, adadvand@cgn.net 

Breakthrough competitive advantage are accomplished by competitive concepts. 

Concept development is costly, hard to manage and may not result in a viable 

solution. This is a method to formulate a problem, effectively and efficiently create 

concepts based on generic cross-industry patented solutions, and prioritize concepts 

to meet predefined goals. Final set of concepts could be optimized and improved. 

■ WA69 

H - Salon F, 6th Floor 

Transportation, Rail 


Chair: Katharina Beygang, University of Kaiserslautern, Department of 

Mathematics, OR Group, Postfach 3049, Kaiserslautern, 67653, Germany, 

beygang@mathematik.uni-kl.de 

1 - Extensions to the Online Delay Management Problem on a 

Single Train Line 

Christiane Zeck, University of Kaiserslautern, Department of 

Mathematics, OR Group, Postfach 3049, Kaiserslautern, 67653, 

Germany, zeck@mathematik.uni-kl.de, Sven O. Krumke, 

Clemens Thielen 

The online delay management problem consists in deciding when to wait for 

delayed passengers in order to minimize the total passenger delay. Viewing this 

problem in the context of game theory, we determine an optimal online strategy. 

We also introduce a new objective function modeling a refund system for delayed 

passengers with the aim of maximizing the profit. For this problem, there cannot be 

a competitive deterministic online algorithm, but we present a 2-competitive 

randomized algorithm. 

2 - Railway Routing Algorithms for Hazardous Materials 

Marc Meketon, Oliver Wyman, 212 Carnegie Center, Princeton, NJ, 

08540, United States of America, Marc.Meketon@oliverwyman.com, 

Paul Stephens 

New Federal Railroad Administration guidelines require examining various 

hazardous material routing alternatives. The shortest path may traverse large 

population centers or other non-desirable areas. Finding a set of acceptable paths is 

challenging for large railroads where the “K-shortest” paths differ insignificantly. 

This research discusses solutions to the important, but difficult, task of identifying 

different paths with acceptable distance and cost that also represent real 

alternatives. 

3 - Efficient Usage of Railway Infrastructure Through 

Pricing Mechanisms 

Arnt-Gunnar Lium, Research Scientist, SINTEF Technology and 

Society, P.O. Box 4760 Sluppen, Trondheim, N 7465, Norway, 

arnt-gunnar.lium@sintef.no, Adrian Werner 

Railroad infrastructure is very costly to develop; hence, increased utilization will 

have a significant positive impact on society. One way of increasing utilization over 

the day is to implement tariffs such that railroad operators adapt their schedules in a 

socio-economical optimal way. We look into how bi-level programming can be 

combined with stochastic service network design to determine socio-economical 

optimal tariffs. 

4 - Extensions to the Train Marshalling Problem 

Katharina Beygang, University of Kaiserslautern, Department of 

Mathematics, OR Group, Postfach 3049, Kaiserslautern, 67653, 

Germany, beygang@mathematik.uni-kl.de, Sven O. Krumke 

Shunting yards, consisting of a hump and a set of parallel classification tracks, play 

an important role in railroad life. They are used for the rearrangement of cars 

according to their destinations. The Train Marshalling Problem consists of 

minimizing the number of classification tracks needed for the rearrangement of a 

given car sequence. It is well known to be NP-complete. We give competitive 

polynomial time algorithms for the online variant as well as lower bounds on the 

competitiveness. 


383 

■ WA70 

H - Salon G, 6th Floor 

Planning and Strategies for Airline Operations 

Sponsor: Aviation Applications 


Chair: Sergey Shebalov, Principal Research Analyst, Sabre Holdings, 

Dallas/Fort Worth, TX, United States of America, 

Sergey.Shebalov@sabre-holdings.com 

1 - Baggage Capacity and Demand Management Becomes Even 

More Sophisticated 

Desmond Di Wang, Northwestern University, 2145 Sheridan Road, 

Evanston, IL, 60208, United States of America, 

diwang2007@u.northwestern.edu, Diego Klabjan 

While sophisticated forecasting management systems exist for the passengers and 

cargo load of an aircraft, this is not the case for baggage despite significant costs 

associate with baggage mishandling. It is thus desirable to detect excessive baggage 

load early during the booking process to avoid dire consequences for both 

passengers and the airline. We present methodologies behind a system for baggage 

load forecasting any number of days before the day of operations. 

2 - Short-term Maintenance Allocation 

Sergey Shebalov, Principal Research Analyst, Sabre Holdings, 

Dallas/Fort Worth, TX, United States of America, 

Sergey.Shebalov@sabre-holdings.com 

Short-term maintenance allocation involves building and supporting maintenance 

schedule for each tail to ensure its operationability. We present a model that creates 

an optimal set of maintenance blocks for a given tail assignment and allocates 

maintenance events to them. We optimize aircraft utilization and maintenance cost 

while maintenance capacity is constrained. We extend our formulation to allow 

overnight aircraft swaps and thus integrate maintenance allocation and tail 

assignment. 

3 - Planeside Manpower Planning at United Airlines 

Feryal Kuran, United Airlines, 1200 E. Algonquin Rd., Elk Grove 

Village, IL, United States of America, Feryal.Kuran@united.com, 

Kumar Satyam 

Planeside Tool is a major component of the Ramp Operations automation strategy at 

United Airlines. Initial phase enables manual assignment of resources to cover 

arrival and departure packages. Model, developed by Enterprise Optimization, 

suggests best assignment of resources to packages and lunches over multiple hours 

considering various factors. Weighting of factors can be adjusted depending on 

needs of operations. Model enables improved coverage of flights and better 

utilization of resources. 

4 - Structured Deplaning: A Simulation and Optimization of 

Implementable Strategies 

Andrew Wald, Northwestern University, 2145 Sheridan Road, 


andrewwald@gmx.com, Diego Klabjan 

Deplaning naturally occurs row by row down the length of an aircraft. Using 

simulation and optimization, we design deplaning strategies (e.g., deplane by group) 

that significantly reduce the overall deplaning time. These evaluations are 

established through field observations and simulation, and have been tested across 

several equipment types. 

■ WA71 

H - Salon H, 6th Floor 

Resource Efficiency and Distribution Models in DEA 

Cluster: In Honor of Bill Cooper 


Chair: Subhash Ray, University of Connecticut, CT, United States of 

America, subhash.ray@uconn.edu 

1 - Using DEA to Help the Regulator Set Promulgated Insurance Rates 

Patrick Brockett, University of Texas, TX, United States of America, 

brockett@mail.utexas.edu, William Cooper, Jing Ai, Charles Yang, 

Linda Golden, Utai Pitaktong 

We offer a methodology for setting efficient promulgated insurance rates illustrated 

through an application to title insurance rate setting. In title insurance, losses 

constitute a tiny percentage of the premium, and expenses dominate the rates. 

Rates are set (in Texas) by the regulator relying upon average expenses and loss 

ratios. This inflates rates and rewards inefficiency, since inefficient agents with 

higher expenses drive up the average expense resulting in higher promulgated rates. 

We use DEA to: identify those title agents that efficiently utilize their resources to 

produce policies, and show how rates can be promulgated based upon the expenses 

of only these efficient agents, thus encouraging efficient management and 

reasonable rates.

WA72 

2 - A DEA Approach to Simulation Output Analysis of an Agent-based 

Model of a Distribution Network 

William Sawaya, Assistant Professor, Texas A&M University, Eng. 

Tech. and Industrial Distribution, College Station, TX, 77845, 

United States of America, sawaya@tamu.edu, Andrew Johnson, Sri 

Nagendra Jayanty 

Simulation output analysis for comparing different systems configurations can be 

complicated by the fact that there are often multiple performance variables of 

interest. This complexity can be further compounded in agent-based system because 

it is conceivable that there are multiple agents that each have their own 

performance information. This research presents a DEA approach analyzing 

simulation output of different system configurations for multiple organizations in a 

distribution network. 

3 - Cost Efficiency in a Model of Production and Distribution 

Subhash Ray, University of Connecticut, CT, United States of 

America, subhash.ray@uconn.edu 

This paper develops a measure of overall cost efficiency in an integrated model of 

production and distribution. The DEA model introduced by Ray et al (2008) for 

multi-location cost minimization in the presence of input price variation across 

locations is combined with the standard transportation model. The principal 

innovation is that optimal quantities produced at different locations and the 

quantities shipped to different destinations are determined simultaneously in a 

unified DEA model that minimizes total production and distribution cost. 

■ WA72 

H - Salon J, 6th Floor 

Joint Session TSL/ SPPSN: Aiding Disaster Relief 

Through Optimization 

Sponsor: Transportation Science and Logistics Society/ Public 

Programs, Service and Needs 


Chair: Irina Dolinskaya, Assistant Professor, Northwestern University, 

2145 Sheridan Road, M235, Evanston, IL, 60208, United States of 

America, dolira@northwestern.edu 

1 - Continuous Approximations for Relief Routing 

Michael Huang, Northwestern University, 2145 Sheridan Rd, 

Evanston, IL, 60208, United States of America, Michael- 

Huang@u.northwestern.edu, Karen Smilowitz 

In relief routing, solutions must be found quickly and should be easy to implement. 

We describe a simple policy and develop analytic approximations to measure its 

effectiveness. We demonstrate the accuracy of the approximations’ by comparing 

the predicted values against simulations of the policy. Finally, we test the solutions 

from the policy against more sophisticated polices generated with a Tabu search. 

2 - A Review on Recent OR Research in Disaster 

Operations Management 

Rajan Batta, Professor and Associate Dean, University at Buffalo 

(SUNY), Department of Industrial & Systems Engg, 438 Bell Hall, 

Buffalo, NY, 14260, United States of America, batta@buffalo.edu, 

Gina Galindo 

Disasters have attracted the attention of OR researches who are interested on 

applying scientific techniques to improve DOM effectiveness, and reduce the 

consequences of disasters on the economy and human lives. In this talk a review on 

recent OR research in DOM is offered and some future research directions are 

proposed. 

3 - Modeling Disaster Relief Networks 

Luis de la Torre, PhD Candidate, Northwestern University, 2145 

Sheridan Road, IEMS, Tech C210, Evanston, IL, 60201, United States 

of America, ledelatorre@u.northwestern.edu, Irina Dolinskaya, 

Karen Smilowitz 

This talk presents our research in last mile operations of disaster relief distribution. 

In particular, we focus on how distribution problems are represented in operations 

research models and characteristics of distribution in practice. We discuss 

implications of problem assumptions in both modeling and implementation. 


384 

■ WA73 

H - Salon K, 6th Floor 

Supply Chain: Reverse Logistics, Cost of Quality 

Sponsor: Transportation Science and Logistics Society 


Chair: Theresa Barker, PhD, University of Washington, Box 352650, 

Seattle, WA, 98115, United States of America, 

barkertj@u.washington.edu 

1 - Cost Sharing for Economic Lot-Sizing Problems with 

Remanufacturing Options 

Mohan Gopaladesikan, School of Industrial Engineering, Purdue 

University, 315 N. Grant Street, Grissom Hall 308, West Lafayette, 

IN, 47907, United States of America, mohang@purdue.edu, 

Nelson Uhan 

We consider a class of cooperative games that model the cost sharing issues that 

arise from the economic lot-sizing problem with remanufacturing options. By 

investigating the properties of various mathematical programming formulations and 

relaxations for the underlying lot-sizing problem, we obtain some insights into the 

existence of cost allocations in the core and the approximate core of these games, as 

well as the algorithmic aspects of computing such cost allocations. 

2 - Robust Design of Computer Remanufacturing and 

Recycling Facilities 

Suzanne Marcotte, Professor, Universite du Quebec a Montreal, 

315 rue Ste-Catherine Est, Montreal, QC, H2X 3X2, Canada, 

Suzanne.Marcotte@cirrelt.ca, Benoit Montreuil 

This presentation will describe the capacity planning of resources required in a 

computer remanufacturing and recycling facility. It will describe the generic process 

and the operations and decisions to be taken. Each operation of this process is 

characterized by sources of uncertainty and variability. It will then present results 

on the capacity required given various level of uncertainty. 

3 - Integrating Cost of Quality in Supply Chain Modeling: 

A Preliminary Study 

Krystel K. Castillo-Villar, Texas Tech University-Tecnologico de 

Monterrey, Box 43061, Lubbock, TX, 79409-3061, 

United States of America, krystel.castillo@ttu.edu, Neale R. Smith, 

James L. Simonton 

This work presents a preliminary methodology to compute the cost incurred by 

various actors within the supply chain due to the cost of poor quality or cost of 

quality (CoQ). We consider a generic consumer goods supply chain, consisting of 

three tiers, namely suppliers, manufacturers, and retailers. The purpose of this work 

is to provide a guide to translate defect rates at supplier, manufacturing plant, and 

retailer to quality costs. The study recommends ways quality engineers can use the 

methodology to make decisions regarding investment in prevention activities, 

rework process, inspection, among others. The practical implications of this research 

are a better selection of suppliers and manufacturing plants based on quality and 

cost considerations and a better understanding of CoQ not just as an internal, but 

also as an external performance measure. The study offers insights based on the 

findings and provides guidelines for future research. 

■ WA74 


Optimal Sensor Location and Deployment 



Chair: Yi-Chang Chiu, University of Arizona, Tucson, AZ, 

United States of America, chiu@email.arizona.edu 

1 - Permanent Traffic Counter Location Problem on 

Transportation Network 

Fatemeh Sayyady, Research Assistant, North Carolina State 

University, 2501 Stinson Drive, 208 Mann Hall, Raleigh, NC, 27695, 

United States of America, fsayyad@ncsu.edu, George List, 

Yahya Fathi, John Stone 

We consider the problem of determining an optimal placement for the traffic 

counters on a large-scale highway network, subject to a budget constraint. We 

formulate the problem as a mixed integer linear programming problem, and show 

that it has structural similarities with the p-median problem as well as the knapsack 

problem. A reasonably fast Lagrangian-heuristic approach is presented to solve large 

size instances of the problem where CPLEX fails to report optimal values.

2 - Traffic Sensor Deployment Under Probabilistic Disruptions and 

Generalized Surveillance Effectiveness Measures 

Xiaopeng Li, University of Illinois at Urbana-Champaign, B156 

Newmark Civil Engineering Laborator, 205 N. Mathews Ave, 

Urbana, IL, 61821, United States of America, li28@illinois.edu, 

Yanfeng Ouyang 

We propose a reliable sensor location model to optimize surveillance effectiveness 

when sensors are subject to site-dependent probabilistic failures, and a general 

effectiveness measure is proposed to encompass most existing measures needed for 

engineering practice. We formulate a compact mixed-integer program and develop a 

variety of solution algorithms. We also propose alternative formulations in the form 

of reliable fixed-charge sensor location models. 

3 - A Continuum Approximation Approach to Reliable Traffic Sensor 

Deployment on Highway Corridors 

Xiaopeng Li, University of Illinois at Urbana-Champaign, B156 

Newmark Civil Engineering Laborator, 205 N. Mathews Ave, 

Urbana, IL, 61821, United States of America, li28@illinois.edu, 

Yanfeng Ouyang 

We propose a continuum approximation framework for the reliable deployment of 

traffic sensors to optimize surveillance effectiveness when sensors are subject to sitedependent 

probabilistic failures. 

4 - Optimal Advance Detector Location for Green Termination Systems 

on High Speed Isolated Intersections 

Lili Du, NEXTRANS Center, Purdue University, 2700 Kent Avenue, 

West Lafayette, IN, 47906, United States of America, 

ldu@purdue.edu, Anuj Sharma, Srinivas Peeta 

This model finds near-optimum solutions very efficiently, and results from the 

approximation method are shown to be very close to those from the discrete 

methods. 

Wednesday, 11:00am - 12:30pm 

■ WB01 


Planning for Uncertainty in Energy Systems 



Chair: Sarah Ryan, Professor, Iowa State University, 3004 Black 

Engineering Bldg., Ames, IA, 50011-2164, United States of America, 

smryan@iastate.edu 

1 - Electric Power System Generation Expansion Planning Problems 

Considering Risk 

David Coit, Associate Professor, Rutgers University, Industrial & 

Systems Engineering, 96 Frelinghuysen Rd., Piscataway, NJ, 08854, 

United States of America, coit@rutgers.edu, Hatice Tekiner, 

Frank Felder 

GEP problems are solved to determine generation options to add and where/when 

to be constructed considering risk. In studies, decision makers are risk neutral; but 

often they are risk averse. We solve a multiobjective optimization problem to 

minimize cost and risk. We define a subset of scenarios to represent the stochastic 

nature. Multiobjective optimization problem is solved to find a Pareto Front. 

2 - Generation Expansion Portfolio Optimization with Stochastic 

Production Tax Credit for Wind Power 

Jo Min, Iowa State University, IMSE Department, 3004 Black, 

Ames, IA, 50011, United States of America, jomin@iastate.edu, 

Jin Lee, Chenlu Lou, Chung-Hsiao Wang 

We construct and analyze a generation expansion portfolio model consisting of 

conventional power plants and windmills. Specifically, a mean - variance utility 

function is optimized under the assumption that the fuel price, electricity price, and 

production tax credit for wind power are random variables. Via parametric quadratic 

programming, we analytically derive various managerial insights with respect to the 

degrees of risk aversion, renewable portfolio standards, and production capacities. 

3 - Computational Issues in Solving Large-Scale Stochastic Grid 

Expansion Problems 

Jean-Paul Watson, Sandia National Laboratories, 

jwatson@sandia.gov, David Woodruff 

Grid generation and transmission expansion problems are frequently expressed as 

stochastic mixed-integer programs, commonly multi-stage due to long-term 

planning horizons. However, these models are very difficult to solve, due to multiple 

exogenous uncertainty sources and scenario tree depth. We investigate the issue of 

decomposition solver performance on such problems, focusing on Progressive 

Hedging. Computational results and solution strategies are reported for several test 

problems. 

INFORMS Austin – 2010 WB02 

385 

4 - Scenario Reduction Methods for Rolling Stochastic Energy 

Planning Programs 

Yan Wang, PhD Student, Iowa State University, 3004 Black 

Engineering, Ames, IA, 50010, United States of America, 

yanwang@iastate.edu, Sarah Ryan 

A medium-term fuel procurement and electricity generation planning problem is 

naturally solved as a rolling horizon series of stochastic programs with evolving fuel 

price forecasts. We examine the accuracy and computational efficiency of a scenario 

reduction heuristic that emphasizes the initial decision rather than the distribution 

of scenarios. 

■ WB02 


Risk Management & Stochastic Programming in 

Gas and Power Systems 

Cluster: Energy: Modeling the Interface Between Markets 

and Operations 


Chair: Qipeng Phil Zheng, Assistant Professor, West Virginia University, 

Industrial & Management Systems Eng, P.O. Box 6070, Morgantown, WV, 

26505, United States of America, Qipeng.Zheng@mail.wvu.edu 

1 - Forecasting PHEV Sales and Recharging Activities 

Lizhi Wang, Iowa State University, 3016 Black Engineering, Ames, 

IA, 50014, United States of America, lzwang@iastate.edu, 

Zhaoyang Duan, Brittni Gutierr 

Sales of plug-in hybrid electric vehicles (PHEVs) and PHEV users recharging 

behavior are two critical factors for studying the potential impact of PHEVs on 

electric power systems. We present novel approaches to making more accurate and 

revealing forecasts of these factors. We will forecast the sales of PHEVs as a function 

of several interacting sub-factors, and forecast PHEV users’ recharging behavior as a 

function of available recharging infrastructures. 

2 - Identification and Prevention for Blackout on Large-scale 

Power Grid 

Hongsheng Xu, University of Florida, 303 Weil Hall, P.O. Box 

116595, Gainesville, FL, 32611, United States of America, 

xuhongsh@ufl.edu 

A blackout is the situation where there is a total loss of power to a relatively wide 

area, and how to quickly identify it in the power grid is key to preventing it from 

cascading. In this paper, we are focusing on detecting the possible blackout using 

our proposed evaluation criteria and algorithms. The results above could lead to 

some strategy guidance for designing and operation power grid. A computational 

study is presented in which we apply our model to the simulated data sets. 

3 - Midterm Coordination of Natural Gas Storage and 

Power Generation 

Cong Liu, Postdoctor, Argonne national Laboratory, 9700 S. Cass 

Ave., Bldg 221, Argonne, IL, 60439, United States of America, 

liuc@anl.gov, Jianhui Wang, Mohammad Shahidehpour, Zuyi Li 

In this talk, we will investigate the midterm coordination of natural gas resources 

and power generation. The objective is to minimize the integrated social cost 

including cost of power and natural gas systems while satisfying their complex 

coupled network constraints. The original problem will be decomposed into subproblems 

for each week. Lagrangian relaxation will be used to relax weekly 

coupling constrains of gas storage reservoirs. 

4 - Expansion Planning Models for Combined Electricity and 

Natural Gas Systems 

Alexey Sorokin, University of Florida, 303 Weil Hall, P.O. Box 

116595, Gainesville, FL, United States of America, sorokin@ufl.edu, 

Vladimir Boginski, Qipeng Phil Zheng 

Natural gas is playing an increasingly important role in global energy market 

because of its environment friendly properties, especially for electricity generation. 

We consider transmission expansion problem for gas and electricity networks, as 

well as for LNG terminal location planning. The multiple stage expansion model is 

proposed.

WB03 

■ WB03 


Environmental Legislation, Carbon Trading and Supply 

Chain Management 

Sponsor: Energy, Natural Resources and the Environment//Forestry 


Chair: Eda Kemahlioglu-Ziya, University of North Carolina-Chapel Hill, 

CB#3490, Chapel Hill, NC, United States of America, 

Eda_KemahliogluZiya@unc.edu 

1 - Strategic Carbon Footprint Labeling in a Supply Chain 

Rob Zuidwijk, Rotterdam School of Management, P.O. Box 1738, 

Rotterdam, Netherlands, RZuidwijk@rsm.nl, Charles Corbett, 

Chien-Ming Chen 

When firms plan to put carbon footprint labels on their products, it is often not 

unambiguous how those carbon footprints should be determined. Current standards 

for carbon footprint reporting also leave room for ambiguity. This gives firms some 

flexibility in how to allocate carbon emissions to different products. In this paper, 

we examine conditions under which that flexibility in fact helps to reduce the firm’s 

total carbon footprint without compromising profits. 

2 - How Does Product Recovery Affect Quality Choice? 

Gilvan Souza, Associate Professor, Indiana University, Kelley School 

of Business, 1309 E 10th St, Bloomington, IN, 47401, 

United States of America, gsouza@indiana.edu, Atalay Atasu 

We study the impact of product recovery (remanufacturing or recycling) on product 

quality, where quality increases market valuation for the product. We find that the 

recovery cost structure and the presence of take-back legislation significantly impact 

quality. Product recovery can be welfare improving, underscoring benefits of 

environmental legislation. 

3 - The Effect of Remanufacturing on New Products 

Vishal Agrawal, Georgia Institute of Technology, 800 W Peachtree St 

NW, Atlanta, GA, United States of America, 

Vishal.Agrawal@mgt.gatech.edu, Atalay Atasu, Koert Van Ittersum 

In this paper, we experimentally investigate the effect of remanufactured products 

on the perceived value of new products. We incorporate this effect to analytically 

investigate an OEM’s strategy in the presence of competition from third-party 

remanufacturers. Our research shows that an OEM may not always benefit from 

preempting third-party remanufacturers. Instead, an OEM may find it more 

profitable to allow third-party competitors to remanufacture its products. 

4 - Complying with Take-Back Legislation: A Cost Comparison and 

Benefit Analysis of Compliance Schemes 

Gokce Esenduran, The Ohio State University, 2100 Neil Avenue, 

Columbus, OH, United States of America, 

esenduran_1@fisher.osu.edu, Eda Kemahlioglu-Ziya 

We compare three compliance schemes (i.e. individual, collective and collective with 

individual financial responsibility) that firms follow to comply with take-back 

legislations. We model each scheme as a two-stage Nash game and find the key 

market/operating conditions that make one preferable to the others. As the most 

cost effective scheme may fall short on environmental benefits, i.e. collection 

rate\treatability level, we identify how environmental benefits compare between the 

three schemes. 

■ WB04 


Decision Analysis III 


Chair: Rebeca Díaz, Full Time Teacher, Technological Institute of Superior 

Studies of Coacalco, Av. 16 de septiempre No. 54, Cabecera, Municipal, 

Coacalco, Edo, de México, México D.F., 55700, Mexico, 

rbkdiazt@hotmail.com 

1 - Integrated Multi-Time-Scale and Multi-Organizational-Scale 

Decision Making Model 

Christian Wernz, Assistant Professor, Virginia Tech, Industrial and 

Systems Engineering, 250 Durham Hall (0118), Blacksburg, VA, 

24061, United States of America, cwernz@vt.edu, Abhijit Deshmukh 

In organizations, hierarchically interacting agents make decisions at different time 

scales. Typically, higher level agents make decisions about strategic variables with 

lower frequency compared to lower level agents, which make decisions about 

operational variables more often. We develop a multiscale decision model for 

hierarchical agents, and present an analysis of three-level agent interactions. 


386 

2 - Selection of Talents Based on Decision Making Theory 

Shaikh Akhlaque-E-Rasul, Concordia University, 1515 St. Catherine 

West, Montreal, Canada, akhlaque1045@hotmail.com, 

Sudhir P. Mudur 

It is always a difficult task to select the best candidate from many. Sometimes, 

biased opinions may skew the selection process. To overcome any bias and to rank 

the candidates, we show how decision makers can apply the decision-making 

theory of material science, which was developed by D. H. Jee. In the present work, 

several case studies from our daily life that benefit from the use of this theory are 

presented. 

3 - Inclusion of Preference-Dependence in Multi-attribute 

Utility Theory (MAUT) 

Johannes Siebert, Akademischer Rat, University of Bayreuth, 

Lehrstuhl Prof. Schluechtermann, Universitaetsstr. 30, Bayreuth, 

95440, Germany, Johannes.Siebert@uni-bayreuth.de 

Decision makers express preferences -here considered as factors- as ratios to 

averages of the alternative. A Taylor expansion of the product of these factors yields 

first and higher order terms. The former depend on each one preference only 

whereas the later depend on two or more preferences and are used to model 

dependencies. The new model is termed aggregate utility factor model (AUFM). In a 

computer-based experiment practicability and consistence in decision making are 

confirmed. 

4 - CUT: A New Multi-criteria Approach for Non-additive 

Concavifiable Preferences 

Nikolaos Argyris, London School of Economics and Political Science, 

Houghton Street, London, N1 0HP, United Kingdom, 

n.argyris@lse.ac.uk, Jose Figueira, Alec Morton 

We propose a new multi-criteria approach for concavifiable preferences: Concave 

UTility (CUT). CUT defines a space of value functions consistent with a DM’s 

expressed preferences. CUT has analogies with existing aggregation-dissagregation 

approaches, e.g. the UTA procedure, however CUT is more general as it does not 

require that preferences are additive. We describe how CUT can be used in an 

interactive setting: pre-ordering a finite set of discrete alternatives and multi-criteria 

optimization. 

5 - Studying the Effect of Improved Individual Skills in Negotiation and 

Collective Decision Making 

Rebeca Díaz, Full Time Teacher, Technological Institute of Superior 

Studies of Coacalco, Av. 16 de septiempre No. 54, Cabecera, 

Municipal, Coacalco, Edo, de México, México D.F., 55700, Mexico, 

rbkdiazt@hotmail.com, Leopoldo Viveros, Mario Chew 

Some questions about group decision making relate to the skills of the members, for 

instance: Does the group make better decisions than those of its most skilled 

member? If the members do not share objectives, how the skills at decision making 

and negotiation influence the decision? To study these issues we developed an 

environment in which the skills of each member can be controlled, using a 

simulated job-sequencing problem as a benchmark and final year engineering 

students as subjects. 

■ WB05 


Stochastic Optimization 


Chair: Asad Ata, Southern Methodist University, P.O. Box: 750123, 

Dallas, TX, 75275-0122, United States of America, ata.asad@gmail.com 

1 - Equity Valuation and Debt Selection via Stochastic Programming 

Davi Valladao, PhD Student, Pontifical Catholic University of Rio de 

Janeiro, Rua Marqu’s de São Vicente, 225, Rio de Janeiro, RJ, 

22451-041, Brazil, davimichel@gmail.com, Geraldo Veiga, 

Alvaro Veiga 

We develop a multistage stochastic programming model for equity valuation and 

debt selection for a firm with a predetermined project portfolio. We consider fixed 

and floating interest rate debt with different maturities. Moreover, the price of each 

corporate bond is given by a risk free valuation multiplied by a discount factor. We 

assume this factor to be a concave piecewise linear function of the new debt issued, 

with each segment based on different leverage rate levels. 

2 - On a Class of Stochastic Programs with Endogenous Uncertainty: 

Algorithm and Applications 

Bruno Flach, PSR, Praia de Botafogo 228/1701, Rio de Janeiro, 

22260020, Brazil, bruno@psr-inc.com 

We study a class of stochastic programming problems in which the probability 

distribution of the random parameters is decision-dependent. We propose a 

convexification technique coupled with a cut-generation algorithm for the MINLP 

formulation and the incorporation of importance sampling concepts into the 

stochastic programming framework so as to allow the solution of large instances. 

The applicability of our methodology is illustrated by an example in the area of 

power systems’ reliability.

3 - A Stochastic Programming Model for Seasonal View Selection 

Problem in Database Management Systems 

Rong Huang, Research Assistant, North Carolina State University, 

484 Daniels Hall, Campus Box 7913, Raleigh, NC, 27695-7913, 

United States of America, rhuang@ncsu.edu, Rada Chirkova, 

Yahya Fathi 

We introduce the stochastic seasonal view selection problem with random query 

sets, and model it as a two-stage stochastic integer programming problem. We 

propose exact and inexact methods for solving the problem and present the results 

of a computational experiment. 

4 - A New Hybrid Method for Solving Multistage Stochastic 

Programming Problems 

Nezir Aydin, Research Assistant, Wayne State University, 4815 

Fourth Street, Room: 1067, Detroit, MI, 48202, United States of 

America, aydin@wayne.edu, Alper Murat, Leslie Monplaisir 

In this study, we propose a creative way of combining the Progressive Hedging 

Algorithm (PHA) and Sampling Average Approximation (SAA) methods for solving 

multi-stage Stochastic Programming (SP) problems such that the exactness and 

speed of attaining a solution can be traded-off. Through extensive experimental 

results, we demonstrate the effectiveness of this hybrid approach over the pure 

strategies (SAA or PHA only) under specific circumstances using multi-product lotsizing 

problem. 

5 - An LP Based State Space Approach to Stochastic Dynamic 

Programming Problem 

Asad Ata, Southern Methodist University, P.O. Box: 750123, Dallas, 

TX, 75275-0122, United States of America, ata.asad@gmail.com, 

Eli Olinick, Chester Chambers, Eli Snir 

A linear programming (LP) approach is discussed to solve a stochastic dynamic 

programming (SDP) problem. An attempt is made to alleviate the curse of 

dimensionality by formulating the problem as a discrete time, infinite horizon, 

stochastic dynamic programming model with a finite state space. The model results 

as an instance of Unichain Markov Decision Process. The LP produces a profit 

maximizing policy and states the likelihood the resulting markov system is in any 

particular state. 

■ WB06 


Tutorial: Robust Vehicle Routing 



Chair: Fernando Ordonez, Associate Professor, University of Southern 

California, Department of Industrial and Systems Eng, 3715 McClintock 

Ave, Los Angeles, CA, 90007, United States of America, fordon@usc.edu 

1 - Robust Vehicle Routing 

Fernando Ordonez, Associate Professor, University of Southern 

California, Department of Industrial and Systems Eng, 3715 

McClintock Ave, Los Angeles, CA, 90007, United States of America, 

fordon@usc.edu 

Vehicle routing problems in many industrial applications must take into account 

uncertain demand, traffic conditions and/or service times. In this tutorial we present 

recent work on the use of robust optimization for vehicle routing problems (VRP) 

under uncertainty. We outline different robust VRP models, depending on the 

formulation, source of the uncertainty, and correlation in the uncertainty. 

Furthermore, we discuss previous computational results that illustrate when such a 

robust model is convenient and when it is not. We show with results in two 

different applications that robust optimization is useful to find a routing plan when 

routes will be adapted to the outcome of the uncertainty. 

■ WB07 


Supply Chain, Closed-loop I 


Chair: Sameer Kumar, Professor of Operations and Supply Chain 

Management, Opus College of Business, University of St. Thomas, 

Mail # TMH 343, 1000 LaSalle Avenue, Minneapolis, MN, 55403-2005, 

United States of America, skumar@stthomas.edu 

1 - Analysis of Bullwhip Effect Under Retailer and Customer 

Forecasting with Price-sensitive Demand Function 

Yungao Ma, School of Management, Xi’an Jiaotong University, 

Shaanxi, China, 710049, Xi’an, 710049, China, 

ma.gao@stu.xjtu.edu.cn, Zhiping Yuan, Yufei Huang 

In a single two-stage supply chain with one supplier and one retailer, we analyze 

the impact of demand forecast by retailer and price forecast by customers on 


387 

Bullwhip Effect (BWE) under moving average (MA) and exponential smoothing 

(ES) forecasting methods. Results show that: first, demand forecast can reduce BWE 

under price-sensitive demand function; second, price forecast can reduce BWE 

under certain conditions when customers consider price fluctuation; besides, BWE 

under ES is not always significant compared to that under MA. 

2 - Product Recoveries in China: Remanufacturing vs. Refurbishing 

Yao Chen, Shanghai Jiaotong University, No.535 Fahua Road, 

Shanghai, Shanghai, China, yaochen514@msn.com, Fangruo Chen 

The markets for recovered products in China consist of both remanufactured and 

refurbished products, where the former must match the quality limit. Relative to 

new products, the two type of recovered products enjoy a cost advantage, but suffer 

from a lower willingness-to-pay by the consumers. We characterize the equilibrium 

market structure when all the three products compete with each other with a 

emphasis on the conditions under which the remanufacturing products can survive 

the competition. 

3 - Variable-level Disassembly Planning for Facilitating 

Remanufacturing Between Different Products 

Yoo S. Hong, Associate Professor, Seoul National University, 599 

Kwanakro, Kwanakgu, Seoul, Korea, Republic of, yhong@snu.ac.kr, 

Changmuk Kang 

In order to resolve the mismatch between return supply of end-of-life (EOL) 

products and demand of a new product, an EOL product has to be disassembled and 

remanufactured at a part level, and used for producing a different product. This 

study solves a problem of planning disassembly of different kinds of products of 

which return and demand change over time. Each product is disassembled into 

variable levels according to its return and demand of serviceable parts. 

4 - Closed Loop Supply Chains for U.S., Japan and EU Auto Industries 

- A System Dynamics Study 

Sameer Kumar, Professor of Operations and Supply Chain 

Management, Opus College of Business, University of St. Thomas, 

Mail # TMH 343, 1000 LaSalle Avenue, Minneapolis, MN, 55403- 

2005, United States of America, skumar@stthomas.edu 

System Dynamics analysis of the U.S., Japan and EU auto industries’ reverse value 

chains was conducted to explore the impact of government regulations, financial 

incentives and market pricing for remanufactured and recycle materials on cash 

flows and use of such raw materials for car manufacturers in these three market 

segments. 

■ WB08 


Location Modeling Applications 

Sponsor: Location Analysis 


Chair: Rajan Batta, Professor and Associate Dean, University at Buffalo 

(SUNY), Department of Industrial & Systems Engg, 438 Bell Hall, Buffalo, 

NY, 14260, United States of America, batta@buffalo.edu 

1 - Ambulance Location, Relocation and Relocation with 

Uncertainty Models 

Bo Zeng, Assistant Professor, University of South Florida, 

Department of IMSE, Tampa, FL, 33620, United States of America, 

bzeng@usf.edu, Shengyong Wang 

Various location models for ambulances have been extensively studied. In this talk, 

we first present a dynamic relocation model that could be combined with any 

existing location models. Then, a stochastic relocation model that includes request 

uncertainties will be presented. Finally, numerical study will be given to show the 

effectiveness of those models on response times in difference situations. 

2 - Centralized Dispatch Approximation for the Transshipment Problem 

Dmitry Krass, Professor, University of Toronto, Rotman School of 

Management, 105 St. George Street, Toronto, ON, M5S 3E6, Canada, 

Krass@Rotman.Utoronto.Ca, Alex Shlakhter 

In the transshipment problem a number of retailers facing stochastic demand must 

place orders before the demand is known, but can transship inventory once the 

demand is realized. We show that by assuming a central depot through which all 

transshipments must flow, the computations of the (approximately) optimal 

ordering policy are greatly simplified. Moreover, the performance of the 

approximate policy is excellent. We also analyze and obtain analytical results for the 

monotone policy case.

WB09 

3 - Facility Location Problem with Network Oligopoly 

Paul Berglund, University at Buffalo, 786 West Ferry Street, Buffalo, 

NY, 14222, United States of America, berglund@buffalo.edu, 

Changhyun Kwon 

We formulate an equilibrium facility location problem on a discrete network, where 

the locating firm acts as the leader in a Stackelberg-Nash-Cournot competitive 

equilibrium problem. To maximize expected profits the locating firm must solve a 

problem with equilibrium constraints. Finding an optimal solution is hard for large 

problems. Therefore a heuristic solution procedure based on simulated annealing is 

presented. 

4 - Locating Temporary Depots to Facilitate a Post-Disaster Relief 

Operation: A Case Study 

Yen-Hung Lin, PhD Candidate, University at Buffalo (SUNY), 438 

Bell Hall, Buffalo, NY, 14260, United States of America, 

yl48@buffalo.edu, Peter Rogerson, Alan Blatt, Marie Flanigan, Rajan 

Batta 

A case study through HAZUS simulation software of a historical earthquake scenario 

is used to evaluate the performance of a disaster relief operation with a distributed 

supply strategy. The distributed supply strategy locates several temporary depots in 

the area impacted by a recent earthquake event so that demand can be fulfilled by 

either temporary depots or the central depot. The items being supplied are of three 

categories (water, food and medicine) and have different priorities. 

■ WB09 


Behavioral Issues 



Chair: Mirko Kremer, Pennsylvania State University, 460 Business 

Building, State College, PA, 16802, United States of America, 

Mirko.Kremer@psu.edu 

1 - Explanations of Newsvendor Biases Do Not Square 

Neil Bearden, Assistant Professor, INSEAD, 1 Ayer Rajah Ave, 

Singapore, 138636, Singapore, Neil.BEARDEN@insead.edu, 

Sameer Hasija 

We show that now-conventional explanations of biases in newsvendor experiments 

do not hold up to close scrutiny. The accounts implicitly assume that decision 

behaviour is invariant with respect to problem framing, but we show that order 

quantities are strongly contingent on the way the problem is presented. Finally, we 

argue for an eliminitavist stance on explanation in behavioural operations: the 

attempt to give simple accounts of decision behaviour in complex problems should 

be abandoned. 

2 - On the Ability to Identify Pareto-improving Supply Contracts 

Mirko Kremer, Pennsylvania State University, 460 Business Building, 

State College, PA, 16802, United States of America, 

Mirko.Kremer@psu.edu, Tony Haitao Cui 

Formal analyses of risk-sharing supply contracts typically focus on identifying 

contracts that induce maximal channel efficiency, while allowing for a flexible 

allocation of channel profits. As a first step toward a better understanding of the 

bargaining process toward such Pareto contracts, we investigate empirically how 

sellers and buyers choose from sets of contracts. We further explore how the ability 

to identify Pareto-improvements is sensitive to the framing of the contract. 

3 - Revenue Sharing versus Buyback Contracts: Influence of 

Supplier Preferences 

Karen Donohue, Associate Professor, University of Minnesota, 

Carlson School of Management, Minneapolis, MN, United States of 

America, donoh008@umn.edu, Yinghao Zhang, Tony Haitao Cui 

Prior analytical research shows that buyback and revenue sharing contracts achieve 

equivalent channel-coordinating solutions when applied in a single supplier-buyer 

setting. More recently, behavioral research suggests that the two contracts do not 

always perform equivalently. We examine how supplier preferences, such as loss 

aversion and time discounting, can lead suppliers to prefer one contract type over 

the other depending on the ratio of overage and underage costs. 

4 - Product Quality Choice and Inventory Risk Under Strategic 

Consumer Behavior 

Robert Swinney, Assistant Professor, Stanford University, 518 

Memorial Way, Stanford, CA, 94305-5015, United States of America, 

Swinney_Robert@GSB.Stanford.Edu, Sang-Hyun Kim 

We analyze a model in which a firm selling a single, seasonal product sets both 

product quality and quantity before selling to a population of forward-looking 

customers. The size of the market is uncertain to the firm, and customers anticipate 

the price path of the product and may strategically delay a purchase to pay a lower 

price. We consider the impact of both demand uncertainty and customer behavior 

on the optimal quality and quantity of the product. 


388 

■ WB10 


Managing the Distribution Channel with Competing 

Products and Supply Risk 



Chair: Yunzeng Wang, University of California-Riverside, A. Gary 

Anderson Graduate School of Mgmt, Riverside, United States of America 

1 - Split Award Auctions for Supplier Retention 

Aadhaar Chaturvedi, IESE Business School, Av. Pearson 21, 

Barcelona, Spain, achaturvedi@iese.edu, Damian Beil, 

Victor Martinez de Albeniz 

Traditionally, buyers use auctions select the lowest-cost supplier. By doing so, they 

might alienate the losing suppliers. This involves a future cost of supplier 

qualification if they defect from the supply base. This paper considers the trade-off 

between the purchasing cost and the qualifying cost paid to maintain the supply 

base. We find the optimal split award that minimizes long-run costs and the optimal 

supply base size that the buyer should maintain. 

2 - Nature of Coordination Contracts for Supply Chain Management: 

Classifications and Structural Results 

Meng Lu, The Chinese University of HK, Shatin, N.T., Hong Kong, 

China, mlu@se.cuhk.edu.hk, Houmin Yan, Suresh Sethi 

In this paper, based on the group decision-making and game theory, we define a 

framework for supply chain contracts and classify into groups. With precise 

mathematical definitions, and subsequently developed structural properties and 

sufficient conditions, we are not only able to measure the goodness of supply 

contracts but also to reveal the nature of the supply coordination. We develop 

indexes to measure coordination contracts in terms of coordination strength and 

decision sequence dependency. 

3 - Heterogeneous and Nonlinear Frontier Analysis: Supply-Chain 

Transaction Cost Perspective 

John Liu, Chair Professor, Hong Kong PolyU, CD 401b, Kowloon, 

Hong Kong - PRC, lgtjliu@polyu.edu.hk, Jason Jianfeng Mao 

Inspired by the works of Williamson (2008), transaction cost economics of supply 

chain management is emerging as a challenging research area, due to heterogeneity 

and non-linearity of transaction costs as incurred in outsourcing operations of SCM. 

We herein develop a degenerative frontier model which incorporates both 

heterogeneity and nonlinearity. We obtain convexity properties, and develop a 

solution method which converts a DF problem into a sequence of convex problems. 

■ WB11 


Empirical Research in Operations Management 



Chair: Vishal Gaur, Associate Professor, Johnson School, Cornell 

University, 321 Sage Hall, Ithaca, NY, 14850, United States of America, 

vg77@cornell.edu 

1 - The impact of New Product Introduction on Plant Productivity in the 

NA Automotive Industry 

Serguei Netessine, Professor, INSEAD, Boulevard de Constance, 

Fontainebleau, 77305, France, serguei.netessine@insead.edu, Manu 

Goyal, Anand Gopal, Matthew Reindorp 

We empirically estimate productivity losses during new product introductions in the 

automotive industry. We show that productivity losses can be mitigated through 

manufacturing flexibility and different forms of organizational learning. 

2 - An Empirical Study of Pricing in the U.S. Automobile Industry 

Antonio Moreno, The Wharton School, 3730 Walnut St, 

Philadelphia, United States of America, amore@wharton.upenn.edu, 

Gerard Cachon, Christian Terwiesch 

Despite the abundant theoretical literature on dynamic pricing, price postponement 

and operational flexibility, there is limited empirical evidence on how firms actually 

adjust their prices, and how operational practices play a role in their pricing 

decisions. Using a detailed transactional dataset of the US auto industry, we study 

the impact of operational strategic decisions on pricing.

3 - Improving Retail Store Performance by Incorporating Traffic 

Characteristics in Labor Planning 

Vidya Mani, The University of North Carolina at Chapel Hill, 

The Kenan-Flagler Business School, Chapel Hill, NC, 27599, 

United States of America, vidya_mani@unc.edu, Saravanan Kesavan, 

Jayashankar Swaminathan 

Managing store labor is critical to improving customer service as well as controlling 

costs for retailers. Using data from a large retailer, we show how labor planning can 

be improved by incorporating store traffic characteristics in managerial staffing 

decisions. 

4 - An Empirical Estimation of the Impact of Airline Flight Schedules on 

Flight Delays 

Vinayak Deshpande, Purdue University, 100 S. Grant St, 


vinayak@purdue.edu, Mazhar Arikan 

Airline flight delays have come under increased scrutiny lately, with FAA data 

revealing that airline on-time performance was at its worst level in 13 years in 

2007. Our goal is to examine the impact of the scheduled time block allocated for a 

flight on on-time arrival performance. We combine empirical flight data published 

by BTS, with the Newsvendor framework from the Operations literature to conduct 

this analysis. Our results show that airlines systematically under-schedule flights. 

■ WB12 


Emerging Topics in Supply Chain Management 


Supply Chain 


Chair: Tingting Cui, University of California-Berkeley, Berkeley, CA, 

United States of America, tingting@ieor.berkeley.edu 

1 - Modeling and Mitigating the Effects of Supply Chain Disruption 

on Wargames 

Shilan Jin, SUNY at Buffalo, 435 Bell, SUNY at Buffalo, 

North Campus, Amherst, NY, 14260, United States of America, 

sjin6@buffalo.edu, Zigeng Liu, Jun Zhuang 

We integrate supply chain risk management with a government-terrorist game in 

war zones. The equilibrium outcomes of wargames depend on the government’s 

resources delivered through military supply chains, which are subject to disruptions. 

We study the government’s optimal pre-disruption strategies, including inventory 

protection, capacity backup protection and the combination. 

2 - Asymmetries of the Pull-to-Center Effect in the 

Newsvendor Experiments 

Tianhu Deng, PH. D student, UC, Berkeley, IEOR Department, 4141 

Etcheverry Hall, Mail Code 1777, Berkeley, United States of America, 

tianhu_deng@berkeley.edu 

It has been frequently observed in newsvendor games, the subjects’ average order 

quantity lies in between the optimal order quantity and the mid-point of the 

demand, in both high profit and low profit settings. This phenomenon is called pullto-center 

effect. Some researchers found that the pull-to-center effect is stronger in 

the high-profit setting than the low profit setting while others found the opposite 

result. We present explanations for this discrepancy. 

3 - Impact of Social Contagion in Make-to-Stock and Make-to-Order 

Supply Chains 

Shan Li, Operations Research Scientist, Amazon.com, 605 5th 

Avenue South, Seattle, WA, 98104, United States of America, 

lisapine@berkeley.edu, Teck Ho, Z. Max Shen 

We first analyze the impact of social contagion in a make-to-stock supply chain. We 

show that an out-of-stock phenomenon that occurs earlier in a product’s life cycle 

always leads to a greater loss in a firm’s customer assets. We then analyze the 

impact of social contagion in a make-to-order spply chain. We demonstrate that a 

lengthy lead time can slow down social contagion, decelerate customer purchases, 

and thus significantly decrease a firm’s total customer assets. 

4 - Flexible Nonhomogeneous Supply Chain Design Under Supply 

Chain Disruptions 

Ye Xu, University of California-Berkeley, Berkeley, CA, 

United States of America, yex207@berkeley.edu, Z. Max Shen 

We look at the flexibility design problem of a general supply chain with unbalanced 

structure, and nonhomogeneous demands and suppliers. Besides demand 

uncertainty, supply disruption is also considered in our model. We discuss solution 

algorithms for a series of models, and show that the marginal value of flexibility 

does not diminish as capacity increases. Rather, more flexibility is encouraged when 

more capacity is available. 


389 

■ WB13 


Dynamically Managing Customer Loyalty and Learning 




Chair: Dan Adelman, Professor, University of Chicago, 5807 South 

Woodlawn Ave., Chicago, IL, 60637, United States of America, 

dan.adelman@chicagobooth.edu 

1 - Dynamic Capacity Allocation to Customers Who Remember 

Past Service 

Adam Mersereau, University of North Carolina, Kenan-Flagler 

Business School, Chapel Hill, NC, United States of America, 

ajm@unc.edu, Dan Adelman 

We study the problem of a supplier dynamically allocating limited capacity among a 

portfolio of customers, where each customer’s orders depend positively on the fill 

rates provided to her in the past. Customers differ from one another in their 

contribution margins, their demand volatilities, and the length of their memories. 

We develop an approximate dynamic programming algorithm that rationalizes the 

fill rates the supplier should target for each customer. 

2 - Optimal Hiring and Retention Policies for Heterogeneous Workers 

with Learning 

Alessandro Arlotto, University of Pennsylvania, 3730 Walnut Street, 

500 Jon M. Hunstman Hall, Philadelphia, PA, 19104, United States 

of America, alear@wharton.upenn.edu, Stephen E. Chick, 

Noah Gans 

We study the hiring and retention of heterogeneous workers that learn over time. 

We formulate the problem as an infinite-armed bandit and characterize the optimal 

hiring and retention policy in detail. We develop approximations that allow the 

efficient implementation of the optimal policy and the evaluation of its 

performance. We present numerical examples that show, among other things, that 

the active screening and monitoring of employees leads to substantial gains. 

3 - Signaling Service Quality in Queues 

Senthil Veeraraghavan, Assistant Professor, Wharton School of 

Business, University of Pennsylvania, 3730 Walnut Street, Jon M. 

Huntsman Hall, Suite 500, Philadelphia, PA, 19104, 

United States of America, Senthilv@wharton.upenn.edu 

We study how a high quality service firm signals quality to differentiate from a low 

quality firm through expensive signaling efforts. Rational consumers try to learn 

quality through these imperfectly advertised signals. Their decisions are based on 

signaling efforts of the firms, congestion in the market, and the service value. We 

show that it is likely that a high quality firms may incur lower revenues due to 

signaling efforts. 

■ WB14 


Supply Chain Management IX 


Chair: Nomesh Bolia, Dr, IIT Delhi, #276 Block III, IIT Delhi, New Delhi, 

DL, 110016, India, nomesh@mech.iitd.ac.in 

1 - Measurement and Optimization of Supply Chain Responsiveness 

Sin-Hoon Hum, National University of Singapore, NUS Business 

School, 15 Kent Ridge Drive, Singapore, 119245, Singapore, 

bizhumsh@nus.edu.sg, Mahmut Parlar 

We consider make-to-order supply chains with multiple stages where each stage is 

completed in a random length of time. We define the responsiveness of such a 

supply chain as the probability that an order placed now will be fulfilled within t 

time units. We optimize the responsiveness of the supply chain by maximizing the 

probability that the order will be fulfilled within some promised time interval 

subject to a budget constraint. 

2 - Behavior of a Remanufacturing System in Presence of Varying 

Suppliers Reliability for New Product 

Suman Niranjan, Assistant Professor, Savannah State University, 

College of Business Administration, 3219 College Street, Savannah, 

GA, 31404, United States of America, suman1130@gmail.com 

In this paper we study a two-echelon remanufacturing system, which focuses on 

what should be the right mix of new and remanufactured components used in the 

manufacturing of a new product, and why should we care about the mix?. 

Moreover we study this problem in the presence of unreliable suppliers for new 

component. We analyze the performance of the system initially by developing a set 

of dynamic equations used in simulation based optimization framework.

WB15 

3 - The Analysis of Performance in One Single and Dual Channels 

Zhaoqiong Qin, Associate Professor, North Carolina A&T State 

University, 1601 E. Market Street, Greensboro, NC, 27411, 

United States of America, zqin@ncat.edu 

Suppliers routinely decide whether to distribute their products through one single 

channel or dual channels. Conventional wisdom says that dual distribution channels 

outperform one single channel based on the whole supply chain’s performance in 

the capacity and the supplier’s profit. However, this paper finds that the channel 

structure including centralized and/or decentralized in the distribution plays an 

important role in these performances. 

4 - Fuzzy Logic Based Methods to Quantify Supply Chain Performance 

Nomesh Bolia, Dr, IIT Delhi, #276 Block III, IIT Delhi, New Delhi, 

DL, 110016, India, nomesh@mech.iitd.ac.in, Pranav Saxena, 

Jalaj Bhandari 

Globalization and dynamic market conditions have forced companies to focus on 

methods to evaluate the performance of their supply chains and improvise where 

needed. Hence there is an increasing interest in quantifying performance. However 

there are a lot of uncertainties in supply chain parameters (hence performance 

measures), and relative importance of these performance measures. We apply fuzzy 

logic to address these issues and develop an appropriate performance index for 

supply chains. 

■ WB15 


Convex Optimization 


Chair: David Papp, Rutgers Center for Operations Research, 640 

Bartholomew Rd, Piscataway, NJ, 08854, United States of America, 

dpapp@rutcor.rutgers.edu 

1 - Convex Relaxation for the Planted k-disjoint-clique Problem 

Brendan Ames, PhD Candidate, University of Waterloo, 200 

University Avenue West, Waterloo, ON, N2L3G1, Canada, 

bpames@math.uwaterloo.ca, Stephen Vavasis 

We consider the k-disjoint-clique problem. For a given graph G, the problem is to 

find within the graph k disjoint cliques that cover the maximum number of nodes 

of G. The k-disjoint-clique problem is NP-hard, but we show that a convex 

relaxation can solve it in polynomial time for certain input instances. The input 

instances for which our algorithm finds the optimal solution consist of k disjoint 

large cliques that are then obscured by noise edges and noise nodes. 

2 - A Convexity Result for an (S-1,S) Inventory Model Under 

Time Limits on Backorders 

Emre Tokgoz, Emre.Tokgoz-1@ou.edu, Hillel Kumin 

The (S-1,S) inventory model with time limits on backorders has previously been 

solved by minimizing a function of two variables, one of which is integer. We 

investigate the convexity of the objective function and develop new convexity 

results for functions with m integer and n continuous variables. 

3 - A Holloway-Inspired Enhancement for the Frank-Wolfe Approach to 

the Convex Hull Problem 

Xinyu Wang, PhD Student, Southern Methodist University, 

P.O. Box 750123, Dallas, TX, 75275-0122, United States of America, 

xwang@smu.edu, Richard Helgason 

The extreme points of a convex hull can be found by using the Frank-Wolfe 

quadratic programming method. We investigate a novel Holloway-Inspired 

enhancement to the Frank-Wolfe method. The enhancement tries to avoid zigzag in 

the later stage of the projection point computation and gives us faster convergence. 

Experimental results indicate the enhancement leads to a significant speedup. 

4 - A Polyhedral Projection Method for Solving Variational Inequalities 

Sudhanshu Singh, PhD Student, UNC Chapel hill, UNC Department 

of STOR, B 44, hanes hall, CB# 3260, UNC Chapel Hill, Chapel hill, 

NC, 27599-3260, United States of America, sssingh@email.unc.edu, 

Shu Lu 

Most projection based methods for solving variational inequalities suffer from the 

drawback that the projection on a general convex set is not easy to find. This talk 

presents a projection method which replaces the feasible set by a polyhedral convex 

set in each iteration. The proposed algorithm uses (sub)gradient information to 

generate the polyhedron. It is easy to implement and its convergence is analyzed 

under some mild assumptions. 


390 

5 - Generalized Sum-of-squares Cones and Their Applications 

David Papp, Rutgers Center for Operations Research, 640 

Bartholomew Rd, Piscataway, NJ, 08854, United States of America, 

dpapp@rutcor.rutgers.edu, Farid Alizadeh, Ricardo Collado 

We consider the cone of sum-of-squares vectors with respect to an arbitrary bilinear 

multiplication in a finite dimensional space. We show that these cones are feasible 

sets of semidefinite optimization problems, extending Nesterov’s results on real 

valued sum-of-squares functions. Different choices of spaces and multiplications 

give rise to a diverse set of applications; we show a few in combinatorial 

optimization, geometric optimization, and shape-constrained statistical estimation. 

■ WB16 


Remanufacturing 


Chair: Pei-Fang Tsai, Assistant Professor, National Taipei University of 

Technology, 1, Sec. 3, Chung-hsiao E. Rd., Taipei, 10608, Taiwan - ROC, 

ptsai@ntut.edu.tw 

1 - Inventory Management in Closed-loop Supply Chains Under 

Non-stationary Demand 

Ibrahim Dogan, Wayne State University, Industrial & Manufacturing 

Engineering, 4815 Third Street, Detroit, MI, 48202, United States of 

America, aq9742@wayne.edu, Ratna Babu Chinnam 

This study aims to analyze remanufacturer’s inventory control policy under nonstationary 

demand. The objective is to decide on virgin product replenishment 

quantities under used product returns. The exact solution to this inventory control 

problem in our setting is complex and time demanding. We offer and analyze a 

number of different sub-optimal policies. 

2 - Used Product Returns Policy Under Demand and 

Return Uncertainty 

Samar Mukhopadhyay, Professor, SungKyunKwan University-GSB, 

53 Myungryun dong 3-ga, Jongno gu, Seoul, 110745, Korea, 

Republic of, samar@skku.edu, Robert Setaputra 

Reusing still usable components from a used up product makes sound 

environmental and economic sense. Uncertainty in the quantity of returns and 

demand complicates the operation. Our decision variable is the return policy, 

characterized by the amount refunded to the consumer. The trade-off is between 

increased revenue due to reduced input cost and increased cost due to higher return 

amount. Optimal return policy and managerial insights on sensitivity analyses will 

be presented. 

3 - Remanufacturing Planning with Variable Quality Returns 

Xiaoning Jin, PhD Student, University of Michigan, 2300 Hayward 

Street, Ann Arbor, MI, 48109, United States of America, 

xnjin@umich.edu, S. Jack Hu, Jun Ni 

There is a need for remanufacturers to grade quality of the returns prior to recovery 

processes because of the quality uncertainty of returns.Since some returns require 

more capacity and cost to bring the unit up to a required quality standard than 

others,quality variability information will have great impacts on remanufacturing 

decisions.We’ll develop a quality-dependent remanufacturing model to obtain the 

optimal control of the remanufacturing quantity that minimize the expected total 

cost. 

4 - Best Partial Disassembly Strategy for Retrievable 

End-of-Life Products 

Pei-Fang Tsai, Assistant Professor, National Taipei University of 

Technology, 1, Sec. 3, Chung-hsiao E. Rd., Taipei, 10608, Taiwan - 

ROC, ptsai@ntut.edu.tw 

A product reaches its end of life when it is malfunctioned or undesirable to the 

users. This research focuses on returned products in two categories: those are 

qualified to be repaired or remanufactured, and those can be retrieved for useable 

parts. The best partial disassembly strategy is proposed and formulated as a 

multicommodity flow problem with applicability established. The objective is to 

obtain maximum potential benefits inherent in the end-of-life production planning.

■ WB17 


Decision Making in Interdependent Systems 


Chair: Kash Barker, Lecturer, University of Oklahoma, 202 W. Boyd, 

Room 124, Norman, OK, 73019, United States of America, 

kashbarker@ou.edu 

1 - Decision Analysis Tool for Assessing Hurricane Impact on Regional 

Workforce Productivity 

Joost Santos, Assistant Professor, Engineering Management and 

Systems Engineering, The George Washington University, 

1776 G Street NW, Rm 164, Washington, DC, 20052, 

United States of America, joost@gwu.edu 

This research develops a workforce recovery model based on input-output analysis 

to estimate sector inoperability and economic losses. Based on our simulated 

hurricane scenarios, service sectors in Virginia suffer the largest workforce 

productivity impact-accounting for nearly 40% of the total economic losses. 

Sensitivity analysis of inoperability and loss reduction objectives can provide insights 

on identification and prioritization of critical workforce sectors to expedite disaster 

recovery. 

2 - Resilience Assessment and Improvement of Urban 

Infrastructure Systems 

Leonardo Dueñas-Osorio, Assistant Professor, Rice University, 6100 

Main Street, MS-318, Houston, TX, 77005, United States of America, 

leonardo.duenas-osorio@rice.edu, Min Ouyang 

This paper proposes an annual resilience metric, which reflects the capacity of 

infrastructure systems to resist, absorb and recover from all possible disruptive 

events. Taking the transmission power grid and gas transmission system of in Harris 

County, Texas, as an example, the effectiveness of different resilience improvement 

measures are analyzed and discussed. This study can provide insight and direction to 

design and retrofit resilient interdependent infrastructure systems. 

3 - Interdependency Models to Compare Industry Preparedness and 

Reactive Strategies to Disruptive Events 

Cameron MacKenzie, Graduate Student, Industrial Engineering, 

University of Oklahoma, 202 W. Boyd, Room 124, Norman, OK, 

73019, United States of America, cmackenzie@ou.edu, Kash Barker 

We use a risk-based interdependency model to quantify the effect of preparedness 

strategies such as maintaining inventory and the effect of reactive strategies such as 

choosing alternate transportation routes or different suppliers if a disruptive event 

occurs. We examine the conditions that incentivize industries to prepare for a 

disruptive event and how those decisions impact their reactions to a supply chain 

disruption. We deploy the model with a case study using actual commodity data. 

4 - Dynamic Analysis of Interdependent Inoperability in Multi-modal 

Transportation Networks 

Raghav Pant, Graduate Student, Industrial Engineering, 


73019, United States of America, Raghav.Pant-1@ou.edu, 

Kash Barker 

We study time-dependent production losses to regions due to disruptions in 

important transportation facilities (e.g., ports), by integrating the Dynamic 

Inoperability Input-Output Model (DIIM) with a network queuing model. Network 

model-driven initial inoperability and recovery time estimates strengthen our ability 

to evaluate supply chain risk management options and enhance risk management 

decisionmaking. Examples using commodity flow data for inland port disruptions 

are illustrated. 

5 - Impact of Disasters on National Freight Flows 

Saniye Gizem Aydin, Graduate Student, Industrial Engineering, 


73019, United States of America, gizemaydin@ou.edu, 

P. Simin Pulat, Guoqiang Shen, Manjunath Kamath, Ricki Ingalls 

Transportation systems are vulnerable to disasters and absolutely vital for the 

economy. There are relatively few studies concerning the regional freight flow 

impact of the disaster, even fewer on national impact. Northridge Earthquake, 

Hurricane Katrina and I-40 Bridge collapse cases are studied. A spatial input-output 

model, distance and traffic volume based changes are used for measuring the 

importance, and the influence on decision making within the interdependent freight 

flow environment. 


391 

■ WB18 


OR in Practice II 

Sponsor: CPMS, The Practice Section 


Chair: Brian Lewis, Vice President, Professional Services, Vanguard 

Software, 1100 Crescent Green, Cary, NC, 27518, United States of 

America, brian.lewis@vanguardsw.com 

Co-Chair: Bjarni Kristjansson, President, Maximal Software, Inc., 933 N. 

Kenmore St., Suite 218, Arlington, VA, 22201, United States of America, 

bjarni@maximalsoftware.com 

1 - An Integrated Framework of Service Quality for Global Delivery of 

Contact Center Services 

Nanda Kambhatla, Senior Manager, Human Language Technologies, 

IBM India - Research, ‘D’ Block, Embassy Golf Links, Koramangala 

Inner Ring Road, Bangalore, 560071, India, kambhatla@in.ibm.com, 

Mayuri Duggirala, Ramana Polavarupu, Dinesh Garg 

We present a framework of provider-perceived service quality for contact center 

services, incorporating key dimensions of service quality based on interviews with 

service providers in contact center services. Our findings indicate that benchmarking 

and error management are significant provider-perceived dimensions of service 

quality in contact center services which predict business performance outcomes. 

Avenues for further research, as well as insights for research and practice are 

suggested. 

2 - Decision Support System for Continuous Production 

Krystsina Bakhrankova, Researcher, Institute of Technology and 

Society - Applied Economics, S. P. Andersens Veg 5, Box 4760 

Sluppen, Trondheim, 7465, Norway, 

Krystsina.Bakhrankova@sintef.no 

The paper develops a model-based decision support system (DSS) for a European 

chemical plant with a multi-stage continuous production process. The system 

comprises two modules - energy cost minimization and output maximization, where 

a gist of the two underlying formulations is presented. The planning tool is tested on 

real data instances - it reflects the essence of the researched production process, 

provides substantial energy cost savings and improved production capacity 

utilization. 

3 - Optimizing Long-range Plans at Novartis 

Brian Lewis, Vice President, Professional Services, 

Vanguard Software, 1100 Crescent Green, Cary, NC, 27518, 

United States of America, brian.lewis@vanguardsw.com 

Long-range strategic planning decisions are not easily modeled with classic 

optimization techniques. Using Monte Carlo simulation-based forecasting, 

simulation optimization, and grid computing, Vanguard Software built a drug 

development pipeline model for Novartis which forecasts long-range R&D 

performance and optimizes strategic decisions such as investments in new drugs. In 

our presentation, we discuss the underlying model, lessons learned, and other 

practical issues. 

■ WB19 


Revenue Management of Opaque and 

Non-traditional Channels 



Chair: Benjamin Marcus, Suffolk University, 8 Ashburton Place, Boston, 

MA, United States of America, bmarcus@suffolk.edu 

1 - Valuation of Opaque Products 

Leo MacDonald, Assistant Professor, Coles College of Business, KSU, 

Kennesaw, GA, United States of America, lmacdon4@kennesaw.edu, 

Benjamin Marcus 

Service providers (airlines, hotels, etc) often use opaque sales channels (Hotwire, 

Priceline) to increase revenues. A fundamental challenge for these providers is 

setting appropriate rates. Set rates too high and no purchase occurs; set rates too 

low and forgo the additional revenue. The focus of this research is to determine 

customer valuation of opaque versus non-opaque services through a choice 

experiment and develop a discrete-choice model to support the decision making 

process.

WB20 

2 - Pricing Opaque and Traditional Channels Jointly 

Benjamin Marcus, Suffolk University, 8 Ashburton Place, Boston, 

MA, United States of America, bmarcus@suffolk.edu 

Despite the benefits to both firms and consumers, there are distinct challenges to 

selling products through opaque channels. We develop a model of a service provider 

selling inventory across a traditional channel and an opaque channel in order to 

identify optimal pricing policies in this setting. In addition, we use this model to 

explore the effects that different characterizations of customer behavior and product 

commoditization on the opaque channel can have on these policies. 

■ WB20 


Pricing and Revenue Management I 


Chair: Craig Sorochuk, Assistant Professor of Decision Science, University 

of Wyoming, 1000. East University Avenue, Department of MGMT and 

MKT (#3275), Laramie, WY, 82071, United States of America, 

csorochu@uwyo.edu 

1 - HOT Lane Pricing for Revenue Generation and Congestion 

Management: An Analysis of Demand Responses 

Lin Qiu, Wilbur Smith Associates, 317 Center St. N, Vienna, 22180, 

United States of America, lin.w.qiu@gmail.com, Lei Zhang 

High Occupancy Toll Lane has attracted significant interests as a means for 

congestion management and revenue generation. Using combined stated-preference 

and revealed-preference data for the I-394 corridor, this study develops models to 

capture the inter-relationships between pricing schemes, travelers’ mode choices 

and dynamic attitudes towards HOT lanes. The estimated attitude and behavior 

changes form a reliable basis for projecting revenues and assessing congestion 

mitigation effects. 

2 - Market Share Characterization Through Scenario Analysis 

Amit Shinde, Research Associate, Arizona State University, 

699 S. Mill Avenue, Tempe, AZ, 85281, United States of America, 

amit.shinde@asu.edu, Mani Janakiram, George Runger 

The supply chain interactions within the high-technology industry are very 

complex. Multiple products with moderate differences in performance and price 

compete within the same market segment. We present data mining models for 

characterizing elements of such supply chains. These models are capable of 

assimilating knowledge from a variety of business scenarios, expert judgment and 

historical trends. 

3 - Designing Public Storage Warehouses with Customer Choice 

Yeming Gong, EM Lyon Business School, 23 Avenue Guy de 

Collongue, Ecully, France, gong@em-lyon.com 

Public storage is a booming industry. A major question is how to design public 

storage facilities to fit market segments to maximize revenue. This paper propose a 

method to design public storage warehouses with considering the choice behavior of 

customers. We solve the problem by column generation. 

4 - Computing Regulated Bertrand-Nash Equilibrium Prices Under 

Mixed Logit Demand 

William Morrow, Assistant Professor, Iowa State University, 2014 

Black Engineering Building, Ames, 50011, United States of America, 

wrmorrow@iastate.edu 

Bertrand-Nash equilibrium prices have been used to analyze mergers, new product 

introductions, and regulatory policy in large differentiated product markets. We 

present a framework with regulatory costs that may not be differentiable, as with 

the Corporate Average Fuel Economy Standards. Two nonsmooth fixed-point 

formulations of the first-order conditions are shown to provide reliable and efficient 

methods for computing equilibrium prices in a large-scale example from the 

automotive industry. 

5 - The Newsvendor Problem with Pricing and Secondary Revenues 

Craig Sorochuk, Assistant Professor of Decision Science, University of 

Wyoming, 1000. East University Avenue, Department of MGMT and 

MKT (#3275), Laramie, WY, 82071, United States of America, 

csorochu@uwyo.edu, John Wilson 

We present an expected profit model for a newsvendor who receives revenues from 

selling primary items as well as from selling secondary items that are only available 

if a primary item has already been purchased. Numerical examples are provided for 

a newsvendor whose primary items for sale are tickets to a performance event and 

whose secondary items for sale are concessions items, parking, etc. 


392 

■ WB21 


Green Supply Chain Management 



Chair: Vipul Jain, Assistant Professor, Indian Institute of Technology 

Delhi, Department of Mechanical Engineering, Indian Institute of 

Technology Delhi, New Delhi, 110016, India, vjain@mech.iitd.ac.in 

1 - A Green Approach to Supplier Selection 

Amit Kumar, Research Scholar, Indian Institute of Technology, 

Department of Mechanical Engineering, Delhi, New Delhi, India, 

akumar@icfi.com, Vipul Jain 

As the climate change movement gathers momentum, there’s a pressing need to 

assess suppliers based on their environmental performance along with other criteria. 

This paper proposes a comprehensive approach based on Data Envelopment 

Analysis with Carbon Footprint monitoring. The approach applies to heterogeneous 

suppliers and incorporates region specific emission compliance standards as well. 

Overall, it encourages suppliers to go green and cut down their emissions to survive 

the competition. 

2 - Carbon Footprint, Information Disclosure, and Shareholder Pressure 

Chien-Ming Chen, UCLA Institute of the Environment, La Kretz 

Hall, Suite 300, Box 951496, Los Angeles, CA, 90095, United States 

of America, cmchen@ioe.ucla.edu, Charles Corbett, Magali Delmas 

This study examines the causal relationship between corporate carbon efficiency, 

voluntary information disclosure, and shareholder’s pressure for greener business 

practices. Our analysis draws on the newly compiled direct and supply chain 

emission inventories of over 1000 public companies in North America. In the 

presentation we will present our preliminary findings. 

3 - Evaluation and Management on Logistics Carbon Emission 

Xiao Qing Wang, IBM Research, Building 19 Zhongguancun 

Software Park, Beijing, China, xqwangxq@cn.ibm.com, Jin Dong, 

Hongwei Ding, Minmin Qiu, Wei Wang 

Logistics carbon emission management is addressed and carbon emissions from 

several key operational stages in logistics industry are studied and evaluated. A 

general logistics carbon emission evaluation framework considering different 

transportation modes, different warehouses and different carry modes is proposed. 

Carbon emission evaluation methods on transportation, storage and carry 

operational stages are presented. 

4 - Measuring Carbon Emissions From Intermodal Freight Operations 

Anthony Craig, PhD Candidate, MIT, 77 Massachusetts Ave, 

E40-222, Cambridge, MA, 02139, United States of America, 

tcraig@mit.edu, Edgar Blanco, Yossi Sheffi 

Estimating the carbon emissions from intermodal shipments is difficult for shippers 

due to limited information and the complexity of intermodal operations. The 

structure of the rail network, terminal locations, and relative efficiency of rail and 

drayage operations all impact the actual emissions. Using data from an intermodal 

freight operator we compare the calculated carbon emissions for a set of shipments 

with the results obtained from popular carbon estimation methods. 

■ WB22 


Workforce Management 



Chair: Foaad Iravani, University of California-Los Angeles, Anderson 

School of Management, 110 Westwood Plaza, Los Angeles, CA, 90095, 

United States of America, firavani@anderson.ucla.edu 

1 - Skill Mix and Cross-Training in Professional Service Firms 

Vincent Hargaden, PhD Student, Rensselaer Polytechnic Institute, 

Industrial & Systems Engineering Dept, 110 8th Street, Troy, NY, 

12180, United States of America, hargav@rpi.edu, Jennifer Ryan 

A comprehensive mixed integer programming model has been developed for the 

workforce planning process in professional service firms. We will present results 

from the model which show the impact of skill mix, skill capability levels and crosstraining 

on key performance metrics such as project completion rates, staff 

utilization and profit.

2 - A Hiring Plan Model for Call Center Management 

Tao Huang, Progressive Insurance, 6300 Wilson Mills Rd, 

Mayfield Village, OH, 44143, United States of America, 

Tao_Huang@progressive.com, Janet Dolohanty, Steve Callitsis 

We have developed a hiring plan model for call center management. The model 

identifies locations and schedules for the hiring need defined by the capacity 

planning and staff-on-hand while taking into account site specific monetary 

variables and nonmonetary constraints. The model outputs the optimal combination 

of schedules that minimizes hiring cost and specifies the agents required to improve 

the peak-hour service level performance. 

3 - The Soft Resource Allocation Problem 

Foaad Iravani, University of California-Los Angeles, Anderson School 

of Management, 110 Westwood Plaza, Los Angeles, CA, 90095, 

United States of America, firavani@anderson.ucla.edu, Sriram Dasu, 

Reza Ahmadi 

We propose optimization models for workforce allocation in a leading software 

company that produces tax software. Every year, the firm struggles with a high 

workload imposed by changes in tax forms announced by the IRS. In this 

competitive market, any delay in the release of the product leads to significant 

losses. We develop models for organizing and staffing the development activities to 

meet the deadline at the lowest cost. 

■ WB23 


Municipal Waste Management, Analytics and 

Optimization 



Chair: Heng Cao, CTO for Business Analytics & Optimization, IBM China 

Research Center, A2/F, Diamond, Zhong Guang Cun Software, 

Haidian District, Beijing, 100193, China, hengcao@us.ibm.com 

1 - Decision Support System for Municipal Solid Waste Collection 

using Forecasting and Optimization 

Tianzhi Zhao, IBM Research - China, Diamond Building A, ZGC 

Software Park, 8 Dongbeiwang West Road, Beijing, 100193, China, 

zhaotzhi@cn.ibm.com, Jun Zhang, Jin Dong, Heng Cao, Wenjun Yin 

Municipal solid waste management (MSWM) is becoming a major issue facing cities 

around the world due to rapid urbanization and growth of population. In this paper, 

an analytics based decision support system is proposed for MSWM. The system 

model is composed of two components, one for waste generation prediction, 

another one for collection vehicle routing optimization. A GIS application is 

integrated into the system to provide route information to as well as map out the 

outputs from the model. 

2 - Disturbance Analysis Model for the Maintenance Plan of Power Grid 

Feng Jin, Dr., IBM Reaserch - China, Bulding 10, 399 Keyuan Road, 

Pudong, Shanghai, 201203, China, jinfsh@cn.ibm.com, Hairong Lv, 

Jun Luo, Wenjun Yin, Jin Dong, Qiming Tian 

To keep the maintenance plan of power gird stable, especially to avoid chainreaction, 

a probability model is proposed to analyze the effected plans once a plan is 

disturbed by various factors. In this model, we consider not only the traditional 

variation of start time, but also the variation of process time and workload under 

the complex grid topology. A case in a typical Chinese power company is studied to 

validate the model. The result shows the plan change rate is greatly reduced. 

3 - An Effort Estimation Model in Project Delivery using Hidden 

Setup Cost 

Saeed Bagheri, IBM T J Watson Research Center, 1101 Kitchawan 

Road, Yorktown Heights, NY, 10598, United States of America, 

sbagher@us.ibm.com, Nianjun Zhou, Krishna Ratakonda 

We discuss the relationship between delivered projects and required effort. In 

particular, we analyze the logarithmic model and its shortcomings in required effort 

estimation for large projects. We introduce the hidden setup cost and its related 

linear model and explain how its existent leads to the above behavior in logarithmic 

models. Our proposed model facilitates effort estimation for project delivery in 

services and manufacturing. We illustrate this, using projects in software 

development. 


393 

■ WB24 


Joint Session SPPSN/ HAS: Emergency Medical 

Services: New Directions 

Sponsor: Public Programs, Service and Needs/ Health Applications 


Chair: Laura McLay, Assistant Professor, Virginia Commonwealth 

University, 1015 Floyd Ave, Box 843083, Richmond, VA, 23284, 

United States of America, lamclay@vcu.edu 

1 - A Markov Chain Model for an EMS System with Repositioning 

Armann Ingolfsson, University of Alberta, Edmonton, AB, Canada, 

armann.ingolfsson@ualberta.ca, Ramon Alanis, Bora Kolfal 

We propose and analyze a Markov chain model of an Emergency Medical Services 

system that repositions ambulances using a compliance table policy, which is 

commonly used in practice. We validate the model against a detailed simulation 

model. We demonstrate that the model provides accurate approximations to such 

performance measures as the response time distribution and the distribution of the 

number of busy ambulances, and that it can be used to identify near-optimal 

compliance tables. 

2 - An Equitable EMS Location Model: Minimizing Envy Toward 

Neighbors’ Chance of Access to Service 

Sunarin Chanta, PhD Student, Clemson University, 203 Freeman 

Hall, Department of Industrial Engineering, Clemson, SC, 29634, 

United States of America, schanta@clemson.edu, Maria Mayorga, 

Laura McLay 

The model is developed for finding optimal locations in order to balance disparity in 

service between zones. The objective is to minimize the sum of “envy” among all 

zones with respect to an ordered set of p operating EMS stations weighted by the 

proportion of demand in each zone. Tabu search is provided to solve the problem. 

The performance of the proposed model is tested and compared to other location 

models such as the p-center and maximal-covering-location (MCLP) problems. 

3 - Estimating Travel Speeds in Road Networks From GPS Data 

Shane Henderson, Professor, Cornell University, School of ORIE, 

Rhodes Hall, Cornell University, Ithaca, NY, 14853, United States of 

America, sgh9@cornell.edu, Dawn Woodard, Brad Westgate, 

David Matteson 

We estimate the speeds vehicles travel on road networks from widely spaced GPS 

readings as recorded by ambulances. We use a Bayesian formulation and estimate 

the posterior distribution of road travel speeds and other quantities using Markov 

chain Monte Carlo (MCMC). In this talk I will sketch the problem and data, 

describe the MCMC algorithm, and present some sample results. 

4 - Fire Department and Other Emergency Medical Service 

John Hall, National Fire Protection Association, Norwood, MA, 

United States of America, jhall@NFPA.org 

People with emergency medical conditions may call an ambulance, possibly fire 

department based, and may seek medical assistance at any of several locations. This 

paper will discuss concepts and data sources and uses for a comprehensive system to 

track the flow of people in and out of emergency medical service. 

■ WB25 


Transportation, Planning I 


Chair: David Novak, Assistant Professor, University of Vermont, 

55 Colchester Ave., Kalkin 310, Burlington, VT, 05405-0157, 

United States of America, dnovak@bsad.uvm.edu 

1 - Exploring Theoretical Properties of Bounded Rational User 

Equilibrium Flow Distributions 

Yingyan Lou, Assistant Professor, The University of Alabama, A127K 

Bevill Building, Box 870205, Tuscaloosa, AL, 35405, 

United States of America, ylou@eng.ua.edu 

This paper investigates theoretical properties of the boundedly rational user 

equilibrium (BRUE) flow set for transportation networks. Probabilistic methods are 

explored to address the uncertainty in link flows characterized by the non-convex 

BRUE flow set. Entropy maximization is used to identify the most likely link flow 

pattern. Finite sampling approaches are also investigated to derive various measures 

from assumptions on the continuous link flow probability space over the BRUE flow 

set.

WB26 

2 - Toolkit for Anticipating and Evaluating Roadway Expansion and 

Tolling Impacts 

Daniel Fagnant, Graduate Research Assistant, The University of 

Texas at Austin, ECJ Suite 6.9, MailCode C1761, Austin, TX, 78712, 

United States of America, annette.perrone@engr.utexas.edu, 

Kara Kockelman, Chi Xie 

We present a new toolkit for forecasting the traffic volumes, travel times, network 

reliability, emissions, and safety impacts of roadway expansion and tolling projects. 

Trip tables are estimated using constrained maximum entropy methods, based on 

link-level traffic counts. Incremental logit models anticipate mode and time of day 

splits. All impacts are monetized and compared over project lifetimes, to generate 

benefit-cost ratios and other success indicators for Austin case studies. 

3 - Simulation of Vessel Traffic and Dredging Impact Analysis in 

Delaware River 

Ozhan Alper Almaz, PhD Student, Rutgers University, Industrial & 

Systems Engineering Departm, 100 Brett Road, Piscataway, NJ, 

08854, United States of America, alperalmaz@gmail.com, 

Tayfur Altiok 

We considered modeling of vessel traffic in the Delaware River Main Channel. A 

high fidelity simulation model was developed to investigate effects of deepening and 

dredging in the River on the navigational efficiency based on several assumptions. 

In this regard, vessel calls to terminals, lightering and barge operations, tidal and 

navigational rules in the River, terminal and anchorage properties and vessel 

profiles were considered. 

4 - Evaluating the Effects of Trip Importance on System-Wide 

Performance in Transportation Networks 

David Novak, Assistant Professor, University of Vermont, 

55 Colchester Ave., Kalkin 310, Burlington, VT, 05405-0157, 

United States of America, dnovak@bsad.uvm.edu 

We introduce measures for evaluating network robustness and identifying and 

ranking the most critical or important links in a transportation network called the 

Network Robustness Index, and comparing disparate networks using a scalable, 

system-wide performance over all links in the network called the Network Trip 

Robustness. We show that the relationships between network robustness, the 

capacity-disruption level, and network connectivity are non-linear and are not 

necessarily intuitive. 

■ WB26 


Joint Session DM/ ICS: Bayesian Computational Issues 

in Data Mining 

Sponsor: Data Mining/ Computing Society 


Chair: Tevfik Aktekin, Assistant Professor of Decision Sciences, University 

Of New Hampshire, 15 Academic Way, Department of Decision Sciences, 

Durham, NH, 03824, United States of America, tevfik.aktekin@unh.edu 

1 - Modeling Brazilian Swap Curve in a Hidden Markov Framework with 

Macroeconomic Variables 

Richard Munclinger, Economist, IMF, 1301 20th Street, NW, Apt. 

107, Washignton, DC, 20036, United States of America, 

richardmunch@gmail.com 

We apply a hidden Markov model of the term structure to modeling Brazilian swap 

rates. We find that multiple regimes are identified in the data and that 

macroeconomic variables improve time series fit without destroying regime 

dependency. This work has two main contributions. Firstly, we include 

macroeconomic variables in conjunction with a hidden Markov framework. 

Secondly, we propose and apply a Bayesian MCMC algorithm to estimate hidden 

Markov models of the term structure. 

2 - Explaining HIV Mortality, Bayesian Spatial Models Applied with 

MCMC Methods 

Rasim Muzaffer Musal, Assistant Professor, Texas State University, 

404 Rio Grande apt 209, Austin, TX, 78701, United States of 

America, rm84@txstate.edu, Tevfik Aktekin 

We propose Bayesian Zero Inflated Poisson models to investigate the effects of 

poverty and inequality on the number of HIV related deaths in NY counties. In 

doing so, we quantify inequality via the Theil Index and Poverty via the ratios of 

the two Census 2000 variables. MCMC methods are utilized in eliciting posteriors. 

We present the computational complexities that are present in these methods and 

emphasize the methods for spatial effects. 


394 

3 - Bayesian Analysis of Discrete Time Queueing Networks with a 

Gridlock Prediction Application 

Toros Caglar, George Washington University, 2201 G St., NW, Funger 

Hall, Suite 415, Washington, DC, 20052, United States of America, 

toros@gwu.edu, Refik Soyer 

Analysis of discrete time queues and their networks have been mostly prevalent in 

the context of computer and communication systems. In this study, we aim to 

utilize a Bayesian approach for the analysis of these queues with an application in 

emergency room (ER) gridlock prediction, which requires transient analysis. We will 

also address the challenges presented by the large state space formed by the ERhospital 

network and the MCMC simulation necessitated by the Bayesian analysis. 

4 - Bayesian State Space Modeling of Mortgage Default Risk 

Tevfik Aktekin, Assistant Professor of Decision Sciences, University of 

New Hampshire, 15 Academic Way, Department of Decision 

Sciences, Durham, NH, 03824, United States of America, 

tevfik.aktekin@unh.edu, Refik Soyer, Feng Xu 

We consider discrete time Bayesian state space models with Poisson measurements 

to model the mortgage default risk at the aggregate level with a stochastic default 

rate and macroeconomic covariates. We discuss parameter updating and estimation 

using Markov chain Monte Carlo methods where the use of the efficient forward 

filtering backward sampling algorithm within a Gibbs sampler is developed. We use 

actual U.S. residential mortgage data and discuss insights gained from Bayesian 

analysis. 

■ WB27 


Network Optimization I 


Chair: Chase Rainwater, University of Arkansas, INEG Department, 

Fayetteville, AR, 72701, United States of America, cer@uark.edu 

1 - Integrated Dynamic Single Facility Location and Inventory 

Planning Problems 

Jiaming Qiu, Student, Rensselaer Polytechnic Institute, 

110 8th St., Troy, NY, 12180, United States of America, qiuj@rpi.edu, 

Thomas Sharkey 

We analyze a class of problems which determines the relocation and inventory plan 

of a single facility over a finite horizon in order to meet dynamic customer 

demands. Dynamic programming algorithms are presented to solve the problem 

under different objective functions and construct the efficient frontier for biobjective 

problems, most of which run in polynomial time. 

2 - Hybrid Wired-cum-wireless Sensor Network Location-allocation 

Problem in Industrial Environment 

Sima Maleki, Graduate Research Assistant, University of 

Tennessee/Industrial and Information Engineering Department, 

Knoxville, Knoxville, TN, 37996, United States of America, 

smaleki@utk.edu, Mohammad Mehdi Sepehri, Hamid Farvaresh, 

Rapinder Sawhney 

A hybrid wired-cum-wireless sensor network consists of a wireless network and a 

wired backbone. The proposed designs consider limitations of wireless 

communication and constraints in industrial applications to minimize the network 

cost over a life of a network. The joint problem of configuring the hybrid network, 

locating the nodes and sensor clustering is formulated as a Mixed Integer Nonlinear 

Programming model. Results illustrate the cost effectiveness and longevity of hybrid 

configuration. 

3 - Social Optimality and Pricing in Shared Computing Centers 

Ishai Menache, postdoc, MIT, 77 Massachusetts avenue, 32-D632, 

Cambridge, MA, 02139, United States of America, ishai@mit.edu, 

Nahum Shimkin, Asuman Ozdaglar 

Motivated by the recent advent of cloud computing facilities that offer online 

computing power on demand, we consider a large service facility that offers 

simultaneous service to a large number of heterogeneous users. Our main concern 

here is in maximizing the social utility, which comprises of the users’ service utility 

minus their delay cost. We show that the social optimum may be achieved by 

simple per-unit pricing, which charges a fixed amount per unit time and resource 

from all users. 

4 - Methodologies for Solving Dynamic Fortification Problems 

Chase Rainwater, University of Arkansas, INEG Department, 

Fayetteville, AR, 72701, United States of America, cer@uark.edu, 

Huy-Nhiem Nguyen, Ed Pohl, Scott J. Mason 

The fortification of critical infrastructure elements is an issue of notable importance. 

However, research in this area has studied the problem of fortification from purely a 

static perspective. This work presents a dynamic model that considers the temporal 

impacts of resource allocation decisions in a potentially changing infrastructure. 

Specifically, we propose decomposition-based solution approaches to solve this 

general class of problems.

■ WB28 


Prognostics and Health Management 

(PHM / Sensors / RFID) 



Chair: Sagar Kamarthi, Associate Professor, Northeastern University, 

360 Huntington Ave, Boston, MA, 02115, United States of America, 

sagar@coe.neu.edu 

Co-Chair: Abe Zeid, Northeastern University, 360 Huntington Ave, 

Boston, 02115, United States of America, zeid@coe.neu.edu 

1 - A Bayesian Inventory Model using Real-time Condition 

Monitoring Information 

Jennifer Ryan, Associate Professor, RPI, Industrial and Systems 

Engineering, 110 8th Street, Troy, NY, 12180, 

United States of America, ryanj6@rpi.edu, Rong Li, Zhi Zeng 

We consider a manufacturer who periodically replenishes inventory for a machine 

part which is used simultaneously on a large set of geographically dispersed 

machines. This part is subject to deterioration, which can be captured by condition 

monitoring. We model the associated degradation signal using a Wiener process. In 

such a setting, we show how real-time sensor data can be used to improve the 

inventory management of spare parts. 

2 - Application of Fuzzy Graph on Sensor Deployment Strategy for 

Fault Diagnosis 

Zhenhua Wu, Research Assistant, TAMU, 3902 College Main Street, 

Apt.806, Bryan, TX, 77801, United States of America, 

wuzhenhua34@tamu.edu, Sheng-jen Hsieh, Jianzhi Li 

I am sorry It only allows 500 characters, I can not submit my abstract. Is there any 

way we can solve it? My email is: wuzhenhua34@tamu.edu Thank you! 

3 - Short Term Performance of Condition Monitored Degrading 

Manufacturing Systems 

Saumil Ambani, University of Michigan, 1210 HH Dow, Ann Arbor, 

MI, 48105, United States of America, sambani@umich.edu, Lin Li, 

Jun Ni 

Steady state performance of manufacturing systems has been studied for over 50 

years, but its short term behavior remains relatively unexplored. In this 

presentation, we focus on the development of an analytical model for predicting the 

short term performance of a condition monitored degrading system. This approach 

allows us to incorporate real time information from the plant floor, such as buffer 

contents and condition of the machines, to make effective short term decisions. 

■ WB29 


Joint Session QSR/ DM: Data Mining for 

Process Monitoring 

Sponsor: Quality, Statistics and Reliability/ Data Mining 


Chair: Myong Jeong, Rutgers University, 640 Bartholomew Road, 

Piscataway, NJ, United States of America, mjeong@rci.rutgers.edu 

1 - Hybrid Novelty Score-Based Control Charts for Multivariate 

Process Monitoring 

Gulanbaier Tuerhong, Korea University, Anam-dong, Seongbuk-gu, 

Seoul, Korea, Republic of, gulambar@korea.ac.kr, Seoung Bum Kim 

We propose a new nonparametric multivariate control chart that can effectively 

handle large amounts of complex data through integration of one of the one-class 

classification algorithm with control chart techniques. Control limits of the proposed 

chart are established based on a bootstrap method. Experimental results with 

simulated data showed that the proposed control chart outperformed Hotelling’s T2 

control charts. 

2 - An Economic Design of the Integrated Process Control 

Minjae Park, Rutgers University, 126 Montgomery St. Apt 2F, 

Highland Park, NJ 08904, United States of America, 

pminj88@gmail.com 

The economic cost model is developed for the integration of statistical process 

control (SPC) and automatic process control (APC). SPC is used to detect special 

causes by monitoring process performance. On the other hand, APC is used to 

improve the process by an adjustment controller. Both are needed to effectively 

keep processes close to target. The long run expected cost is also suggested as the 

criteria to evaluate the performance of economic cost model. 


395 

3 - Detection of Potential Failure Wafers Based on Fail Bit Counts Data 

Seung Hoon Tong, Principal Engineer, Samsung Electronics Co., Ltd., 

San#16 Banwol-Dong, Gyeonggi-Do, Hwasung-City, 445-701, Korea, 

Republic of, shtong@samsung.com, In Kap Chang, Jeong Hee 

Hwang, Kun Han Kim, Seung Sik Jung 

We present a method for detection of potential failure wafers based on fail bit 

counts data from wafer electrical test stage in semiconductor manufacturing. The 

number of fail bit each wafer has superior capability for classifying potential failure 

wafers than wafer yield itself. We developed two quality measures considering the 

magnitude and clustering level of fail bit patterns and showed real applied 

examples. 

4 - Logistic Regression-Based Control Charts for Fault Identification 

Jihoon Kang, Korea University, Anam-dong, Seongbuk-gu, Seoul, 

Korea, Republic of, joker404@hanmail.net, Seoung Bum Kim 

Identification of process variables that contribute to out-of-control signal has been 

one of important problems in statistical process control. In the present study we 

integrated a logistic regression algorithm with control chart techniques for fault 

identification in multivariate process control. Simulation studies demonstrated the 

accuracy and efficiency of the proposed fault identification method. 

■ WB30 


Statistics/Quality Control I 


Chair: Michael Wood, Principal Lecturer, University of Portsmouth, 

Strategy and Business Systems, Richmond Building, Portland St, 

Portsmouth, PO1 3DE, United Kingdom, michael.wood@port.ac.uk 

1 - A New Monitoring Scheme for Various Types of Mean Shifts 

Chang-Ho Chin, Assistant Professor, Kyung Hee University, 

1 Seocheon-dong, Yongin-si, 446-701, Korea, Republic of, 

chin@khu.ac.kr, Taek-Jin Jeong 

The EWMA control chart is more sensitive to small mean shifts than to large ones as 

opposed to the Shewhart control chart. The right selection of control charts to the 

underlying process is directly linked with the timely detection of mean shifts. For 

processes possibly with various types of mean shifts, we propose a control charting 

scheme of selecting and applying a right control chart to the process. Simulation 

results show the proposed scheme outperforms conventional ones. 

2 - An Adaptive Sequential Methodology for n-Dimensional Quadratic 

Response Surface Optimization 

Adel Alaeddini, Wayne State University, Department of Industrial 

Engineering, Detroit, MI, 48202, United States of America, 

dz3027@wayne.edu, Kai Yang, Alper Murat 

Despite their ability in modeling a wide range of process and product design, RSM 

techniques are not the most effective methodologies for applications with limited 

resources. In this paper, we develop an adaptive methodology for response surface 

optimization (ASRSM) for expensive experiments with noisy data requiring high 

design performance. We also show that in terms of both design optimality and 

experimentation efficiency it compares favorably with optimal designs. 

3 - Modeling Data using Kalman Filtering: A Proposal for SPC 

Andre Korzenowski, Prof. Ms., PPGEP/UFRGS, Av. Osvaldo Aranha, 

99 - 5° andar, Porto Alegre, RS, 90.035-190, Brazil, 

andre@korzenowski.com, Marcelo Portugal, Carla ten Caten 

The aim of this paper is to compare forecasting in univariate time series models 

applied in Statistical Process Control (SPC). A structural model in state space via 

Kalman’s filtering and an ARIMA model with exogenous variables were fitted. The 

modeling was applied in real data series of five different products in a plastic 

production. The results obtained indicate that structural model had better 

performance than ARIMA model when compared by RMSE. 

4 - Problems with Statistical Methods in Management Research - and a 

Few Solutions 

Michael Wood, Principal Lecturer, University of Portsmouth, Strategy 

and Business Systems, Richmond Building, Portland St, Portsmouth, 

PO1 3DE, United Kingdom, michael.wood@port.ac.uk 

A case study of a typical journal paper leads to three conclusions. (1) The value of a 

statistical approach is seriously limited by various factors: e.g. difficulties of 

generalizing to contexts other than the sample studied. (2) The conventional 

hypothesis testing format makes results almost meaningless: instead, I suggest using 

confidence levels for hypotheses - and suggest two ways of doing this (one a 

bootstrap method on a spreadsheet). (3) The analysis should be far more userfriendly.

WB31 

■ WB31 


Forecasting II 


Chair: Nitin Shenoy, Teaching Assistant, Texas Tech University, Texas Tech 

Industrial Engineering Department, Lubbock, Tx, 79409-3061, United 

States of America, nitin.shenoy@ttu.edu 

1 - Wrong Response Functions: Their Detection and Implications 

Steven Shugan, Professor, University of Florida, 2030 NW 24th 

Avenue, Gainesville, FL, 32605, United States of America, 

sms@ufl.edu 

This paper first explains why simple tests are unable to detect wrong response 

functions. Next, the paper shows that wrong response functions use dampening and 

inflated parameters to filter error variance to obtain better predictions. This 

dampening is problematic when decisions require the correct response. Fortunately, 

it is possible to detect dampening with specific tests for dampening. This paper 

shows how to do that and demonstrates the effectives of those tests. 

2 - Role of Forecast Effort on Supply Chain Profitability Under Various 

Information Sharing Scenarios 

Linda (Xiaowei) Zhu, West Chester University of PA, 312A 

Anderson, West Chester, United States of America, 

xzhu@wcupa.edu, Xuemei Su, Samar Mukhopadhyay, Xiaohang Yue 

We analyze several forecast systems and their impact on forecast accuracy, forecast 

costs and profit. Specifically, we consider a supplier who sells a product to a buyer 

in a single selling season. Three different forecast systems, namely, Non-Information 

Sharing, Information Sharing, and Supplier Forecasting were studied. We derive 

optimal price and forecast accuracy level and discuss forecast variance and the 

related forecast costs. 

3 - Unpacking the Future: A Nudge Towards Wider Interval Forecasts 

Kriti Jain, Doctoral Student, INSEAD, 1, Ayer Rajah Avenue, 

Singapore, Singapore, kriti.jain@insead.edu, Kanchan Mukherjee, 

Neil Bearden, Anil Gaba 

With few exceptions, forecasters tend to underestimate uncertainty, for example, a 

typical analyst’s 90% confidence intervals for future quantities will likely capture 

only 50-60% of the actual realizations. Using a series of lab and field experiments, 

we show that unpacking the distal future into intermediate more proximal futures 

has a substantial effect on subjective forecasts. We refer to this phenomenon as the 

time unpacking effect and show that it persists with expert judgments as well. 

4 - The Welfare Implications of Carbon Taxes and Carbon Caps: A 

Look at U.S. Households 

Kara Kockelman, The University of Texas in Austin, ECJ Suite 6.9, 

Austin, TX, United States of America, kkockelm@mail.utexas.edu, 

Binny Paul, Sumala Tirumalachetty 

Household expenditure and vehicle choice data are used here to anticipate the 

economic impacts of energy taxes versus household-level carbon-emissions caps 

(with trading) across different income classes. Translog utility models were 

estimated and then maximized subject to money and carbon budget constraints. 

U.S. trade accounts were used to infer the carbon footprints of 9 goods’ sectors, with 

vehicle choice and fuel economy modeled in detail. Cap and trade was found less 

regressive. 

5 - Propane Demand Modeling for Residential Sectors - 

A Regression Analysis 

Nitin Shenoy, Teaching Assistant, Texas Tech University, Texas Tech 

Industrial Engineering Department, Lubbock, Tx, 79409-3061, 

United States of America, nitin.shenoy@ttu.edu, Milton Smith 

When winter space heating contributes significantly to the propane demand system; 

it is useful, for forecasting purposes, to separate total demand system in to two 

components: Large houses and Small houses. Examination of historical data 

indicates that the propane demand depends largely on weather conditions. 

Regression analysis techniques were used to show the combined effect of these 

weather conditions on propane demand. 


396 

■ WB32 


Computing Economic Equilibria 



Chair: James Orlin, Professor, Massachusetts Institute of Technology, E53- 

363, Cambridge, MA, 02139, United States of America, jorlin@mit.edu 

1 - A New Convex Program for Fisher Markets and Convergence of 

Proportional Response Dynamics 

Nikhil Devanur, Microsoft Research, 1 Microsoft Way, Redmond, 

WA, United States of America, nikdev@microsoft.com 

We give a new convex program that captures the equilibrium of spending constraint 

utilities, thus resolving an open problem. The new program also demystifies the 

convergence properties of “proportional response” dynamics: a simple, distributed 

way to converge to market equilibria. We show that the proportional response 

dynamics is equivalent to gradient descent on the new convex program, with KLdivergence 

instead of Euclidean distance. 

2 - Equilibrium Computation for Low Dimensional Markets 

Amin Saberi, Stanford University, Terman Engineering Building, 

Room 317, Stanford, CA, 94305, United States of America, 

saberi@stanford.edu 

I will describe a model of markets in which every good has a value or desirability in 

each of the k dimensions. The utility of every agent in a bundle is a function of the 

values of the goods in that bundle in each dimension. In this model, I will present 

polynomial-time approximation schemes for computing the equilibria when k is 

bounded. Joint work with Costis Daskalakis 

3 - Improved Algorithms for Computing Fisher’s Market Prices 

James Orlin, Professor, Massachusetts Institute of Technology, 

E53-363, Cambridge, MA, 02139, United States of America, 

jorlin@mit.edu 

Irving Fisher developed a simple model of a market economy. Nevertheless, it is still 

non-trivial to develop efficient algorithms for determining the market clearing 

prices. We develop a combinatorial algorithm for computing the market equilibrium 

that runs in O(n^4 log n) time, improving the previous best bound of O(n^8 log U). 

■ WB33 


Business Analytics and Optimization Practices 



Chair: Jin Dong, IBM Research - China, Building 19 ZGC Software Park, 

8 Dongbeiwang West Road, Beijing, 100193, China, dongjin@cn.ibm.com 

1 - Supply Risk Management using Approximate 

Dynamic Programming 

Lei Zhao, Tsinghua University, Department of Industrial Engineering, 

Tsinghua University, Beijing, 100084, China, lzhao@tsinghua.edu.cn, 

Xiaobo Zhao, Weijun Ding, Jiarui Fang, Jan Fransoo 

We consider a production system that is subject to failures at part suppliers. 

Mitigation planning and emergent capacity can be used to hedge against the risk at 

a higher cost. We model and solve the problem using finite-horizon approximate 

dynamic programming. 

2 - Mail Performance Measurement Study Process Improvements via 

Business Analytics & Optimization at IBM 

Hua Ni, IBM, 8000 Grainger Ct, Springfield, VA, 22153, United 

States of America, huani@us.ibm.com, Christine Friesz 

IBM conducts an ongoing measurement study of single-piece First-Class Mail service 

performance for the United States Postal Service. In this talk, we will present recent 

process improvement initiatives designed to strengthen the quality and operational 

efficiency of the measurement study, which leveraged Business Analytics and 

Optimization (BAO) methodologies such as optimization, simulation, and 

predicative analytics. We will also share the challenges and lessons learned from 

these efforts. 

3 - Using Simulation in Global Supply Network Rationalization 

Changrui Ren, IBM, Building 19 ZGC Software Park, 8 Dongbeiwang 

WestRoad, Haidian District, Beijing, China, rencr@cn.ibm.com, 

Xu Yang, Jin Dong, Qinhua Wang, Bing Shao, Miao He 

This presentation will introduce a real case showing how simulation could be the 

key enablement for a supply chain network rationalization project in the 

pharmaceutical industry. An IBM tool - Supply Chain Process Modeler (SCPM) has 

been applied in this project. Five major simulation scenarios are designed based on 

the pain points the client has, and numerical results show that simulation has 

addressed the key issues and provided the client accountable evaluation results for 

decision-making.

4 - WISE-BPM: Accelerating Blueprint for Rollout Project in SAP 

System Implementation 

Qinhua Wang, IBM Research - China, Building19, Zhongguancun 

Software Park, Beijing, 100193, China, wangqinh@cn.ibm.com, 

Bing Shao, Miao He, Changrui Ren, Jin Dong 

In SAP system implementation, business process model of a rollout project should 

be conducted through localizing the global template. Simultaneously, global 

template changes should be synchronized to all rollout projects, which causes 

conflicts, information loss and duplication easily. WISE-BPM addresses this problem 

through an additional version control system to rollout global template changes, a 

localization flag and a mapping system to avoid information duplication and loss 

respectively. 

■ WB34 


Stochastic Optimization and Equilibrium Problems: 

Analysis and Stochastic Approximation Algorithms 



Chair: Uday Shanbhag, Asst. Professor, University of Illinois at Urbana 

Champaign, Urbana, Il, United States of America, udaybag@illinois.edu 

1 - Simulation-based Optimization in the Presence of Convexity 

Eunji Lim, Assistant Professor, University of Miami, University of 

Miami, Coral Gables, United States of America, lim@miami.edu 

We consider the problem of computing a response surface when the underlying 

function is known to be convex. We introduce a methodology that incorporates the 

convexity into the function estimator. The proposed response surface estimator is 

formulated as a quadratic program and exhibits convergence properties as a global 

approximation to the true function. Numerical results will illustrate the convergence 

behavior of the proposed estimator and its potential application to simulation 

optimization. 

2 - Single Timescale Regularized Stochastic Approximation Schemes 

for Monotone Stochastic Nash Games 

Jayash Koshal, Department of Industrial & Enterprise Systems 

Engineering, UIUC, 117 Transportation Building, 104 S. Mathews, 

Urbana, 61801, United States of America, koshal1@illinois.edu, 

Angelia Nedich, Uday Shanbhag 

We consider the distributed computation of equilibria arising in monotone stochastic 

Nash games over continuous strategy sets. We develop single timescale projectionbased 

stochastic approximation schemes for computing equilibria when the 

associated gradient map is merely monotone and establish its global convergence. In 

an extension where players choose their steplengths independently we claim the 

convergence of the scheme if the deviation across their choices is suitably 

constrained. 

3 - On the Characterization of Solution Sets of Smooth and Nonsmooth 

Stochastic Nash Games 

Uma Ravat, Graduate Student, University Of Illinois, Urbana, IL, 

61801, United States of America, ravat1@illinois.edu, 

Uday Shanbhag 

Solution sets of deterministic Nash games over continuous strategy sets can be 

characterized using variational analysis. A direct application of such results to 

stochastic regimes is challenging as the expectation yields a far less tractable 

nonlinear function. We develop an analytical framework to examine existence of 

stochastic Nash equilibrium under general probability measures in possibly 

nonsmooth regimes. We apply it to classes of Nash-Cournot games with risk and 

stochastic constraints. 

4 - Optimal Coverage of Rectangular Request Areas using Multiple 

Rectangular Targets 

Manish Bansal, PhD Student, Texas A&M University, 3131 TAMU, 


bans1571@neo.tamu.edu, Kiavash Kianfar 

We consider positioning k target rectangles on 2-dimensional plane to partially 

cover a set of existing rectangular areas (requests) to maximize total coverage 

reward. Applications include camera surveillance and imaging. We show this 

problem is NP-hard and present a novel branch-and-bound algorithm over a 

reduced solution space to solve the problem exactly. To our knowledge no 

algorithm has been proposed before for this problem and our algorithm is memory 

and performance efficient. 


397 

■ WB35 


Advances in Anomalous Diffusion II 



Chair: Iddo Eliazar, Professor, Holon Institute of Technology, P.O. Box 

305, Holon, 58102, Israel, eliazar@post.tau.ac.il 

1 - Anomalous Mixing and Reaction Induced by Spatially 

Fractional Dispersion 

David Benson, Colorado School of Mines, 1500 Illinois St., Golden, 

CO, 80401, United States of America, dbenson@mines.edu, 

Diogo Bolster, Tanguy Le Borgne, Marco Dentz 

Long-range mass transfer changes the character of mixing of two fluids of different 

chemical composition and the consequent chemical reaction. For mixing-limited 

equilibrium reactions following the space-fractional Advection-Dispersion equation 

(fADE), the scalar dissipation and global reaction rates decay as power-laws in time. 

As opposed to the Fickian (local) transport model, local reaction rates are not zero 

where the concentration has zero gradient. 

2 - Diffusive Processes Run with Non-linear Clocks: Complexity, 

Ergodicity and Fractional F-P Equations 

John Cushman, Professor, Purdue University, 550 Stadium Mall 

Drive, West Lafayette, IN, 47906, United States of America, 

75674.1670@compuserve.com 

We study stochastic processes run with deterministic, but non-linear clocks. The 

clock is used to stretch or compress the process. The clock does not change the 

fractal dimension or complexity of the process. Ergodicity is analyzed for several 

processes and compared to their classical analogs. Fokker-Planck equations for 

several of the processes are derived and in some cases shown to possess fractional 

derivatives. A number of misconceptions concerning power-law second moments 

are discussed. 

3 - Influence of Anomalous Diffusion on the Robustness and Time 

Evolution of Morphogen Gradients 

Katja Lindenberg, Distinguished Professor, University of California, 

San Diego, Department of Chemistry & Biochemistry, 9500 Gilman 

Drive MC 0340, La Jolla, CA, 92093-0340, United States of America, 

klindenberg@ucsd.edu, Santos B. Yuste, Enrique Abad 

Crowded cellular environments are subdiffusive. The usual picture of morphogen 

gradients evolving as a result of a source of normal diffusive particles (morphogens) 

coupled to a degradation mechanism that removes them from the system must be 

modified. We aim to investigate how subdiffusive motion of the morphogens affects 

the evolution and robustness of the concentration profile against changes in the 

degradation rate. We find interesting fluctuation buffering effects caused by 

subdiffusion. 

■ WB36 


Education I 


Chair: Thomas Groleau, Associate Professor of Business Administration, 

Carthage College, 2001 Alford Park Drive, Kenosha, WI, 53140, 

United States of America, tgroleau@carthage.edu 

1 - Vertical Integration: Results From a Cross-Course 

Student Collaboration 

David Lewis, University of Massachusetts Lowell, 1 University 

Avenue, College of Management, Lowell, MA, 01854, 

United States of America, David_Lewis@uml.edu, Thomas Sloan 

We report the results of a cross-class project involving Sophomore-level students in 

a Management Science (MS) class with Junior-level students in an Operations 

Management (OM) class. The students formed virtual teams and developed a 

simulation model of a call center. Results were strongly linked with the presence or 

absence of a team champion. 

2 - Approach to Improve the Expense Efficiency of Research Capacity 

of Universities 

Jongwoun Youn, PhD Candidate, KUBS(Korea University Business 

School), Anam-Dong, Seongbuk-Gu, Seoul, Korea, Seoul, Kr, 136- 

701, Korea, Republic of, yjw333@korea.ac.kr, Kwang-Tae Park 

Competitiveness of university can be expressed by education and research capacity. 

Research capacity is an essential indicator for estimating the development of 

university as well as that of our society. Thus, we want to propose the approach to 

improve the expense efficiency of research capacity.

WB37 

3 - Teaching Decision Tree Classification using Microsoft Excel 

Thaddeus Sim, Assistant Professor of Business Administration, 

Le Moyne College, 1419 Salt Springs Road, Syracuse, NY, 13214, 

United States of America, simtk@lemoyne.edu, George Kulick, 

Kaan Ataman 

Data mining is the extraction of useful patterns from data to aid with decision 

making, and decision makers are increasingly viewing data mining as an essential 

analytical tool. Unfortunately, data mining does not get as much attention in the 

traditional OR/MS classroom as other more popular areas. We discuss our 

experiences in teaching a popular data mining method in an undergraduate OR/MS 

elective course, and outline a procedure to implement the decision tree algorithm in 

Microsoft Excel. 

4 - A Different IE Teaching Experience: Active Learning 

Ali Kefeli, PhD Candidate, North Carolina State University, 

2717 Brigadoon Dr., Apt 24, Raleigh, NC, 27606, 

United States of America, akefeli@ncsu.edu, Michael G. Kay 

In this work, we take findings of a Felder and Brent (2008) and provide a critical 

analysis of a senior level industrial engineering class at North Carolina State 

University. We investigate each mistake laid out by Felder and Brent and argue how 

avoiding them resulted in a better educational outcome. We provide sample 

assessment tools such as class assignments and midterm questions, as well as active 

learning exercises, technological tools and student evaluations. 

5 - Pre-Analytics with Cash for Clunkers 

Thomas Groleau, Associate Professor of Business Administration, 

Carthage College, 2001 Alford Park Drive, Kenosha, WI, 53140, 

United States of America, tgroleau@carthage.edu 

In summer 2009 nearly 700,000 new vehicles were purchased through the U.S. 

government Cash Allowance Rebate System (CARS). The resulting dataset provides 

a freely available pre-analytics training ground for a variety of classes. Using basic 

desktop productivity software, students can practice data cleansing, data organizing, 

and basic statistics on a very large dataset. Student can also prepare the data for 

more detailed analysis with specialized geospatial or statistical software. 

■ WB37 


Queueing Models of Call Centers 



Chair: Itai Gurvich, Assistant Professor, Kellogg School of Management, 

Northwestern University, 2001 Sheridan Rd., Evanston, IL, 60208, 

United States of America, i-gurvich@kellogg.northwestern.edu 

1 - Approximate Dynamic Programming Techniques for Skill-based 

Routing in Call Centers 

Sandjai Bhulai, VU University Amsterdam, Department of 

Mathematics, Faculty of Sciences, Amsterdam, 1081 HV, 

Netherlands, sbhulai@few.vu.nl, Dennis Roubos 

We consider the problem of dynamic multi-skill routing in call centers. We obtain 

near optimal dynamic routing policies that are scalable with the size of the problem 

instance and can be computed online. The algorithm is based on approximate 

dynamic programming techniques. We compare the performance with 

decomposition techniques. Numerical experiments demonstrate that our method 

outperforms leading routing policies and has close to optimal performance. 

2 - Comparing Erlang C and A for Modeling Real Call Centers 

Tom Robbins, Assistant Professor, East Carolina University, 1006 

Gemstone Circle, Winterville, 28590, United States of America, 

Robbinst@ecu.edu, DJ Medeiros, Terry Harrison 

Erlang C is a widely used model for call centers. Erlang C makes many assumptions 

with one of the most problematic being no abandonment. Many authors have 

recently advocated the use of the more complicated Erlang A model. We present the 

results of a simulation study that compares the performance of these models over a 

range of realistic call center conditions. We find that while the Erlang A model is 

often more accurate, it is not always so. We also find the models have different 

biases. 

3 - Routing to Manage Resolution and Waiting Time in Call Centers 

with Heterogeneous Servers 

Yong-Pin Zhou, Michael G. Foster School of Business, Box 353200, 

University of Washington, Seattle, WA, 98195-3200, United States of 

America, yongpin@uw.edu, Kevin Ross, Geoff Ryder, Vijay Mehrotra 

In many call centers, agents exhibit very different performance for the same call 

type, where performance is defined by the average call handling time and the call 

resolution probability. We explore strategies for determining call routing policies, 

where call assignments are dynamically determined based on the specific attributes 

of the agents and/or the current state of the system. We test several strategies using 

data obtained from a financial service firm and present empirical results. 


398 

4 - Call Centers with Hyperexponential Patience 

Alex Roubos, VU University Amsterdam, Department of 

Mathematics, De Boelelaan 1081A, Amsterdam, 1081 HV, 

Netherlands, aroubos@few.vu.nl 

We show that customers’ patience can be modeled by the hyperexponential 

distribution; which allows an exact analysis. A framework is developed in order to 

compute all kinds of practical service levels. This framework utilizes the recursive 

relation between the queue lengths at successive service completion epochs. Our 

approach shows overall better performance compared to current algorithms. 

Moreover, the computation times are short and our approach can therefore readily 

be applied in practice. 

■ WB38 


First-order Optimization Methods and 

Their Applications 




Chair: Zaiwen Wen, NSF Math Institutes’ postdoc, United States of 

America, zw2109@columbia.edu 

1 - Bundle-type Methods Uniformly Optimal for Smooth and 

Non-smooth Convex Optimization 

Guanghui Lan, Assistant Professor, University of Florida, 

Department of Industrial and Systems Engineering, Gainesville, FL, 

United States of America, glan@ise.ufl.edu 

The study of Bundle or level-type methods has been focused on non-smooth convex 

programming (CP) problems. In this talk we present new bundle-type methods 

which are optimal, not only for non-smooth, but also for smooth convex 

programming problems. Surprisingly, the optimal rates of convergence are obtained 

without even requiring any smoothness information. We also demonstrate the 

superior practical performance of these methods over existing optimal methods for 

convex programming. 

2 - Fast Splitting and Alternating Linearization Methods for 

Convex Optimization 

Shiqian Ma, Columbia University, 500 W. 120th Street, Mudd Blvd, 

Room 313, New York, NY, 10027, United States of America, 

sm2756@columbia.edu, Donald Goldfarb 

In this talk, we present two classes of splitting and alternating direction methods for 

which we can obtain iteration complexity bounds. The basic and accelerated 

versions of these methods require 1/e and 1/e^(1/2) iterations to obtain an eoptimal 

solution, respectively. These complexity results are the first ones that have 

been given for splitting and alternating direction type methods. 

3 - A Unified Approach for Minimizing Composite Norms 

Necdet Aybat, Columbia University, IEOR Department, 

Columbia University, New York, NY, United States of America, 

nsa2106@columbia.edu, Garud Iyengar 

FALC is a first-order augmented Lagrangian algorithm to solve 

min{mu1|X|_*+mu2|C(X)-d|_1:A(X)=b},where X in R^{m*n},by inexactly solving a 

sequence of problems.FALC converges to the optimal X_* if it is unique. For all e>0, 

iterates are e-feasible, e-optimal in O(1/e) iterations.FALC can also solve: 

min{mu1|X|_p+mu2|C(X)-d|_q+:A(X)=b, F(X)-G is psd, |H(X)-h|_r

2 - A Probabilistic Comparison Between Split Cuts and Type 1 

Triangle Cuts 

Qie He, School of Industrial & Systems Engineering, Georgia 

Institute of Technology, 765 Ferst Drive NW, Atlanta, GA, 30332, 

United States of America, qie.he@gatech.edu, George L. Nemhauser, 

Shabbir Ahmed 

The nontrivial facets for mixed-integer programs (MIP) with two rows and two 

integer variables are split (or Gomory) cuts, three types of triangle cuts, and 

quadrilateral cuts. This talk presents a probabilistic comparison of split and type 1 

triangle cuts. Under a reasonable distribution of the problem parameters of the MIP, 

we show that the average performance of a single split cut is better than a single 

type 1 triangle cut in terms of dominance and volume cut off from the linear 

relaxation. 

3 - A Polyhedral Study of the Triplet Fromulation for Single Row Facility 

Layout Problem 

Sujeevraja Sanjeevi, PhD Student, Texas A&M University, TAMU 

3131, College Station, TX, 77843-3131, United States of America, 

sujeevraja@tamu.edu, Kiavash Kianfar 

We present a polyhedral study of the triplet formulation for the Single Row Facility 

Layout Problem, introduced by Amaral (Dis. App. Math., 2009). We show that this 

polytope is of dimension n(n-1)(n-2)/3, where n is the number of facilities. We then 

prove that several valid inequalities (VIs) proposed by Amaral are facet-defining. 

This provides a theoretical support for the strength of the LP lower bounds obtained 

by Amaral. We also present a generalized class of VIs that encompass Amaral’s VIs. 

4 - Finding Good MIP Solutions by Restricted Tree Search 

Menal Guzelsoy, Georgia Institute of Technology, 

mguzelsoy@gatech.edu, George L. Nemhauser, Martin Savelsbergh 

Starting with the branch-and-bound tree associated with the solution of a restricted 

mixed integer program (MIP), i.e., a MIP in which some variables are fixed, we 

expand certain nodes of the search tree by using dual information to selectively free 

previously fixed variables in the hope of quickly finding improved solutions. 

■ WB40 


Panel Discussion: COIN-OR Technology Forum 

Cluster: John Forrest-fest | COIN-OR 10th (Joint Cluster Computing) 


Moderator: Ted Ralphs, Associate Professor, Lehigh University, 200 West 

Packer Avenue, Bethlehem, PA, 18015, United States of America, 

ted@lehigh.edu 

1 - Panel Discussion: COIN-OR Technology Forum 

Panelists: Ted Ralphs, Associate Professor, Lehigh University, 200 

West Packer Avenue, Bethlehem, PA, 18015, United States of 

America, ted@lehigh.edu, Lou Hafer, Simon Fraser University, 

Burnaby BC V5A 1S6, Canada, lou@cs.sfu.ca, William Hart, Sandia 

National Laboratories, PO Box 5800, Albuquerque NM, United States 

of America, wehart@sandia.gov, Kipp Martin, University of Chicago, 

5807 South Woodlawn, Chicago IL 60637, United States of America, 

kmartin@chicagobooth.edu 

Following up on last year’s successful forum, this panel discussion will be an 

opportunity for users and developers of COIN-OR software to discuss recent and 

future developments within COIN-OR. If you want to get involved, provide 

feedback, or just learn about COIN-OR, please join us! 

■ WB41 


Vehicle Routing II 


Chair: Sam Thangiah, Professor, Slippery Rock University, 250 ATS, 

Computer Science Department, Slippery Rock, PA, 16057, 

United States of America, sam.thangiah@sru.edu 

1 - Backhaul Vehicle Routing Problem with Time Constraint 

Yuanyuan Dong, Southern Methodist University, P.O. Box 750123, 

Dallas, TX, 75275-0123, United States of America, 

njdyy03@gmail.com, Junfang Yu 

A backhaul vehicle routing problem has been studied in which maximum profit is 

desired during backhaul trip but travel time is constrained. A heuristic algorithm, 

adapted from genetic algorithm, has been developed and implemented for the 

problem. Empirical study has been performed to show the efficiency and 

effectiveness of the algorithm. 


399 

2 - Hybrid Genetic Algorithm for the Split Delivery Vehicle 

Routing Problem 

Joe Wilck, The University of Tennessee - Knoxville, 411 East 

Stadium, Knoxville, TN, United States of America, jwilck@utk.edu, 

Gautham Rajappa, Michael Vanderlan 

The SDVRP allows customers to be assigned to multiple routes. A genetic algorithm 

is presented where results from a construction heuristic are used to rank the 

importance of certain node relationships. If nodes 2 and 9 appear on the same route 

in many solutions, and those solutions have significantly lower distances when 

compared to solutions where nodes 2 and 9 are not on the same route, then the 

nodes 2 and 9 have a higher probability of being placed on the same route in future 

generations. 

3 - Solving Large Scale Dial-A-Ride Problem using a Two-Stage 

Heuristic Based on Clustering-Routing 

Taehyeong Kim, University of Maryland, Dept of Civil & 

Environmental Eng, 1173 Glenn L. Martin Hall, College Park, MD, 

20742, United States of America, tommykim@umd.edu, Ali Haghani 

In this paper, a static DARP model with time varying travel times and multiple 

depots is formulated. Also, a heuristic methodology using two-stage is proposed to 

solve this problem. At first stage, initial solution is constructed using clusteringrouting 

algorithm. And then, it is improved at next stage. The model is used to solve 

a real world problem that is provided by the MTA in Baltimore, MD and the results 

of implementing the model are compared with those of current system operations. 

4 - Multi-depot, Multi-Destination, Mix-Fleet Vehicle Routing Problem 

with Real-Life Constraints 

Sam Thangiah, Professor, Slippery Rock University, 250 ATS, 

Computer Science Department, Slippery Rock, PA, 16057, 

United States of America, sam.thangiah@sru.edu, Joseph Forsythe 

This paper presents a real-life multi depot, multi-destination, mix-fleet vehicle 

routing problem with 1200 pickup locations, 2000 customers and 60 vehicles. The 

precise distances between pickup locations are obtained using digitized road maps 

instead of Euclidean distance. Heuristics for solving the problem and results for the 

real-life data set are reported. 

■ 

WB42 


Computational Optimization and Applications II 




Chair: Ali Ekici, Assistant Professor, University of Houston, Department of 

Industrial Engineering, E221A Engineering Building 2, Houston, TX, 

77204, United States of America, aekici@Central.UH.EDU 

1 - On the Parallel Computation of Individual Penalties in 

Scheduling Jobs 

Irinel Dragan, Professor Emeritus of Mathematics, University of 

Texas at Arlington, U.T.Arlington, Mathematics, Arlington, TX, 

76019-0408, United States of America, dragan@uta.edu 

For scheduling n jobs on a single machine, with a common due date. and given 

weights, we want to find a schedule which minimizes the total penalty. We build a 

scheduling game and we propose to compute in parallel the individual penalties as 

the Shapley Value expressed by means of the Average per capita formula 

(Dragan,1992). 

2 - Experience in Developing Heuristic Algorithms to Solve Large-scale 

Non-convex NLP Problem 

Vladimir Krichtal, Senior Development Engineer, Transpower NZ, 96 

The Terrace, P.O. Box 1021, Wellington, New Zealand, 

Vladimir.Krichtal@transpower.co.nz, Conrad Edwards 

Transpower, as the New Zealand power system operator, uses an LP dispatch and 

pricing model. The model is derived from a non-convex NLP one, with quadratic 

circuit losses approximated by piece-wise linear functions. Sometimes the solver 

does not solve the resulting LP model correctly. The types of incorrect solutions are 

identified. Some of these incorrect solutions are prevented via market rule changes. 

Other can be fixed using heuristic iterative MIP algorithms. 

3 - Cutting Stock Problem with Setup Costs 

Ali Ekici, Assistant Professor, University of Houston, Department of 

Industrial Engineering, E221A Engineering Building 2, Houston, TX, 

77204, United States of America, aekici@Central.UH.EDU, 

Azadeh Mobasher 

We study the weighted cutting stock problem which is a more general version of the 

well-known cutting stock problem. The objective in weighted cutting stock problem 

is to minimize total production cost including both waste and setup costs. Since the 

problem is NP-hard, we develop heuristic algorithms to find good solutions and test 

their effectiveness on randomly generated instances.

WB44 

■ WB44 


Appointment Scheduling in Healthcare 



Chair: Serhan Ziya, University of North Carolina, 356 Hanes Hall 

CB# 3260, Chapel Hill, NC, 27599, United States of America, 

ziya@email.unc.edu 

Co-Chair: Nan Liu, Mailman School of Public Health, Columbia 

University, 600 W 168th ST, 6th Floor, New York, NY, 10032, United 

States of America, nl2320@columbia.edu 

1 - A Comparison of Traditional and Open-access Appointmentscheduling 

Policies 

Rachel Chen, University of California at Davis, Graduate School of 

Management, Davis, CA, 95616, United States of America, 

rachen@ucdavis.edu, Lawrence Robinson 

Under traditional scheduling, patients are booked in advance, but may not show up. 

Under open-access scheduling, a random number of patients call in the morning to 

make an appointment for that same day. Thus under either policy the number of 

patient arrivals will be random, for different reasons. We find that the open-access 

schedule in general outperforms the traditional schedule, except when patient 

waiting time is held in little regard or when no-show probability is small. 

2 - Adaptive Appointment Systems with Patient Preferences 

Wen-Ya Wang, University of Minnesota, 111 Church St SE, 

Minneapolis, MN, 55455, United States of America, 

wenya@ie.umn.edu, Diwakar Gupta 

We propose a framework for the design of the next generation of appointment 

systems that dynamically learn and update patients’ preferences, and use this 

information to improve booking decisions. Analytical results leading to a partial 

characterization of an optimal booking policy are presented. Examples show that 

heuristic decision rules, based on this characterization, perform well and reveal 

insights about tradeoffs among a variety of performance metrics important to clinic 

managers. 

3 - Controlling Demand for Appointment-based Services in the 

Presence of No-show Customers 

Nan Liu, Mailman School of Public Health, Columbia University, 

600 W 168th ST, 6th Floor, New York, NY, 10032, 

United States of America, nl2320@columbia.edu, Serhan Ziya 

Customer no-shows is a major problem for many service systems that work with 

appointments. In this talk, we model the scheduled appointments as a single-server 

queue and investigate how to maximize system throughput taking into account 

waiting time dependent no-shows. We also study how to adjust the system design 

in response to changes in customer no-show behavior. Both analytical and 

numerical results will be presented. 

■ WB45 


Modeling and Optimizing Quality of Care and Efficiency 

of Delivery 



Chair: Eva Lee, Professor & Director, Georgia Institute of Technology, 

Center for Operations Research in Medici, Industrial & Systems 

Engineeriing, Altanta, GA, 30332-0205, United States of America, 

eva.lee@gatech.edu 

1 - Quantifying Reductions in Variability of Intraoperative Time From 

Meta-analysis of Trial Results 

Franklin Dexter, Professor, University of Iowa, Department of 

Anesthesia, 6JCP, Iowa City, IA, 52242, United States of America, 

franklin-dexter@uiowa.edu 

Electronic medical record data were used to facilitate meta-analysis of randomized 

clinical trials. For interventions affecting anesthetic durations, usually there are 

many (> 20) clinical trials each with small N (

4 - Influenza Vaccine Supply Chain: Role of Consumption Externality 

and Yield Uncertainty 

Kenan Arifoglu, PhD Student, Northwestern University, 2145 

Sheridan Road, Room C 210, Evanston, IL, 60208, United States of 

America, kenanarifoglu2011@u.northwestern.edu, Sarang Deo, 

Seyed Iravani 

We consider the inefficiency in flu vaccine supply chain and study the impact of two 

critical factors: yield uncertainty (supply side) and self-interested consumers 

(demand side). Contrary to previous economic models, we find that consumers may 

demand more vaccinations than is socially optimal when they jointly consider the 

availability and infection externalities. We study two partially-centralized supply 

chains to investigate the benefits of government interventions on demand and 

supply side. 

■ WB47 


Project Scheduling and Risk Management 



Chair: Richard Wendell, Professor, Katz Graduate School of Business, 

University of Pittsburgh, Pittsburgh, PA, 15260, United States of America, 

wendell@katz.pitt.edu 

Co-Chair: Timothy Lowe, Professor, Tippie College of Business, 

University of Iowa, Iowa City, IA, United States of America, 

timothy-lowe@uiowa.edu 

1 - New Product Design and Pricing in a Duopoly Market with 

Forward-looking Consumers 

Ted Klastorin, Burlington Northern/Burlington Resources Professor, 

Department of Information Systems & Operations Mgt, Michael G 

Foster School of Business, University of Washington, Seattle, WA, 

98195-3200, United States of America, tedk@u.washington.edu, 

Aysun Ozler, Yong-Pin Zhou 

We study the introduction of a new product into a durable goods duopoly market. 

In our model, an innovator firm begins to develop a product; after a random time 

period, information about this product leaks and an imitator firm begins to develop 

a competing product. Consumers are forward looking and make purchases based on 

the available product and their expectations on future products. We derive 

implications for both profit-maximizing firms with respect to the design and pricing 

of the products. 

2 - Planning Stochastic Projects Under the Threat of a Disruptive Event 

Gary Mitchell, Assistant Professor, Pamplin School of Business 

Administration, University of Portland, 5000 N. Willamette Blvd, 

Portland, OR, 97035, United States of America, mitchelg@up.edu, 

Ted Klastorin, Issariya Sirichakwal 

We consider the issue of planning a stochastic project when there is a threat of an 

exogenous disruptive event that may stop work on one or more tasks (possibly the 

entire project). Our goal is to minimize expected total project costs, including 

resource costs, overhead/indirect costs, and penalty costs. We analyze a model to 

determine when the project manager should take proactive steps during the 

planning phase or wait and take contingent actions after a disruption occurs. 

3 - Agility in Projects - Theoretical and Computational Results 

Karolina Glowacka, Assistant Professor, Stevens Institute of 

Technology, Howe School of Technology Management, Hoboken, NJ, 

07030, United States of America, kglowack@stevens.edu, Richard 

Wendell, Timothy Lowe 

In this research we show that a lack of agility can be a significant factor in project 

delays. We characterize the concept of agility in projects, show how agility can have 

a significant impact on the likelihood of achieving a target-time for a project, and 

give some general properties on agility with respect to a project’s structure. We also 

discuss how and where to build agility into key project activities. 

4 - Integration of Project Management and Software Development 

Processes in a Complex Project 

Laura Anderson, Manager, Advanced Estimating & Infrastructure 

Solutions, IBM Research - Almaden, 650 Harry Road, San Jose, CA, 

95120, United States of America, laurac@almaden.ibm.com, 

Ruoyi Zhou 

Project management continues to increase in complexity, with special challenges 

due to agile methodologies and geographically dispersed teams. We describe our 

experience in software project management for a medium sized research and 

development project using IBM’s Rational Jazz. We discuss the observed 

advantages of Jazz in systematizing the overall process, with narrative observations, 

quantitative measurements, and objective measures of the value of such a project 

management system. 


401 

■ WB49 


Systems Analysis and Design 



Chair: Vijay Khatri, Indiana University, 1309 E. 10th Street, BU 560F, 

Bloomington, IN, 47405, United States of America, vkhatri@indiana.edu 

Co-Chair: Jeffrey Parsons, jeffreyp@mun.ca 

1 - A Cognitive Perspective on Developing and Interpreting 

Conceptual Models 

Palash Bera, Assistant Professor, Texas A&M InternationalUniversity, 

5201, University Blvd., Laredo, TX, 78041, United States of America, 

palash.bera@tamiu.edu 

Conceptual models are often used to document features of the domain to be 

reflected in the Information Systems. Developing and interpreting conceptual 

models can be considered as cognitive tasks performed by modelers. In a laboratory 

setting, this study contrasts the cognitive difficulties faced by two groups of 

modelers- one who develops the models and the other who interprets the same 

models. The results indicate that the cognitive difficulties faced by the groups are of 

different nature. 

2 - Business Informational Sabotage: An Exploration into Incidence 

Rates and Causes 

John Erickson, University of Nebraska at Omaha, College of 

Business, 6001 Dodge Street, Omaha, NE, 68182, United States of 

America, johnerickson@mail.unomaha.edu, George Gresham, 

John Hafer 

This study represents an exploratory effort into the occurrence rates of nineteen 

types of organizational information sabotage, moral attitudes towards the acts and 

the saboteurs. Roughly 40% of the respondents indicated awareness of sabotage and 

up to 31% admitted committing one/some acts of sabotage. Demographic variables 

were tested for significance. 

3 - Experiments in Paired Software Development 

Radha Mahapatra, Associate Professor, University of Texas at 

Arlington, 701 S. West St., Arlington, TX, 76019-0437, 

United States of America, mahapatra@uta.edu, Sridhar Nerur, 

VenuGopal Balijepally, George Mangalaraj 

A core practice in recently popularized agile software development methods is 

paired development, where two developers work together to jointly design and 

develop application systems. We have conducted a series of experiments to 

understand the efficacy of paired development vis-á-vis the traditional practice of 

software development by developers working individually. Findings from these 

experiments will be presented. 

■ WB50 


Information Systems I 


Chair: Zsolt Ugray, Associate Professor, USU, 3515 Old Main Hill, Logan, 

United States of America, Zsolt.Ugray@usu.edu 

1 - Bundling of Information Sources using the Value of Information 

Pantea Alirezazadeh, PhD Student, University of Connecticut, 2100 

Hillside rd, Unit 1041, Storrs, CT, 06269, United States of America, 

pantea.alirezazadeh@business.uconn.edu, Fidan Boylu, Ram Gopal 

We present a method for acquiring information from multiple information sources 

with different reliabilities based on the incremental value of information. We 

evaluate the value of information that can be provided by third-party data providers 

and measure the value of different combinations of information sources based on 

their contribution to expected utility and provide heuristic solutions to select an 

optimal combination of these information sources. 

2 - Understanding Time Inconsistent Preferences in Real Options 

Based it Investment Justification 

Ram Kumar, UNC- Charlotte, 9201 University City Blvd, Charlotte, 

NC, United States of America, rlkumar@uncc.edu, 

Sarah Khan 

The literature on the use of Real Options in IT (and other) investment justification 

assumes that decison makers will be able to exercise options optimallly. However, in 

practice, decision makers can have time inconsistent preferences. The effects of 

these preferences on the use of Real Options Analysis is analyzed. This analysis has 

important implications for decison makers.

WB51 

3 - Multicriteria Model for Selecting Information Systems Based on the 

PROMETHEE Method 

Jônatas Almeida, Federal University of Pernambuco, Recife-PE, 

Brazil, jonatasaa@yahoo.com.br, Ana Paula Costa, Adiel Almeida 

This paper discusses the importance of the process of selecting Information Systems 

(ISs) integrated with an IS planning methodology. A model which integrates an 

adapted version of an IS planning methodology called Business System Planning 

(BSP) and the multicriteria method PROMETHEE II is presented. As a contribution, 

this paper identifies the need for ISs to start from the company’s strategic vision, 

including how value is to be added to the business, considering aspects for the 

organization. 

4 - Franchisee Attitudes Toward the Adoption of Advanced Internet 

Technologies and Innovations 

Zsolt Ugray, Associate Professor, USU, 3515 Old Main Hill, Logan, 

United States of America, Zsolt.Ugray@usu.edu, Kelley O’Reilly 

We present findings from a case study exploring franchisees’ attitudes and 

perceptions of using advanced Internet technologies and innovations. We illustrate 

that the dual role of owner-operators, in which they are both decision makers and 

users of technologies, provides a dichotomy of perspectives that yield insights into 

aspects of business leadership, customer service, and operational proficiency. 

■ WB51 


Analyses for Rotary Wing Aircraft 



Chair: Caolionn O’Connell, Institute for Defense Analyses, 4850 Mark 

Center Drive, Alexandria, VA, 22311, United States of America, 

coconnel@ida.org 

1 - Rotary Wing Aircraft Capacity for Operations in Afghanistan 

Matthew Grund, Research Analyst, CNA, 4825 Mark Center Drive, 

Alexandria, VA, 22311, United States of America, grundm@cna.org, 

Greg Cox 

Senior leaders in the Department of Defense have expressed concern that there is 

insufficient rotary wing capacity in Afghanistan to effectively carry out the required 

missions. Our work broadly explores three ways of increasing rotary wing capacity 

in Afghanistan: increasing the number of aircraft in theater, increasing the hours 

flown by aircraft already in theater, and increasing the efficiency of flight hours 

already being flown by aircraft in Afghanistan. 

2 - Utility Assessment of Rotary Wing Aircraft 

Dana Paterson, Senior Analyst, Naval Air Systems Command, 

Bldg 2109, Room S211, Patuxent River, MD, 20670, 

United States of America, dana.paterson@navy.mil 

Decision makers faced with making a choice among a finite set of rotary-wing 

aircraft, aircraft designs or aircraft concepts must have information products 

describing the utility of each member of the set. This briefing is a description of a 

senior analyst’s method beginning with identification of the decision to be made and 

ending with an exposition of his preferred method for describing the decision space 

for the decision maker. 

3 - A Forensic Analysis of Cost Growth in Rotary Wing Programs 

Caolionn O’Connell, Institute for Defense Analyses, 4850 Mark 

Center Drive, Alexandria, VA, 22311, United States of America, 

coconnel@ida.org 

An analysis of cost growth in rotary wing (RW) programs and how it compares to 

other major defense acquisition programs (MDAPs). The brief will also include a 

discussion of possible root causes for RW cost growth and a suggested cost 

estimating methodology for future programs. 

4 - Army Aviation Force Structure 

Steven Stoddard, US Army, Burke, VA, United States of America, 

steven.stoddard@us.army.mil 

The Army has the responsibility to provide manned, trained, and equipped units to 

support operations. We explain how the Army manages the supply of helicopter 

units to meet as much demand as possible. We answer three questions: 1. 

What happens if the Army adds more helicopters? 2. What happens if the Army 

uses its helicopter units more often? 3. What happens if the Army alters its active - 

reserve mix? 


402 

■ WB52 


Assessing Terrorism Probabilities and All-Hazards Risk 

in the U.S. Department of Homeland Security 

Cluster: Homeland Security and Defense 


Chair: Steve Bennett, Asst. Director, Risk Analytics Division, Office of 

Risk Management and Analysis, U.S. Department of Homeland Security, 

Washington, DC, 20528, United States of America, steve.bennett@dhs.gov 

1 - Relative Probabilities of Terrorist Attacks: Rational Adversaries with 

Uncertain Value Tradeoffs 

Evan Levine, DHS Office of Risk Management and Analysis, 

2929 Connecticut Ave NW, Apt 709, Washington, DC, 20008, 

United States of America, evan.levine@dhs.gov 

Many analyses conducted to inform counterterrorism decisions depend on estimates 

of the relative probabilities of different attack types. We describe a method of using 

uncertainty in utility function value tradeoffs to model the adversary’s decision 

process and solve for the relative probabilities of attacks in closed form. The process 

we describe is an extension of value-focused thinking, and is suitable for application 

outside of counterterrorism, including general business decision-making. 

2 - Informing All-hazards Decision-making at the U.S. DHS: 

Constructing an Appropriate Scenario Set 

Julie Waters, Risk Analyst, Office of Risk Management and Analysis, 

U.S. Department of Homeland Security, Washington, DC, 20528, 

United States of America, julie.waters@dhs.gov, Evan Levine, 

Steve Bennett 

The Homeland Security National Risk Assessment (HSNRA) methodology, developed 

by DHS, uses an order-of-magnitude estimation technique to quantify the dominant 

risks across the disparate hazards confronted by the Nation. In this presentation, we 

discuss how the space of possible events relevant to homeland security is discretized 

into scenarios that form the units of analysis for frequency and consequence 

elicitation and the process by which dominant scenarios are identified. 

3 - Eliciting Probabilities From the Intel Community to Support 

Terrorism Risk Assessments at U.S. DHS 

Natasha Hawkins, Risk Analyst, U.S. Department of Homeland 

Security, Washington, DC, 20528, United States of America, 

natasha.hawkins@dhs.gov, Tony Cheesebrough, Steve Bennett 

Probability elicitations for terrorism involve engaging intelligence community 

experts who generally communicate threat judgments in qualitative terms. This 

presentation will discuss challenges and opportunities in applying expert elicitation 

methods to the elicitation of intelligence information in DHS and provide an 

opportunity for the OR/MS community to provide input and suggestions for 

improvement and enhancement. 

4 - RAPID: Supporting Risk-informed Strategic Policy and Resource 

Allocation Decisions at the U.S. DHS 

Debra Elkins, Section Chief, Risk Assessments and Analysis, U.S. 

Department of Homeland Security, U.S. Department of Homeland 

Security, Washington, DC, 20528, United States of America, 

debra.elkins@dhs.gov, Steve Bennett, Tony Cheesebrough, 

Natasha Hawkins 

The Risk Assessment Process for Informed Decision-making (RAPID) is a risk 

assessment conducted within the DHS Office of Risk Management and Analysis that 

supports strategic policy and budgetary decision-making across the U.S. Department 

of Homeland Security. This presentation will describe the technical risk analysis 

methodologies used in RAPID and provide an opportunity for the OR/MS 

community to provide input and suggestions for improvement and enhancement.

■ WB53 


Operations/Marketing Interface I 


Chair: Baokun Li, Dr., Southwestern University of Finance and 

Economics, 555 Liutai Ave, Wenjiang District, Chengdu, 611130, China, 

flyinghorse1967@yahoo.com 

1 - Market-based Joint Decisions on Price, Delivery Time, Service 

Level, and Supplier Selection 

Li Qian, South Dakota State University, Solberg Hall 115B, 

Brookings, 57006, United States of America, li.qian@sdstate.edu 

This paper models the demand as a linear or log-linear function of attributes 

including price, guaranteed delivery time, and service level. With stochastic delivery 

time, the service level is not always binding at the minimal value reserved by the 

manager or the market, as assumed in most literature. A market-oriented approach 

for supplier selection or investment is proposed in consideration of operation 

performances in cost, delivery time, service level, and/or quality in an integrated 

manner. 

2 - Cooperative Advertising in a Dynamic Durable Goods Supply Chain 

Anshuman Chutani, UT Dallas, School of Management, SM 30, 

Richardson, TX, 75080, United States of America, 

anshuman.chutani@student.utdallas.edu, Suresh Sethi 

We analyze dynamic advertising and pricing policies in a durable goods supply chain 

with co-operative advertising. We consider a stackelberg game where manufacturer 

announces the wholesale price and its share in the retailer’s advertising 

expenditure. The retailer responds with its optimal advertising and pricing policies. 

The sales dynamics follows the model suggested by Sethi, Prasad and He (2008). We 

analyze two different demand specifications, linear and iso-elastic. 

3 - Can We All Get Along? Incentive Contracts to Bridge the Marketing 

and Operations Divide 

Kinshuk Jerath, Assistant Professor, Carnegie Mellon University, 

5000 Forbes Ave, Pittsburgh, PA, 15213, United States of America, 

kinshuk@cmu.edu 

The marketing and operations management arms in a firm must work in 

coordination. However, a major source of conflict is that marketing compensation is 

usually heavily weighted towards sales whereas operations compensation is usually 

heavily weighted towards expense reduction. In this paper, we invoke agency 

theory to determine compensation plans for sales and operations managers to 

coordinate their activities in the best interests of the firm. 

4 - Computing Reasonable Allocations in Coalition Games 

Baokun Li, Dr., Southwestern University of Finance and Economics, 

555 Liutai Ave, Wenjiang District, Chengdu, 611130, China, 

flyinghorse1967@yahoo.com 

We introduce a new kind of coalition gain allocation, Proportion Allocation, which 

is inspired from the Coarsening at Random mechanism in statistics. Based on the 

Proportion Allocation and traditional Shapley value, a reasonable class of allocations 

is dened and related algorithm is given for calculating them. Besides, a simulated 

annealing method is constructed to solve for the stable allocation, nucleolus, for 

coalition games. 

■ WB54 


Characterizing Uncertainty and Learning for 

Technology Choice Decisions 



Chair: Thomas Rand-Nash, Doctoral Candidate, Massachusetts Institute of 

Technology Materials Systems Laboratory, 77 Massachusetts Avenue, 

Cambridge, MA, 02139, United States of America, trand@mit.edu 

1 - Learning-Derived Cost Evolution in Materials Selection 

Trisha Montalbo, MIT, 77 Massachusetts Ave, E38-420, Cambridge, 

United States of America, trisha@mit.edu, Richard Roth 

We investigate the impact of considering cost evolution due to learning by doing in 

the selection of materials for a manufacturing firm. A multi-product, multi-period 

selection framework is developed to analyze the problem because single-product 

selection methods are unable account for benefits realized through shared learning 

among products. The use of test beds as a strategy for introducing new materials to 

a firm is also evaluated. 


403 

2 - Cost-Optimal Reinspection Plans 

Hadi Zaklouta, Graduate Student Researcher, MIT Materials Systems 

Laboratory, 292 Main Street, E38-435, Cambridge, MA, 02139, 

United States of America, zaklouta@mit.edu, Randolph Kirchain, 

Richard Roth 

This paper compares single and double inspection plans effect on cost and quality. A 

cost model is proposed that accounts for appraisal and internal/external failure 

costs. All units are tested with one or two tests. In plan RR, rejects are retested and 

replaced after rejection. In plan AR, accepts are retested and rejects replaced. Nonconformance 

rate and test error rates are known. The AR plan is best for high 

warranty/scrap cost ratio; the RR plan for low; a single test plan otherwise. 

3 - Learning as a Driver Technology Choice Decisions in Manufacturing 

Thomas Rand-Nash, Doctoral Candidate, Massachusetts Institute of 

Technology Materials Systems Laboratory, 77 Massachusetts Avenue, 

Cambridge, MA, 02139, United States of America, trand@mit.edu 

This work explores process technology decision making in the presence of learning 

effects in manufacturing, and hopes to characterize the conditions under which 

learning-related production cost effects impact technology choice decision making. 

Relevant factors considered include learning rates, production volume as a function 

of demand, market structure, and budget constraints. 

4 - Shifting Grounds: How Industry Emergence Changes the 

Effectiveness of Knowledge Creation Strategies 

Sebastian Fixson, Babson College, Babson Park, MA, 

United States of America, sfixson@babson.edu, Won Hee Lee 

The knowledge management literature has identified various aspects, advantageous 

and disadvantageous, of both inward-looking and outward-looking knowledge 

creation strategies. With a longitudinal empirical study we explore the dynamics of 

firms’ knowledge creation strategies during the period of industry emergence. We 

find that the emergence of an industry changes the effectiveness of the different 

knowledge creation strategies. 

■ WB55 


Sustainability and NPD 



Chair: Cheryl Druehl, Assistant Professor, George Mason University, 

4400 University Dr, MS 5F4, Fairfax, VA, 22030, United States of 

America, cdruehl@gmu.edu 

1 - New Business Models to Enable Clean and Renewable Generation 

in the Electric Power Industry 

Edward Anderson, Associate Professor, University of Texas, 

1 University Station B6500, Austin, TX, 78733, United States of 

America, Edward.Anderson@mccombs.utexas.edu, Geoffrey Parker 

Over the coming decades, electric power companies must transform their business 

models to accommodate the smart grid and growing demand for clean renewable 

energy. Many researchers are examining aspects of this problem, such as power 

companies’ infrastructure portfolios, consumer behavior, pricing structures, etc. To 

complement this analysis, we build and analyze a top-down systems model of a 

typical power company and its market “ecosystem” using the system dynamics 

methodology. 

2 - Does “To Go Green” Translate into Profitability? 

Asoo Vakharia, Professor, University of Florida, Department of ISOM, 

Gainesville, FL, 32611-7169, United States of America, 

asoo.vakharia@warrington.ufl.edu, Arda Yenipazarli 

Consumers are frequently integrating “green” product attributes when making 

purchasing decisions. We propose a firm level analytical model to enable decisions 

on whether to upgrade the “green” content of an existing product, replace the 

existing product with a “green” product, or provide a portfolio of an existing and 

“green” product. 

3 - Reviving the Electric Car Movement: Developing Green 

Infrastructure for Sustainable Transportation 

Cheryl Druehl, Assistant Professor, George Mason University, 4400 

University Dr, MS 5F4, Fairfax, VA, 22030, United States of America, 

cdruehl@gmu.edu, Michael Naor 

An innovative new business model for sustainable transportation is described. A 

case study approach is used to study the unique product and service development 

process required to electrify the automobile industry, focusing on a clean tech 

company building innovative green infrastructure for sustainable transportation in 

Israel.

WB56 

■ WB56 


Human Decision-Making and Social 

Network Simulation 

Sponsor: Simulation Society 


Chair: Young-Jun Son, Professor, The University of Arizona, Systems and 

Industrial Engineering, Tucson, AZ, 85721, United States of America, 

son@sie.arizona.edu 

1 - Hyper-Networks and Their Properties 

W. K. Victor Chan, Assistant Professor, Rensselaer Polytechnic 

Institute, ISE Dept, CII 5015, 110 8th Street, Troy, NY, 12180-3590, 

United States of America, chanw@rpi.edu, Cheng Hsu 

A hyper-network is an integration of multi-layered (role-based) connections of 

members in a community, such as the Internet and an ecosystem. In this talk, we 

formally define hyper-networks and present their analytical properties. We show 

that hyper-networks can reveal otherwise hidden social structures and provide 

estimation formulae for determining average vertex-vertex distances and average 

vertex degrees. 

2 - Simulation-based Workforce Assignment Considering Position in a 

Social Network 

Nurcin Celik, PhD, The University of Arizona, 1127 E James E. 

Rogers Way, Tucson, AZ, 85721-0020, United States of America, 

nurcinkoyuncu@gmail.com, Dong Xu, Hui Xi, Young-Jun Son, 

Robin Lemaire, Keith Provan 

A novel modeling framework is proposed to help managers devise optimal 

workforce assignments that consider both short and long-term aspects of the 

organizational social network. The framework involves the evaluation module (via 

agent-based simulation) to calculate the position and equivalence values between 

each pair of workforce members and the assignment module (via multi-objective 

optimization) to select an optimal workforce mix. The framework is illustrated with 

the Kuali organization. 

3 - Dynamic Control and Simulation of Manufacturing Process with 

Different Forms of Uncertainties 

Hyunsoo Lee, Texas A&M University, 3131 TAMUS, College Station, 

TX, 77843, United States of America, hsl@neo.tamu.edu, 

Amarnath Banerjee 

We show a method for capturing and dynamically controlling uncertainties under 

changes in manufacturing condition. Ambiguity and variance type uncertainties are 

incorporated into a Fuzzy colored Petri Net with stochastic time delay model and it 

is controlled using a new and effective learning method for simulation-based 

optimization. A semiconductor process with serial and batch machines is controlled 

and a given objective is achieved using the suggested simulation based optimization 

method. 

■ WB57 


Financial Optimization in Energy and 

Communication Systems 

Sponsor: Financial Services Section 


Chair: Jim Bander, Credit Risk Optimization Manager, 

Toyota Financial Services, 3200 W Ray Road, Chandler, AZ, 85226, 

United States of America, jim.bander@gmail.com 

1 - Energy Portfolio Investment with Entry Decisions 

Zhen Liu, Assistant Professor, Missouri University of Science & 

Technology, United States of America, zliu@mst.edu, Scott Grasman, 

Jianjun Deng 

We formulate energy portfolio problems as an optimization problem to maximize 

long-term profit through stochastic control and numerical analysis methods, and 

solve the following problems: (1) the optimal time to build a new alternative green 

energy power generating plant, and (2) the optimal dispatch from the existing coal 

plant and the new plant. 

2 - Reducing Price Volatility of Electricity Consumption for a Firm’s 

Energy Risk Management 

Xiaohua Wu, Rensselaer Polytechnic Institute, 903 Peoples Ave 

Apt 3, Troy, NY, 12180, United States of America, wux4@rpi.edu, 

Aparna Gupta 

Smart Grid technologies allow firms to manage electricity price volatility and 

fluctuations. We develop a framework to assess strategies for optimally shifting peak 

load consumption using distributed storage systems of Smart Grids. Risk 

management is achieved by optimal investment in storage systems and peak load 

shifting under stochastic electricity prices. 


404 

3 - A Regime-switching Approach to the Valuation of Weather Options 

M.I.M Wahab, Ryerson University, Toronto, ON, M5B 2K3, Canada, 

wahab@ryerson.ca, R. S. Elias, L. Fang 

Regime-switching processes are used to model the stochastic behavior of 

temperature with aim of valuation of temperature-based weather options. Three 

different models are used to predict the expected heating degree days (HDD) and 

cooling degree days (CDD) that play a crucial role in valuation of weather options. 

Temperature dataset from Toronto, Canada, is used for the analysis. Results 

demonstrate that one of the models captures temperature dynamics more accurately 

than the other two models. 

4 - Path-Vector Contracting: Profit Maximization and Risk Management 

Praveen Kumar, Rensselaer Polytechnic Inst, 110 8th Street, 

Troy, NY, United States of America, muthup@rpi.edu 

We consider an Internet Service Provider’s problem of providing end-to-end (e2e) 

services with bandwidth guarantees, using a path-vector based approach. The spotpricing 

framework for e2e bandwidth guaranteed services utilizes a path contracting 

strategy by formulating it as a stochastic optimization problem with the objective of 

maximizing expected profit subject to risk constraints. 

5 - Flash Crashes - Volatility, Smart Order Routing, and 

Market Fragmentation 

Bruce Weber, London Business School, Regent’s Park, London, 

NW14SA, United Kingdom, bweber@london.edu 

Advances in IT have lowered costs but also fostered trading practices that may 

negatively impact the entire market. A national market system for securities with 

multiple fragmented liquidity pools is simulated. We analyze how trading rules at 

different exchanges and software for smart order routing can lead to illiquidity and 

“flash crashes.” Under certain conditions, high frequency trading can lead to 

securities market aberrations, and curbs on computer-driven trading may be 

warranted. 

■ WB58 


Finance-Theory and Empirics 


Chair: Vinod Cheriyan, Georgia Institute of Technology, Industrial and 

Sytems Engineering, 765 Ferst Drive NW, Atlanta, Georgia, 

vinod.cheriyan@gatech.edu 

1 - The Empirical Research of Media Effect in China Stock Market 

Yahui Zhang, The School of Management, Xi’an Jiaotong University, 

#28 Xianning West Road, Mailbox 1875, Xi’an, China, 

kailey@stu.xjtu.edu.cn, Leiming Fu, Difang Wan 

This research explores the existence of media effect in China stock market through 

event research and analyses the affecting factors. Results show that media effect is 

significant and presented as negative cumulative abnormal return within the event 

window, which can be mitigated by better corporate governance. CAR is positively 

correlated with the B/M ratio of equity, financial leverage, ROA and position 

accumulation significantly, the affect of firm size and news type are not significant. 

2 - MIDAS Instruments for Multiple Parameters 

Stephen Goldberger, PhD candidate, UNC, University of North 

Carolina, Chapel Hill, NC, 27514, United States of America, 

sgoldber@email.unc.edu 

Many Time Series models in Econometrics are dependent on the condition that an 

error term is expected to be zero given all information available at the beginning of 

time t. I extend the MIDAS (Mixed Data Sampling) Instruments framework 

developed by Eric Ghysels and Johnathan Wright to create a GMM estimator for a 

vector of parameters dependent on this moment condition. 

3 - Can Oil Prices Help Estimate Commodity Futures Prices? The 

Cases of Copper and Silver 

Gonzalo Cortazar, P. Universidad Catolica de Chile, Vicuna 

Mackenna 4860, Santiago, Chile, gcortaza@ing.puc.cl, 

Francisco Eterovic 

We use prices of long term oil futures contracts to estimate copper and silver prices. 

We show that the Cortazar et al (2008) multi-commodity model applied to oilcopper 

and oil-silver which have low correlation seems not to be effective. We then 

propose a modified multi-commodity model that uses the non-stationary long term 

process of oil to help estimate long term copper and silver futures prices, achieving a 

much better fit than using available individual or multi-commodity models.

4 - A Bounded-rational Model of Price Bubbles and Business Cycles 

Vinod Cheriyan, Georgia Institute of Technology, Industrial and 

Sytems Engineering, 765 Ferst Drive NW, Atlanta, GA, 

vinod.cheriyan@gatech.edu, Anton Kleywegt 

Various markets exhibit growth and collapse in prices that are sometimes called 

bubbles. Related to that is the notion of business cycles. We consider a model in 

which investors behave reasonably, although with imperfect expectations, that 

attempt to provide insight into the formation of bubbles and business cycles. We 

also present results on the convergence of the process to an attractor that describes 

a business cycle, and consider where in the cycle most time is spent. 

■ WB61 


Operations Management VI 


Chair: Andrew Kach, Doctoral Student, New Mexico State University, 

Department of Management, P.O. Box 3001, Las Cruces, NM, 88003, 

United States of America, akach@nmsu.edu 

1 - Advance Demand Information, Capacity Restrictions and 

Customer Prioritization 

Bisheng Du, PhD Student, Aarhus University, Fuglesangs Alle 4, 

Department of Business Studies, Aarhus, 8210, Denmark, 

bisd@asb.dk, Christian Larsen, Alan Scheller-Wolf 

We study a single supplier with fixed capacity selling products to buyers having 

different priorities. The buyers can place pre-orders before their demand is observed, 

and can also issue additional orders upon observing updating demand information. 

Since the supplier guarantees delivery of pre-ordered goods (these are not 

constrained by her capacity) buyers with lower priorities may consider pre-ordering 

in order to secure inventory. We find optimal policies for the supplier and buyers. 

2 - Centrality and Heterogeneity During the Evolution of 

Inter-organisational Networks 

Nuno Oliveira, PhD Student, LSE, Houghton Street, London, 

United Kingdom, n.r.oliveira@lse.ac.uk 

Although extant research has reported a linkage between network centrality and 

heterogeneity, the understanding on the co-evolution of both variables is little. For a 

inter-organisatitional network of a 48 million British pounds building project, we 

show that network heterogeneity decreases whilst network centrality has a U-shape 

throughout a 3-year period. Implications for researchers and practitioners are also 

presented. 

3 - Security as a Moderator of Stress on Airline Performance 

Andrew Kach, Doctoral Student, New Mexico State University, 

Department of Management, P.O. Box 3001, Las Cruces, NM, 88003, 

United States of America, akach@nmsu.edu, Jeffrey Teich 

Heightened security in airports is used as a preventative action to reduce passenger 

engagement in terrorist acts or disruptive behavior; however, high safety measures 

may have an adverse impact on airline performance. More specifically: How does 

increasing levels of security impact the relationship between passenger stress levels 

on airline performance? 

■ WB62 


Aviation Applications IV 


Chair: Ioannis Simaiakis, PhD Candidate, Massachusetts Institute of 

Technology, 77 Massachusetts Avenue, Cambridge, MA, 02139, United 

States of America, ioa_sim@mit.edu 

1 - Future Aircraft Network and Schedules 

Yan Shu, Graduate Student, Georgia Institute of Technology, 686 

Cherry Street, Atlanta, 30332, United States of America, 

yshu@gatech.edu, Ellis Johnson, Barry Smith, John-Paul Clarke 

We propose a three-step approach to build flight schedule from scratch. We 

implement our algorithms into solutions. 

2 - Optimizing Staffing Plans at Airports 

Prem Kumar Viswanathan, Scientist, Transport and Mobility 

Laboratory, Ecole Polytechnique Federale de Lausanne, GC B3 435, 

EPFL, Ecublens, VD, 1015, Switzerland, prem.viswanathan@epfl.ch, 

Michel Bierlaire 

Minimizing operating costs for maintaining ground personnel at airports is a 

complex problem due to uneven flight activities, passenger service expectations and 

staffing inflexibilities due to shift durations. In this work, we develop a method to 

find optimal shift timings that considers non-productive time due to activity 

changeovers, the mix of full-time and part-time workers and passenger waiting time 

criteria. 


405 

3 - A Colored Stochastic Petri Net Based Approach to Performance 

Analysis of JFK International Airport 

Poornima Balakrishna, Sensis Corporation, 11111 Sunset Hills Rd, 

Suite 130, Reston, VA, 20190, United States of America, 

pbalakri@sensis.com 

We model the airport departure process using colored stochastic Petri nets and 

airport surveillance data. This formal specification of the probabilistic airport system 

is then analyzed through simulation. The use of surveillance data in the model 

enables identification of bottlenecks on the airport surface including gate area and 

taxiway intersections. We measure congestion through analysis of queue lengths, 

delays and resource utilization and report on both airline and airport performance. 

4 - Estimation of Airport Performance Metrics 

Ioannis Simaiakis, PhD Candidate, Massachusetts Institute of 

Technology, 77 Massachusetts Avenue, Cambridge, MA, 02139, 

United States of America, ioa_sim@mit.edu, Hamsa Balakrishnan 

Operational throughput and taxi times are two key metrics of airport performance. 

In this work, we show how the maximum throughput capacity of an airport can be 

represented as a function of arrival and departure demand. We also illustrate how 

unimpeded taxi times may be estimated by representing taxi time as a function of 

takeoff and the queues. We use convex-optimization to estimate these metrics 

without assuming the form of the solutions, and by only imposing operational 

constraints. 

■ WB63 


Game Theory and Homeland Security 

Sponsor: Decision Analysis 


Chair: Jun Zhuang, Assistant Professor, University at Buffalo, SUNY, 403 


jzhuang@buffalo.edu 

1 - Mixed Strategy Nash Equilibria in Symmetric Signaling Games 

Barry Cobb, Virginia Military Institute, 335 Scott Shipp Hall, 

Lexington, United States of America, cobbbr@vmi.edu, Atin 

Basuchoudhary, Gregory Hartman 

Signaling games are characterized by asymmetric information where the more 

informed player has a choice about what information to provide to its opponent. 

Decision trees are used to derive the Nash equilibrium strategies for signaling games 

where the more informed player has an arbitrary number of possible types. The 

technique is demonstrated by analyzing an interactive game between a terrorist and 

a governmental agency. 

2 - Solving Massive Scale Security Games with Arbitrary Schedules: 

A Branch and Price Approach 

Milind Tambe, University of Southern California, 3737 Watt Way, 

PHE 410, Los Angeles, CA, 90089, United States of America, 

tambe@usc.edu, Manish Jain, Fernando Ordonez, Chris Kiekintveld, 

Erim Kardes 

Security games are used in deployed decision-support tools in use by LAX police 

and the Federal Air Marshals Service. The algorithms used to solve these games find 

optimal randomized schedules to allocate security resources for infrastructure 

protection. Unfortunately, the state of the art algorithms fail to scale to large 

problems with arbitrary scheduling constraints. We introduce ASPEN, a branch-andprice 

approach that overcomes this limitation. 

3 - Quantifying Unobserved Attributes in Expert Elicitation of 

Terrorist Preferences 

Chen Wang, University of Wisconsin-Madison, 3237 Mechanical 

Engineering, 1513 University Avenue, Madison, WI, 53706, 

United States of America, cwang37@wisc.edu, Vicki Bier 

A terrorist values targets according to a multi-attribute utility function in which 

some attributes are unknown to the defender. A group of experts ranks potential 

targets by their attractiveness to the terrorist. The defender infers the weights of the 

known attributes, and the importance of the unobserved attributes, using either 

probabilistic inversion or Bayesian density estimation. 

4 - Modeling Arbitrary Layers of Continuous Level Defenses in Facing 

with A Strategic Attacker 

Jun Zhuang, Assistant Professor, University at Buffalo, SUNY, 403 


jzhuang@buffalo.edu, Mohsen Golalikhani 

We propose a class of models for the optimal assignment of defensive resources in a 

game between a defender and an attacker. The novelty of our model is that we 

allow the defender to assign her continuous-level defensive resources to any subset 

(or arbitrary layers) of targets due to functional similarity or geographical proximity. 

The results show that our model could significantly increase the defender’s payoff, 

especially when the cost of defense is high, or the attack cost is intermediate.

WB64 

■ WB64 


Health Care, Modeling and Optimization VI 


Chair: Song Chew, Assistant Professor, Southern Illinois University 

Edwardsville, Edwardsville, Edwardsville, IL, 62026, United States of 

America, schew@siue.edu 

1 - Simulating Clinics: The Challenge of Data Collection and Analysis 

Helida Dodd, President, Dodd Consulting Group, 10223 SW 89 St, 

Miami, FL, 33176, United States of America, helida@doddcg.com, 

Martha Centeno 

We discuss a simulation study of an Endoscopy center, challenges, and lessons 

learned. The goal of the project was to find ways to increase throughput to 80 

patients/day. It was quickly apparent that the stumbling block would be the lack of 

data available. Some data is collected via the information system, but there is no 

reporting application to measure performance. Data is also collected manually 

sporadically in different areas, but it is not readily analyzed because of the lack of 

manpower. 

2 - Simulation-based Optimal Staffing Policies Under Cyclic Demand 

Mina Loghavi, University of Tennessee, 331 SMC, 916 Volunteer 

Blvd., Knoxville, TN, 37996, United States of America, 

mloghavi@utk.edu, Robin Clark, Charles Noon, Bogdan Bichescu 

We consider the objectives of minimizing staffing cost and maintaining acceptable 

patient waiting times in an emergency department (ED). We employ simulation and 

optimization to explore the best policies for ED staffing over repeating cycles of 

stochastic demand. In contrast to SIPP-based methods, this approach is feasible in 

periods when demand temporarily exceeds capacity and when additional constraints 

are present. We present insights from computational analysis on real-world datasets. 

3 - Setting Staffing Levels Under Time-varying Demand in the Context 

of an Emergency Department 

Mieke Defraeye, PhD Student, K.U.Leuven, Naamsestraat 69, 

Leuven, Belgium, mieke.defraeye@econ.kuleuven.be, 

Inneke Van Nieuwenhuyse 

When determining capacity levels in a healthcare system (such as an emergency 

department), a key feature that has to be taken into account is the time-varying 

demand for service. Due to these fluctuations in demand, determining capacity 

levels to achieve a certain service level is often complicated. In this presentation, an 

approach to determine staffing levels that results sufficiently small waiting times, 

will be presented. 

4 - Applying Inventory Control Practices Within the Sisters of Mercy 

Health Care Supply Chain 

Server Apras, University of Arkansas, 1343,N.Leverett,15, 

Fayetteville, AR, 72703, United States of America, sapras@uark.edu 

The goal of this research is to lay a foundation for the application and acceptance of 

more advanced inventory control practices within the healthcare supply chain.The 

project examines the demand characteristics and optimal control policies for bulk 

pharmaceuticals within the Sisters of Mercy’s network to compare to the current 

ordering and inventory control strategies to document potential cost savings.Also,a 

multiechelon inventory analysis examines the benefits of centralized inventory 

control 

5 - Outpatient Appointment Scheduling using Genetic Algorithms 

Song Chew, Assistant Professor, Southern Illinois University 

Edwardsville, Edwardsville, Edwardsville, IL, 62026, 

United States of America, schew@siue.edu 

Outpatient appointment scheduling has been an active area of research. The goal of 

the research is to strike a balance between patient waiting time, and doctor idle time 

and overtime. The objective of our work is to determine the optimal total number of 

patients for a clinical session, and the optimal number of patients assigned to each 

time slot in the session so as to minimize the total cost using genetic algorithms. 


406 

■ WB65 


Inventory Management VII 


Chair: Haitao Li, University of Missouri - St. Louis, 229 CCB, One 



1 - Making Better Fulfillment Allocation Decisions on the Fly 

Jason Acimovic, Massachusetts Institute of Technology, 

77 Massachusetts Ave, E62-459, Cambridge, MA, 02139, 

United States of America, acimovic@mit.edu, Stephen Graves 

E tailers manage complicated distribution networks, serving customers with 

heterogeneous service time requirements. What is the best way to fulfill each of 

these customers’ orders for items with low inventory levels? We partner with an etailer 

to examine this question. We find the value of the improvement gap by 

comparing a greedy strategy with an ex post facto optimization. We then develop an 

approximate dynamic programming heuristic and evaluate its performance on toy 

and actual examples. 

2 - Two-stage, Two-product, Capacitated Supplier Problem with 

Uncertain Demand 

Ramesh Bollapragada, Associate Professor, College of Business, San 

Francisco State University, 1600 Holloway Avenue, San Francisco, 

CA, 94132, United States of America, rameshb@sfsu.edu, Laoucine 

Kerbache, Kai Luo 

We investigate a finite-horizon two-product (expensive and cheaper product from 

capacitated local and far-away suppliers, respectively), one retailer problem with 

two-stages, and uncertain demand. In the first stage, we determine the order 

quantity vector to place with suppliers, and in the second stage we determine the 

allocation of inventory to the two products given limited shelf-space. Optimal and 

heuristic solutions are discussed. 

3 - Optimal Control of a Manufacturing/Remanuacturing Systems with 

Quality Grade Differentiation 

Morteza Pourakbar, Erasmus University Rotterdam, BurgOudlaan 50, 

Rotterdam, Netherlands, pourakbar@few.eur.nl, Mohsen Elhafsi, 

Saif Benjaafar 

We consider a manufacturing/remanufacturing system where each demand is 

coupled with the return of an item that may be remanufacturable. Returned items 

differ in their quality grades with grades affecting remanufacturing cost and time. 

Decisions must be made regarding whether or not to accept a returned item and 

whether to produce a new item or remanufacture a returned one and if so from 

which grade. 

4 - New Model and Heuristics for Safety Stock Placement in General 

Acyclic Supply Chain Networks 

Haitao Li, University of Missouri - St. Louis, 229 CCB, 

One University Blvd, St. Louis, MO, 63121, 

United States of America, lihait@umsl.edu, Dali Jiang 

We model the safety stock placement problem in general acyclic supply chain 

networks as a project scheduling problem, for which the constraint programming 

(CP) techniques are both effective and efficient in finding high quality solutions. We 

further integrate CP with a genetic algorithm (GA), which improves the CP solution 

quality significantly. The performance of our hybrid CP-GA algorithm is evaluated 

on randomly generated test instances. CP-GA is able to find optimal solutions to 

small problems in fractions of a second, and near optimal solutions of about 5% 

optimality gap to medium size problems in less than two minutes on average. 

■ WB66 


Facilities Planning and Design 


Chair: Li Zhang, Royal Bank of Scotland, 399 Main Ave., Apt 614, 

Norwalk, CT, 06851, United States of America, lieezhang@gmail.com 

1 - Framework for Measuring Rationale Clarity of Collaborative 

Design Decisions 



States of America, john.m.chachere@nasa.gov, John Haymaker 

Designers often must convey clear rationale supporting design decisions. We define 

rationale as a set of assertions about components (managers, stakeholders, 

designers, gatekeepers, goals, constraints, alternatives, and analysis) with variable 

clarity (coherent, concrete, connected, consistent, credible, certain, and correct). We 

relate these definitions in the Rationale Clarity Framework to enhance objectivity in 

evaluating tools and processes for design decision making.

2 - Order and Inventory Management System for a Food Bank 

Arsalan Paleshi, PhD Student, University of Louisville, Department 

of Industrial Engineering, JB Speed School of Engineering, 

Louisville, KY, 40292, United States of America, 

a0pale01@louisville.edu, Bulent Erenay, Trivikram Rao 

This project creates an Excel tool to support a food bank’s need for a less labor 

intensive, inexpensive ordering and inventory management system. Additionally 

application of lean principles like KANBAN, ANDON and visual management to 

control inventory are suggested. The extension of similar methods for other cases is 

also discussed. 

3 - A Study of Spine Layout for Semiconductor Manufacturing Plant 

Under the Multi-Floor Environment 

Chikong Huang, Professor, National Yunlin University of Science & 

Technology, 123 University Rd., Sec. 3, Department of Industrial 

Management, Touliu, Yunlin, 640, Taiwan - ROC, 

huangck@yuntech.edu.tw, Ming-Ru Tsai 

This study focuses on the spine layout in multiple floors by arranging workstations 

along several closed-loop moving tracks. The objective is to minimize the horizontal 

and vertical handling costs. The model and solution algorithm using Tabu search are 

developed and they are also verified by a numerical example. 

4 - Layout Analysis and Lead-time Calculations with PEM Forklifts 

Abhijit Gosavi, Missouri University of Science and Technology, 219 

Engineering Management, Rolla, MO, 65409, United States of 

America, gosavia@mst.edu, Suzanna Long, Scott Grasman 

PEM forklifts utilize hydrogen-based fuels and have been shown, in certain 

conditions, to be more economical than traditional battery-powered forklifts. We 

analyze the impact of using PEM forklifts on manufacturing layout and the lead 

time, including material-handling time, calculations. We use simulation to compare 

layouts that use PEM forklifts and battery-powered forklifts on the basis of the lead 

time and the empty travel time. 

5 - Kernel Method of Performance Estimation in Autonomous Vehicle 

Storage and Retrieval Systems 

Li Zhang, Royal Bank of Scotland, 399 Main Ave., Apt 614, Norwalk, 

CT, 06851, United States of America, lieezhang@gmail.com, 

Changjian Huang 

In this study, a systematic simulation combined with kernel learning method is 

proposed for use in Autonomous Vehicle Storage and Retrieval Systems (AVS/RSs). 

The proposed method reflects the nonlinear relationship between system 

performance and factors and provides universal usage in performance estimation for 

various system designs. Simulation results indicate that the developed models 

provide accurate and computational efficient estimations. 

■ WB67 


Simulation II 


Chair: Byung-In Kim, Associate Professor, Pohang University of Science 

and Technology, Nam-Gu, Hyoja-Dong, POSTECH, Pohang, Korea, 

Republic of, bkim@postech.ac.kr 

1 - Revenue Management Under Customer Choice Behavior 

Marco Bijvank, University of Montreal, CP 6128 Succ Centre-Ville, 

Pavillon André-Aisenstadt, Montreal, QC, H3C 3J7, Canada, 

bijvankm@iro.umontreal.ca, Pierre L’Ecuyer 

Observations from practice and recent research suggest a shift to customer oriented 

revenue management. During this presentation we discuss how to design a 

simulation tool to incorporate very general demand processes and different discrete 

choice models. We also show how this tool can be applied in practice to understand 

the customer choice decision making. 

2 - Formation Control Strategies for Groups of Mobile 

Autonomous Agents 

Wei Zhao, IBM Research - China, Diamond Building,ZGC Software 

Park, Beijing, China, wzhaow@cn.ibm.com, Wenjun Yin, Jin Dong, 

Bin Zhang, Ming Xie, Long Wang 

This paper presents the control strategies for multiple mobile autonomous agents to 

achieve leader-following formations. Each follower agent establishes a Bezier 

trajectory between its current position and that of its leader agent. Considering the 

nonholonomic properties of the agent, the optimization of Bezier curve’s curvature 

to choose appropriate scale factor is conducted by penalty function method. 

Simulations and experimental results show the effectiveness of our control 

strategies. 

3 - Demonstrating Simulations of Inventory and Production 

Management using VBA in Microsoft Excel 

Amanda Baty, Texas Tech University, 3424 Frankford Ave 

Apt 8B, Lubbock, TX, 79407, United States of America, 

amanda.baty@ttu.edu, Dr. Rafael Moras 


407 

Based on simulations by Dr. Paul Jensen and W.G. Lesso, “Advanced PNG Game” is 

a single product simulation using spreadsheets. The standard deviation of demand, 

lead time, and production level can be adjusted, for a cost, in each of the 52 

simulated weeks. The objective is to minimize total cost. 

4 - Agent-based Simulation for Emergency Evacuation of 

Heterogeneous Populations in a High-rise Building 

Jeongin Koo, PhD Student, POSTECH, San31, Hyoja-Dong, Namgu, 

Pohang, Korea, Republic of, xession@postech.ac.kr, Yong Seog Kim, 

Byung-In Kim 

An efficient evacuation plan is critical for the safety of residents in high-rise 

buildings during catastrophic events. Although there are some research works on 

building evacuation using simulation, most of them do not consider the 

handicapped people. This research presents an agent-based simulation of 

heterogeneous populations including various types of people with disabilities for a 

24-story real world building evacuation. Using the model, several evacuation plans 

are tested and evaluated. 

■ WB68 


Innovation/Entrepreneurship II 


Chair: David Gomulya, Foster Business School, University of Washington, 

MacKenzie 355, Box 353200, Seattle, WA, 98195, 

United States of America, dgomulya@u.washington.edu 

1 - Institutional Conditions and Venture Capital Investment in 

Developing Countries 

Theodore Khoury, Oregon State University, Bexell 422B, Corvallis, 

OR, 97331, United States of America, 

ted.khoury@bus.oregonstate.edu, Marc Junkunc, Santiago Mingo 

Focusing on entrepreneurial ventures in developing countries, we explore how 

venture capital (VC) investments are shaped by firms’ stage of development and 

host country institutional conditions. Using a panel of 443 VC investments occurring 

in Latin America over 12 years, we find that larger investments are affiliated with 

later stage firms, higher transaction costs and lower political hazard risk. Also, 

political hazards moderate the relationship between development stage and 

investment size. 

2 - New Ventures and Timing of Alliance Formation: The Dynamics of 

Temporal Congruence and Contingency 

David Gomulya, Foster Business School, University of Washington, 

MacKenzie 355, Box 353200, Seattle, WA, 98195, United States of 

America, dgomulya@u.washington.edu 

While alliances have been shown to increase new-venture survival, the literature 

remains silent regarding the effect of timing of alliance formations. Related 

literatures regarding timing have also been unable to explain their conflicting 

findings, which show the effect of timing can range from positive to negative. To fill 

gaps, I develop a novel model based on temporal changes during pre- and postformation 

phases of an alliance. I show the effect can indeed range from positive to 

negative. 

■ WB69 


Joint Session TSL/ SPPSN: Risk and Recovery of 

Transportation Networks to Disaster 




Chair: Nicholas Lownes, Assistant Professor, University of Connecticut, 

261 Glenbrook Rd., U-2037, Storrs, CT, 06269, United States of America, 

nlownes@engr.uconn.edu 

1 - Resilience of Fright Transportation Networks 

Xiaodong Zhang, University of Maryland, College Park, MD, 20742, 

United States of America, xzhang@umd.edu, Elise Miller-Hooks, 

Reza Faturechi 

The problem of measuring network resilience in transportation networks and 

determining optimal pre-event remedial actions is formulated as a two-stage 

stochastic program. The formulation explicitly recognizes that post-disaster 

performance depends not only on the inherent capability of the system to absorb 

externally induced changes, but also on the actions that can be taken in the 

immediate aftermath of the disaster to restore system performance. An L-shaped 

method is proposed for its solution.

WB70 

2 - Debris Operations 

Kael Stilp, Georgia Institute of Technology, 765 Ferst Drive, NW, 

Atlanta, GA, United States of America, mstilp3@isye.gatech.edu, 

Monica Villarreal, Antonio Carbajal, Ozlem Ergun, Pinar Keskinocak 

Debris operations following a disaster is a lengthy and costly process, involving 

multiple interrelated stages. Each stage contains a unique set of social fairness 

concerns with which need to be considered as well the computationally difficult 

objectives of efficiency. We present an encompassing set of models with 

computational analysis of each stage and discuss their interrelated nature. 

3 - Many-to-many Transportation Network Risk Assessment 

Qixing Wang, University of Connecticut, 261 Glenbrook Road, Unit 

2037, Storrs, CT, 06269, United States of America, 

qiw09005@engr.uconn.edu, Nicholas Lownes 

This work presents an extension to game-theoretic network risk models through the 

many-to-many application. A model that can be applied to large-scale networks is 

presented along with application results. Results provide a measure of link risk and 

can be used as decision support for sensor placement strategies. 

■ WB70 


Emerging Applications in Aviation 



Chair: Wenwei Cao, Doctoral Student, Georgia Institute of Technology, 

765 Ferst Dr NW, GA-Tech ISyE, Atlanta, GA, 30332-0205, United States 

of America, cww@gatech.edu 

1 - The Role of Air Travel in the Worldwide Spread of Vector 

Borne Diseases 

Lauren Gardner, Graduate Research Assistant, University of Texas at 

Austin, Earnest Cockrell Jr. Hall, 6.204, Austin, TX, 78712, United 

States of America, lmgardner@mail.utexas.edu, David Fajardo, 

Travis Waller 

Increased passenger air traffic has lead to an increased risk of importation and 

establishment of vector-borne diseases. As such, it is critical to be able to identify the 

risk associated with air travel routes between vector-compatible environments. This 

problem requires modeling the interaction between three overlapping networks: a 

vector survivability network, a social network and an airport network. We study the 

the role of the airport network in connecting the vector-survivability network. 

2 - Crew Rostering in Fractional Airlines 

Wenwei Cao, Doctoral Student, Georgia Institute of Technology, 765 

Ferst Dr NW, GA-Tech ISyE, Atlanta, GA, 30332-0205, United States 

of America, cww@gatech.edu, Ellis Johnson, Ozlem Ergun 

We consider a crew rostering problem in fractional airlines where rosters are 

generated through a bidding process. Instead of creating bid-lines beforehand, the 

management can only control the capacities on schedule lines and vacation periods. 

We propose an optimization based framework for capacity decision-making. Within 

the framework, a model incorporating training considerations is solved via column 

generation. 

3 - Transforming US Army Supply Chains: An Analytical Architecture for 

Enterprise Management 

Greg Parlier, Institute for Defense Analyses, Madison, AL, 

United States of America, gparlier@ida.org 

A comprehensive analytical architecture to enable US Army Logistics 

Transformation is presented, incorporating an “engine for innovation” to accelerate 

and sustain continual improvement. Strategic management challenges are 

addressed, including decision support systems, human capital investment needs, 

organizational design and strategic alignment for a learning organization. 

■ WB71 


Health and Education Applications of DEA 



Chair: Emmanuel Thanassoulis, Aston University, Operations & 

Information Management Group, Birmingham, AL, 

United States of America, e.thanassoulis@btinternet.com 

1 - Analysis by DEA of Health Referral Costs in England 

Emmanuel Thanassoulis, Aston University, Operations & Information 

Management Group, Birmingham, AL, United States of America, 

e.thanassoulis@btinternet.com, Maria C. Portela, Mike Graveney 


408 

The paper is drawn from an analysis by DEA of patient referral costs on behalf of a 

Primary Care Trust (PCT) in England. The paper focuses on General Practitioners 

(GPs) contracted to the PCT. GPs are the gateway to health services in England. 

Clinical decisions by GPs heavily influence referral costs incurred by the PCT. The 

paper formulates models to determine the scope for savings in referrals costs and 

decomposes them into those attributable to mix of in and out patient referrals and 

those attributable to price differentials between in and out patient cases. 

2 - A Decade of Clinical Productivity Change and the Determinants of 

Relative Clinical Efficiency in Pennsylvania Coronary Artery Bypass 

Graft Programs 

Jon Chilingerian, Brandeis University, The Heller School for Social 

Policy and Management, South Street, Waltham, MA, 02454-9110, 

United States of America, chilinge@brandeis.edu 

We present evidence on the growth of clinical productivity change in hospitals 

performing coronary artery bypass grafts (CABGs) over a decade. Analyzing 

outcome and severity-adjusted hospital clinical in Pennsylvania between 1994-2004, 

frontier methodology was used to measure and evaluate clinical productivity 

change, medical progress (i.e., shifts in the clinical frontiers), and relative clinical 

efficiency over time. In Pennsylvania, the average hospital’s clinical productivity for 

CABG surgeries grew by more than 30%. We found strong support, however, that 

medical-technical progress, rather than improvement in clinical efficiency, was the 

underlying reason for the growth. Factors associated with clinical efficiency 

included: employing hospitalists; having a higher percent of salaried physicians in 

relation to FTEs on salary; and strategic choices that focused on open heart 

cardiovascular surgeries. All of these factors are associated with performance and 

are under a clinical department’s control. 

3 - DEA in the 100 Largest U.S. Public School Districts 

N.K. Kwak, Saint Louis University, Department of Decision Sciences 

and MIS, St. Louis, MI, United States of America, kwakn@slu.edu, 

Walter Garrett 

This paper reports on a study of the 100 largest U.S. public school districts, which 

are together responsible for educating 22 percent of all public school students. Using 

data from the U.S. Department of Education, we use DEA to assess the efficiencies 

of those districts and to identify common characteristics of efficient districts. The 

results may be useful for benchmarking inefficient districts to improve their 

performance. 

■ WB72 


Facility Logistics III 



Chair: Ananth Krishnamurthy, University of Wisconsin, 1513 University 

Ave, Madison, United States of America, ananth@engr.wisc.edu 

1 - Modeling Inventory Visibility in Supply Chain Networks 

Sandeep Srivathsan, Research Associate, Oklahoma State University, 

School of IE&M, 322 EN, Stillwater, OK, 74078, United States of 

America, sandeep.srivathsan@okstate.edu, Manjunath Kamath 

We develop stochastic models of two-echelon supply chains where a retailer has 

visibility of production facilities’ net inventory levels and vice versa. Order 

placement and fulfillment policies depend on the net inventory levels and the 

models developed can be used to obtain insights into the benefits of inventory 

visibility on overall supply chain performance. We present models developed under 

Markovian assumptions and then discuss strategies to develop models for more 

general settings. 

2 - Evaluation of Decision Criteria to Select Efficient Order 

Picking Systems 

Detlef Spee, Detlef.Spee@iml.fraunhofer.de, Tim Geissen, 

Michael ten Hompel 

Order picking systems contribute to the competitive capability of companies. 

Increased requirements regarding performance and quality aspects necessitate 

flexible systems that meet present and future demands. The paper focuses on the 

decisions which are made during the planning process to select a proper order 

picking method. Therefore, relevant order picking methods are systematized and 

selection criteria are evaluated by the use of a market study amongst suppliers of 

order picking systems. 

3 - Estimating the WIP on a Conveyor Based Material Handling System 

with Multiple Stations 

Dima Nazzal, Assistant Professor, University of Central Florida, 4000 

Central Florida Blvd., Orlando, FL, 32816, United States of America, 

dnazzal@mail.ucf.edu, Vernet Lasrado 

We present a method to improve the estimates of the work in process (WIP) on a 

closed loop conveyor. We estimate the total traveling WIP and the WIP waiting to 

be loaded on the conveyor. A probabilistic distribution has been derived for the 

loading process and we test this model with a detailed simulation of a 

semiconductor manufacturing facility.

■ WB73 


Reducing Transportation Emissions and GHG 



Chair: Hakob Avetisyan, PhD Student, University of Maryland, College 

Park, MD, United States of America, havetisy@umd.edu 

1 - Freeway Congestion Mitigation and the Emissions 

Minimization Problem 

Alex Bigazzi, Portland State University, Portland, OR, 

United States of America, bigazzi@pdx.edu, Miguel Figliozzi 

We present a freeway control problem that not only minimizes delays but also takes 

emissions into account. We formulate several variants of a freeway traffic control 

problem. Utilizing real-world sensor data from Portland Oregon and calibrated 

emissions models we present and analyze solution results under different traffic 

conditions and levels of elasticity. 

2 - Modeling a Novel Method for Reducing Transportation Greenhouse 

Gas Emissions 

Erica Wygonik, University of Washington, Seattle, WA, 

United States of America, ewygonik@uw.edu, Anne Goodchild 

A model is developed to evaluate parameters effecting emissions to consider the 

environmental impacts of aggregating personal travel into shared-use services. This 

model indicates whether shared-use vehicles show significant benefit to the 

environment over personal vehicles, and under what conditions those emissions 

savings would be realized, using grocery store shopping in Seattle, Washington as 

the first case study to quantify and compare the total environmental impacts. 

3 - Greener Construction of Transportation Construction Projects 

Through Optimization 

Hakob Avetisyan, PhD Student, University of Maryland, College 

Park, MD, United States of America, havetisy@umd.edu, 

Elise Miller-Hooks, Suvish Melanta 

Optimization-based techniques are presented that permit a contractor to develop an 

equipment-usage plan that adheres to current environmental standards and 

anticipated new regulations, accounting for recent laws that might affect 

construction, and possible future carbon tax or cap and trade programs. These 

techniques aid contractors in trading off project cost, duration and resulting GHG 

emissions in bid development and aid contractors in making green construction 

decisions. 

■ WB74 


Public Transit IV 



Chair: Luca Quadrifoglio, Assistant Professor, Texas A&M Univeristy, 

CE/TTI bldg - Room 301I, 405 Spence St., College Station, TX, 77843, 

United States of America, lquadrifoglio@civil.tamu.edu 

1 - A Logit-based Assignment on a Transit Schedule Hypergraph 

Mark Hickman, Associate Professor, University of Arizona, Civil 

Engineering, 1209 E. Second St., Bldg 72, Tucson, AZ, 85721-0072, 

United States of America, mhickman@email.arizona.edu, 

Hyunsoo Noh 

A hyperpath has been a common network structure in transit assignment. Using the 

hyperpath, we propose a logit-based transit assignment on a transit scheduled 

network. We introduce a link-based and time-expanded network and two 

hyperpath algorithms, and we investigate the stochastic user equilibrium situation 

under hard capacity constraints. 

2 - Evaluating the Use of “Transfers” for Improving Zoning 

Paratransit Systems 

Chung-Wei Shen, Texas A&M University, CE/TTI Building, Room 

309C, 3136 TAMU, College Station, TX, 77840, United States of 

America, tzungwei0610@yahoo.com.tw, Luca Quadrifoglio 

For paratransit agencies, the zoning strategy divides their service area into smaller 

zones to different provider to simplify the management. After dropping customers 

off for cross zone trips, the vehicles will return without passengers on-board 

resulting in deadheading miles. Using transfers on boundaries is a promising method 

to decrease deadheading miles while applying zoning strategies. This study evaluates 

the performance of transfers on zoning paratransit systems through simulation 

method. 

INFORMS Austin – 2010 WC01 

409 

3 - The Design of Single-Line Demand Adaptive Systems: 

An Evaluative Framework 

Fausto Errico, CIRRELT, 2920, Chemin de la tour, Montreal, QC, H3T 

1J4, Canada, Fausto.Errico@cirrelt.ca, Teodor Gabriel Crainic, 

Federico Malucelli, Maddalena Nonato 

Demand-Adaptive Systems (DASs) display features of both traditional fixed-line bus 

services and purely on-demand systems such as dial-a-ride. A DAS bus line serves a 

given set of compulsory stops according to a predefined schedule. On the other 

hand, passengers may issue requests for transportation between two optional stops, 

inducing detours in the vehicle routes. The design of a DAS line is a complex 

planning process. We propose, evaluate and compare several possible design 

strategies. 

Wednesday, 1:30pm - 3:00pm 

■ WC01 


Technology Policy and Energy Markets 



Chair: Ekundayo Shittu, Assistant Professor, Tulane University, 7 

McAlister Dr., New Orleans, LA, 70118, United States of America, 

eshittu@tulane.edu 

1 - Volatility Pricing for Renewable Energy Sources 

Xiaoyue Jiang, Assistant Professor, Tulane University, 

7 McAlister Dr, New Orleans, LA, 70118, United States of America, 

xjiang@tulane.edu, Geoffrey Parker, Ekundayo Shittu, Anjali 

Sheffrin 

In recognizing the negative effects of volatility to reliability of the power grid, we 

have proposed an analytical framework that models capacity volatility by 

capacity@risk and prices volatility through a QoS-centric capacity market 

mechanism. In this work, we will apply this framework to renewable sources 

including solar and wind. We will price each type of sources based on their volatility 

characteristics. Real data from one of the ISOs in the US will be used to gain 

practical insights. 

2 - Flexible the Better? On the Design of RPS in Presence of Green 

Consumers and Emissions Trading 

Yihsu Chen, Assistant Professor, University of California Merced, 

Merced, CA, 95343, United States of America, 

yihsu.chen@ucmerced.edu, Lizhi Wang 

Emissions trading, green pricing programs and renewable portfolio standard (RPS) 

are three concurrent policies implemented in US to reduce reliance on fossil fuel 

and GHG emissions. Despite their differences in policy targets, they are closely 

related and integrated with competitive electric markets. We examines market 

outcomes of two aspects of RPS design in this talk: bundling and double counting. 

3 - The Influence of Market Structure and Policy on Technology 

Portfolio Investments 

Ekundayo Shittu, Assistant Professor, Tulane University, 

7 McAlister Dr., New Orleans, LA, 70118, United States of America, 

eshittu@tulane.edu, Geoffrey Parker, Xiaoyue Jiang 

We study firms’ incentives to invest in a portfolio of technologies under different 

markets and environmental policies. The impacts of policies are critical to 

understanding how competition and energy production mix. We pay attention to 

the representation of the technologies in the portfolio because of implications on 

cost. We demonstrate intriguing results that describe how investments and adoption 

incentives are shaped by strategic interactions between market structure and 

regulatory policy.

WC02 

■ WC02 


Energy I 


Chair: Le Xie, Assistant Professor, Texas A&M University, 3128 TAMU, 

216M Zachry Bldg, College Station, TX, 77843, United States of America, 

lxie@mail.ece.tamu.edu 

1 - Why Your Plug-in Vehicle with a 40-mile Battery Pack May 

Only Go 25 

Orkun Karabasoglu, PhD Candidate, Carnegie Mellon University, 

5000 forbes ave. 402 Scaife hall, Pittsburgh, PA, 15213, 

United States of America, karabasoglu@cmu.edu, Jeremy Michalek 

We investigate the effects of driving patterns on the range, gasoline consumption, 

greenhouse gas emissions, and lifecycle costs of conventional, hybrid, and plug-in 

hybrid vehicles. We find that drive cycle (style) can have a greater effect on 

consumption, cost and emissions than vehicle design or charging frequency. Plug-in 

vehicle range drops 35% under aggressive, rather than standard cycles, and drive 

cycle can affect which vehicles are optimal. 

2 - Stochastic Control for Smart Grid Users with Flexible Demand 

Yong Liang, PhD Student, University of California-Berkeley, 1117 

Etcheverry Hall, Berkeley, CA, 94720, United States of America, 

yongliang@berkeley.edu, Z. Max Shen, Alan Sanstad 

We propose a stochastic optimal control model for smart grid users to make the 

optimal energy usage decisions incorporating energy consumption and generation. 

The main feature of this model is its ability to dynamically adjust consumptions by 

responding to the pricing signals from the electricity grid, deal with stochastic new 

job arrivals as well as scheduling the jobs based on their own deadlines. The model 

leads to a dynamic programming problem, which is solved using ADP techniques. 

3 - Quantifying the Economic Impact of Variable Energy Forecast on 

Power System Scheduling 

Le Xie, Assistant Professor, Texas A&M University, 3128 TAMU, 

216M Zachry Bldg, College Station, TX, 77843, United States of 

America, lxie@mail.ece.tamu.edu, Haifeng Wang, Jin Dong, 

Wenjun Yin 

The integration of variable resources such as wind and photovoltaic has posed 

fundamental challenges to the operation of electric energy systems. We attempt to 

quantify the value of variable energy forecast in improving the efficiency of power 

system scheduling. We demonstrate in a realistic system that only by explicitly 

valuing near-term prediction and inter-temporal variations in the system scheduling 

model could the economic potential of these variable resources be fully utilized. 

■ WC03 


Environmental Operations 


Chair: Bin Dai, HKUST, A601, HKUST, Kowloon, Hong Kong - PRC, 

dbbudstar@gmail.com 

1 - Stochastic Cost Estimation Approach for Full-Scale Reverse 

Osmosis Desalination Plants 

Seong-Hee Kim, Associate Professor, Georgia Institute of Technology, 

765 Ferst Dr, Atlanta, GA, 30332, United States of America, 

skim@isye.gatech.edu, Pyung-Kyu Park, Jae-Hong Kim, 

Varun Gandhi, Pranay Mane, Hoon Hyung, Chuljin Park 

A stochastic approach was developed to estimate the construction and operation 

cost of a seawater reverse osmosis (SWRO) desalination plant. The stochastic cost 

model was further coupled with a process simulation model that predicts 

performance measures such as water production rate and produced water quality. 

The case study demonstrates the effectiveness of the coupled model in ranking and 

comparing a large number of design and operating conditions for the full-scale 

SWRO plant. 

2 - Self-Insurance: The Case of the Canadian Oil Sands 

Kalinga Jagoda, Assistant Professor, Mount Royal University, 

4825 Mount Royal Gate SW, Calgary, AB, T3E7N9, Canada, 

kjagoda@mtroyal.ca, Pamini Thangarajah 

In Canadian oil sands much of the negative publicity is around land pollution and 

there are calls for establishment of disaster relief program to manage hazardous 

events. Using the theory of public goods, we analyze the self insurance aspects of 

such program and the efficient provision levels. 


410 

3 - A Study on Setting Recycling Subsidy for Waste PCs in Taiwan 

Through Bi-level Nonlinear Programming 

Hsu-Shih Shih, Professor, Tamkang University, Taiwan, ROC, 

151 Ying-Chuan Rd., Tamsui, Taipei, 25137, Taiwan - ROC, 

hshih@mail.tku.edu.tw, Chia-Wei Hsu, Bo-Han Huang 

The study investigates a recycling subsidy decision for waste PCs in Taiwan through 

bi-level nonlinear programming (BLNP) for environmental protection. The upperlevel 

unit is RFMB (Recycling Fund Management Board) which hopes to balance 

the recycling fund and others by controlling the subsidy et al.; the lower-level’s is 

the recycling industry which likes to get maximum profits by controlling the 

recycling rate. The BLNP model can help obtain an optimal solution for the 

recycling system. 

4 - Design of Solar Radiation Management by Projecting Man-made 

Particles to Counter Global Warming 

Ka Shek Lee, Associate Professor, Department of Industrial 

Engineering and Logistics Managemen, The Hong Kong University of 

Science and Technology, IELM, HKUST, Kowloon, Hong Kong - PRC, 

nlee@ust.hk, Bin Dai 

Currently, mitigation efforts to counter global warming such as carbon emission 

reduction are proving time inefficient, therefore it desires for an alternative to cool 

the earth on an emergency basis. In this study, we proposed to project atmospheric 

man-made particles to manage the solar radiation. Fluid dynamics and scattering 

theory are employed to model particles’ life time and radiation forcing to determine 

particles’ size, projection height and amount required for selecting materials. 

5 - Performance Analysis on Introducing Intercity High-speed 

Railway System 

Bin Dai, HKUST, A601, HKUST, Kowloon, Hong Kong - PRC, 

dbbudstar@gmail.com, Ka Shek Lee 

Carbon emission reduction contributes to mitigate global warming. Transportation 

accounts for 20% global carbon emission and electricity-powered railway system is 

carbon efficient. In this study, discrete choice model is employed to capture the 

travel demand shift by introducing high-speed railway system. Performance analysis 

is conducted to evaluate the carbon emission reduction, energy and time saving 

conditioning on intercity distance and city size. Results show its attractiveness. 

■ WC04 


Sustainability I 


Chair: Thomas Sloan, University of Massachusetts Lowell, 

1 University Avenue, College of Management, Lowell, MA, 01854, 

United States of America, Thomas_Sloan@uml.edu 

1 - Managing Pollution in Oligopolistic Markets with 

Sustainability Constraints 

Sung Hoon Chung, The Pennsylvania State University, 244 Leonhard 

Bldg., University Park, United States of America, sxc447@psu.edu, 

Robert Weaver 

We present a differential variational inequality framework to consider the 

equilibrium patterns of oligopolistic markets through pollution taxes and quotas for 

sustainability. An algorithm is proposed to compute the oligopolistic firms’ 

equilibrium output rate, shipping pattern, and pollution flow with or without the 

central authority. A numerical example is also presented to illustrate the use of our 

theory. 

2 - Technology Base Second Tier Distribuitors Sustainability in Mexico 

Guillermo Torres, PhD in Administration, Tecnologico De Estudios 

Superiores De Coacalco, Av. 16 de Septiembre # 54, Coacalco de 

Berriozàbal, 55700, Mexico, chapultepec19@hotmail.com, 

Eduardo G. Hernandez-Martinez 

At present, it has become increasingly hard for businesses to survive in a globalized 

market, when traditional differentiators must adjust to new circumstances since 

they are less efficient as opposed to new scenes. This study is focused on identifying 

competitiveness of small business in Mexico, who are specialized in trading and/or 

distributing computer equipment, that increasingly find it more difficult to compete 

against the large chains. 

3 - A Stochastic Linear Programming Model for the Management of 

Emission Rights 

Justyna Dyduch, Cracow University of Economics, Department of 

Industrial and Ecological, ul. Rakowicka 27, 31-510 Kraków, Cracow, 

Poland, dyduchj@uek.krakow.pl 

The management of emission rights means using them to cover enterprise’s 

emissions, selling or buying them on the market, banking and borrowing them. We 

describe a multi-stage recourse model that optimizes the production, the use and 

acquisition of emission rights. The model is established to maximize enterprise’s 

total profit.

4 - Sustainable and Maintainable: An Equipment Maintenance Model 

with Environmental and Economic Inputs 

Thomas Sloan, University of Massachusetts Lowell, 1 University 

Avenue, College of Management, Lowell, MA, 01854, United States 

of America, Thomas_Sloan@uml.edu, Joseph Sarkis 

An MDP model using traditional economic measures and environmental measures is 

used to optimize equipment maintenance decisions. Increased deterioration leads to 

more energy usage, more scrap, and greater environmental burdens. Maintenance 

reduces these burdens but has environmental and economic impacts. The impacts of 

using different cleaning solvents to perform maintenance are estimated using data 

from the Toxic Use Reduction Institute (TURI). 

■ WC05 


Dynamic Programming/Control I 


Chair: Srinivasa Puranam, Rutgers University, 1 Washington Park, 

Newark, NJ, 07102, United States of America, karti@pegasus.rutgers.edu 

1 - Further Insight on Optimizing Taboo Criteria in Markov 

Decision Processes 

Michael N. Katehakis, Professor, Rutgers Business School, 

1 Washington Park, Newark, United States of America, 

mnk@andromeda.rutgers.edu, Srinivasa Puranam 

Optimization of systems is often based on costs associated with the states of the 

system. However, in many applications it is difficult to determine costs for all states. 

In such situations, one could consider maximizing taboo criteria such as the taboo 

mean return times for a propitiously defined set of taboo states. This is a hard 

problem and well-known methods can not be applied. However, we can provide 

further insight and heuristics for this problem. 

2 - Risk-Averse Control Problem for Undiscounted Infinite Horizon 

Markov Decision Models 

Ozlem Cavus, Rutgers Center for Operations Research, 640 

Bartholomew Road, Piscataway, NJ, 08854, United States of 

America, ozlem_cavus@yahoo.com, Andrzej Ruszczynski 

The Markov risk measure is introduced and used to formulate risk-averse control 

problems for finite and infinite horizon models by Ruszczynski. In this study, this 

new concept is used to formulate and solve an infinite horizon insurance problem 

where one of the control decisions is to purchase insurance. Furthermore, riskaversion 

is introduced for undiscounted infinite horizon Markov control models 

with absorption. 

3 - Managing Data Quality Risk in Accounting Information Systems 

Manuel Nunez, Associate Professor, School of Business, UConn, 

2100 Hillside Road Unit 1041, Storrs, CT, 06269, United States of 

America, mnunez@business.uconn.edu, Xue Bai, Jayant Kalagnanam 

The quality of data contained in accounting information systems has a significant 

impact on internal business decision-making and external regulatory compliance. 

We present a methodology for managing the risks associated with the quality of 

transactional data in accounting information systems. This methodology models the 

error introduction and propagation process in transactional data flow, and finds 

optimal control policies to mitigate data quality risks using a Markov decision 

process. 

4 - Vehicle Routing with Traffic Congestion and Drivers’ Driving and 

Working Rules 

Marco Schutten, University of Twente, Fac. Management and 

Governance, P.O. Box 217, Enschede, 7500 AE, Netherlands, 

m.schutten@utwente.nl, Leendert Kok, Erwin Hans, Henk Zijm 

We develop a solution method for the VRPTW, time-dependent travel times, and 

driving hours regulations. The major difficulty of this VRPTW extension is to 

optimize each vehicle’s departure times to minimize the duty time of each driver. 

We propose a restricted dynamic programming heuristic for constructing the vehicle 

routes, and an efficient heuristic for optimizing the vehicle’s departure times for 

each (partial) vehicle route, such that the complete algorithm runs in polynomial 

time. 

5 - Optimal Bidding in Sequential Auctions with Random 

External Demand 

Srinivasa Puranam, Rutgers University, 1 Washington Park, Newark, 

NJ, 07102, United States of America, karti@pegasus.rutgers.edu, 

Michael N. Katehakis 

We consider the problem of sequentially bidding in N auctions of identical items 

when items acquired are sold in a secondary market. The demand size and the sales 

price are random variables with known distributions. The objective is to acquire 

items through a sequence of auctions and sell them in the secondary market at 

maximum expected profit. We present a Markov decision processes model for this 

problem and study monotonicity properties of the optimal bids, for several cases of 

interest. 


411 

■ WC06 


Tutorial: Searching and Hiding on Networks 



Chair: J. Cole Smith, Professor, University of Florida, Industrial and 

Systems Engineering, 303 Weil Hall, P.O. Box 116595, Gainesville, FL, 

United States of America, cole@ise.ufl.edu 

1 - Searching and Hiding on Networks 

J. Cole Smith, Professor, University of Florida, Industrial and Systems 

Engineering, 303 Weil Hall, P.O. Box 116595, Gainesville, FL, 

United States of America, cole@ise.ufl.edu 

The problem of deploying and controlling a set of searchers on a network to locate a 

hidden target is referred to as a search game. There are countless variations of this 

game, depending on, e.g., whether or not the target is mobile, and the searchers’ 

communication and control capabilities. This tutorial will examine classical results in 

this field, discuss contemporary search literature, and explore research directions in 

this field as driven by future security challenges. 

■ WC07 


Supply Chain Optimization II 


Chair: Haitao Li, University of Missouri - St. Louis, 229 CCB, One 



1 - Game Theoretic Analysis of Supply Chain 

Satish Tyagi, Wayne State University, Department of Industrial, 

Manufacturing Engineering, Detroit, MI, 48202, 

United States of America, styagi.nifft@gmail.com, Kai Yang 

The aim of this paper is to exploit the salient features of Game Theory in modeling 

and to investigate SC functioning under various alliances. In order to give a realistic 

view to model, a comprehensive objective function has been formulated by 

combining different objectives. This paper introduces a novel approach Gaussian 

Particle Swarm Optimization which is embedded with beneficial attributes viz. (1) 

Gaussian probability distribution, and (2) Time Varying Acceleration Coefficients. 

2 - Measuring the Bullwhip Effect in the Supply Chain Based on the 

Demand Forecasting Coordination Level 

Seong-Hyun Nam, Associate Professor, University of North Dakota, 

Management Department, Grand Forks, ND, 58202-8377, 

United States of America, snam@business.und.edu 

This paper studies the role of coordination in relation of the bullwhip effect. It seeks 

to derive a bullwhip effect measure using a stochastic optimal control theory. In 

particular, we measure how much the supply chain bullwhip effect mitigation 

depends on the level of supply chain coordination of demand forecast and develop 

the strategic coordination policy. 

3 - Heuristic Procedures for Biomass-to-Biorefinery Supply Chains 

Ambarish Acharya, PhD Candidate, Mississippi State University, Dept 

of Industrial & Systems Engineering, Mississippi State, MS, 39762, 

United States of America, ama206@msstate.edu, Daniela Gonzales, 

Sandra Eksioglu 

The objective of this research is modeling and solving coordinated biomass-tobiorefinery 

supply chain design and management problem. We discuss a number of 

special cases of this problem where the location, capacity and number of facilities is 

known. We propose solution procedures to solve the general problem and its special 

cases. 

4 - Optimizing the Supply Chain Configuration for 

Make-to-Order Manufacturing 

Haitao Li, University of Missouri - St. Louis, 229 CCB, 

One University Blvd, St. Louis, MO, 63121, 

United States of America, lihait@umsl.edu, Keith Womer 

Configuring an MTO supply involves determining both sourcing and scheduling 

supply chain activities to meet customer’s demand in a cost effective and time 

efficient way. We develop an optimization model and algorithm for optimally 

configuring MTO supply chains. Managerial insights are derived and discussed.

WC08 

■ WC08 


Facility Location I 


Chair: Priyanka Verma, Indian Institute of Technology Kanpur, IME 

Department, Kanpur, 208016, India, priyankav08@gmail.com 

1 - A Location-allocation-local Search Procedure for Territory Design 

with Multiple Balance Constraints 

Roger Z. Ríos-Mercado, Universidad Autónoma de Nuevo León, 

CIDET-FIME, AP 111 - F, Cd. Universitari, San Nicolàs de los Garza, 

66450, Mexico, roger@yalma.fime.uanl.mx 

In this talk, a commercial territory design problem motivated by a real-world 

application in the bottled beverage distribution industry is addressed. A novel 

location-allocation-local search algorithm is presented and fully evaluated over a 

wide range of instances. The results indicate the excellent performance of some of 

the procedure components. It also shown how the algorithm finds design plans of 

significantly better quality than those currently handled by the company. 

2 - Hospital Location Planning in Singapore 

Kok-Choon Tan, Assoc Prof, National University of Singapore, NUS 

Bisiness School, 15 Kent Ridge Drive, Singapore, 119245, Singapore, 

kokchoon@nus.edu.sg, Joe SIM 

Singapore is currently served by 6 public hospitals which provide affordable and 

good quality healthcare. This research aims to enhance the health ministry’s 

capabilities in planning new acute and community hospitals, which are costly 

investments. We will describe the use of Urban OR techniques to develop a decision 

support system that allows policy makers to compare the impact among different 

choices of new hospital locations with the corresponding sizes and capabilities. 

3 - Enabling Easy Consumer Access to Services and Products 

Baris Hasdemir, UMASS, Isenberg School of Management, 

Finance and Operations Management, Amherst, 01003, 

United States of America, hasdemir@som.umass.edu, Agha Ali 

Enabling better access to services for spatially dispersed populations is pertinent in 

today’s fiscally constrained socio-political landscape. A network of centers provides 

better access if the centers are located so as to serve maximal populations within 

each of multiple threshold distances. Computational studies using a model with 

distance differentiation of the population involving 1,829 model instances reveal 

insights about possible access for the 224M people living in 23K places in the US. 

4 - Facility Location and Relocation Problem (FLRP-U) 

Under Uncertainty 

Ayse Durukan Sonmez, PhD Candidate, University of Houston, 

Department of Industrial Engineering, Engr. Bldg 2, Houston, TX, 

77204, United States of America, adurukan@uh.edu, Gino Lim 

FLRP-U is a facility location and relocation problem that considers future demand 

changes as well as uncertainties in number of future facilities. The objective is to 

minimize the sum of the initial weighted distance and the expected future distance, 

within a given budget for opening and closing of facilities. We propose a 

decomposition algorithm that can produce near optimal solutions for FLRP-U. Our 

numerical results compare the performance of our algorithm with exact solution 

techniques. 

5 - Single/two Stage Warehouse Location Problem: Solution Technique 

and Strong, Weak, Hybrid Formulations 

Priyanka Verma, Indian Institute of Technology Kanpur, IME 

Department, Kanpur, 208016, India, priyankav08@gmail.com, 

RRK Sharma 

For the single stage uncapacitated warehouse location problems, hybrid 

formulations are shown to be the best performer against weak and strong 

formulations for large sized problems. Vertical decomposition approach is presented 

for the single and two stage capacitated warehouse location problem. Large sized 

complex warehouse location problem is decomposed into the smaller and relatively 

simpler versions of the capacitated plant location problems by the use of vertical 

decomposition approach. 


412 

■ WC09 


Retail Operations and Assortment Planning 



Chair: Felipe Caro, UCLA Anderson School of Management, 

110 Westwood Plaza, Suite B-420, Los Angeles, CA, 90095, 

United States of America, fcaro@anderson.ucla.edu 

1 - Dynamic Assortment Strategies for Variety-Seeking Consumers 

Dorothee Honhon, University of Texas at Austin, IROM, B6500, 

Austin, TX, United States of America, 

Dorothee.Honhon@mccombs.utexas.edu, Gurhan Kok 

In this project, we first characterize the static optimal assortment for a retailer with 

variety-seeking consumers. We then characterize the optimal assortment sets in the 

dynamic strategy that offers possibly different assortments each period. We show 

that it is possible to generate a high level of satisfaction at the customer level by 

having a mixed assortment strategy. 

2 - Dynamic Assortment Customization with Limited Inventories 

Fernando Bernstein, Duke University, 1 Towerview Dr., Durham, 

NC, 27708, United States of America, fernando@duke.edu, 

Gurhan Kok, Lei Xie 

We consider a retailer with limited inventories of a category of substitutable 

products and heterogeneous customer preferences. Customers arrive sequentially 

and the firm decides which subset of the products to offer to a customer depending 

on the customer type, the inventory levels and the time-to-go in the season. We 

show that limiting the choice set of some customers can significantly increase 

profitability. 

3 - Dynamic Assortment Models: A Portfolio Approach 

Felipe Caro, UCLA Anderson School of Management, 

110 Westwood Plaza, Suite B-420, Los Angeles, CA, 90095, 

United States of America, fcaro@anderson.ucla.edu, Rene Caldentey 

We investigate optimal dynamic assortment planning strategies for a retailer with 

limited shelf space. The retailer can choose among basic and fashion items with low 

and high risk (and return) respectively. We present two models within this setting. 

One is theoretical where we explicitly model the vogue trend as a stochastic process 

that the retailer tries to follow. The second model has a similar objective but a much 

simpler formulation intended for easy implementation in practice. 

■ WC10 


Track and Trace Technologies in Supply Chains 



Chair: Gary Gaukler, Texas A&M University, TAMU-3131, College Station, 

United States of America, gaukler@tamu.edu 

1 - Item-level RFID Tagging and Inventory Record Accuracy 

John Aloysius, University of Arkansas, WCOB 226 University of 

Arkansas, Fayetteville, United States of America, 

JAloysius@walton.uark.edu, Bill Hardgrave, Sandeep Goyal 

Previous research has demonstrated that case-level RFID tagging can improve 

inventory record accuracy for consumer packaged goods. The increased visibility 

provided by item-level tagging however enables tracking items in-store right up to 

the point-of-sale. We report the results of experiments in the field that investigate 

the potential of item level tagging in the retail store. 

2 - The Impact of Supply Network and Product Characteristics on 

Tracking Technology Assimilation 

Rahul Basole, Tennenbaum Institute, Georgia Tech, 760 Spring Street 

NW, Atlanta, GA, 30332, United States of America, 

rahul.basole@ti.gatech.edu, Maciek Nowak 

This research examines the influence of supply network and product characteristics 

on the extent of tracking technology (TT) adoption and use. We develop and test a 

theoretical model through a global survey of supply chain executives. Our results 

provide insights into how factors such as product value, handling risk, and supply 

network complexity impact the level of TT assimilation. This study yields important 

implications for the management of emerging IT in a supply chain operations 

context.

3 - The Impact of Product Contamination in the Food Supply Chain 

Vijaya Chebolu-Subramanian, Texas A&M University, 3131 TAMU, 


cheb12@neo.tamu.edu, Gary Gaukler 

In this talk we discuss the impact of product contamination in a food supply chain 

(e.g., e. coli contamination of spinach). We consider a real world situation in which 

a contaminated food product has entered a supply chain and is being sold at the 

retailer. We develop a quantitative model to quantify the overall cost of the 

contamination event and evaluate the effect of the product and supply chain 

attributes on the performance metrics of the model. 

■ WC11 


Supply Chain Issues & Sustainability 



Chair: Chelliah Sriskandarajah, The University of Texas at Dallas, 800 

West Campbell Road, SM30, School of Management, Richardson, TX, 

75080, United States of America, chelliah@utdallas.edu 

1 - Optimal Life Cycle Inventory Management 

Burcu Keskin, Assistant Professor, University of Alabama, Alston 

Hall, Tuscaloosa, AL, United States of America, bkeskin@cba.ua.edu, 

Charles Schmidt 

We consider inventory replenishment of a product with nonstationary, stochastic 

demand that moves probabilistically from one stage of the life cycle to the next, 

each with its own demand and cost information, until the end-of-life. The planning 

horizon is neither infinite nor of a known finite length. Via analytical analysis, we 

show that stage-dependent base stock policy is optimal. Via numerical analysis, we 

show the impact of not modeling the random product life on cost and inventory 

levels. 

2 - On the Tradeoff Between Remanufacturing and Recycling 

Tharanga Rajapakshe, University of Texas at Dallas, 800 West 

Campbell Road, SM30, School of Management, Richardson, TX, 

75080, United States of America, tharanga@utdallas.edu, 

Srinagesh Gavirneni, Milind Dawande, Chelliah Sriskandarajah 

Motivated by our interactions with two Dallas-based reverse-logistics firms, we 

analyze the tradeoff between two well-known product-recovery approaches: 

recycling and remanufacturing. Our analysis exploits the supply- and demand-side 

implications as well as product design characteristics of these approaches. We 

provide a complete theoretical characterization of the tradeoff and develop rich 

insights on the influence of ability of sustainability and disposal costs. 

3 - Inventory Models for Medium-size Depository Institutions Under the 

New Federal Reserve Policy 

Yunxia Zhu, The University of Texas at Dallas, 800 West Campbell 

Road, SM30, School of Management, Richardson, TX, 75080, 

United States of America, yunxia.zhu@student.utdallas.edu, Milind 

Dawande, Chelliah Sriskandarajah 

We study two new multi-period models — designed specifically to capture the 

operations of a medium-size Depository Institution — that emerge from its objective 

to minimize the total cost incurred in managing the inventory of cash over a finite 

planning horizon under the new Federal Reserve policy. We develop several 

managerial insights from a comprehensive test bed and demonstrate a procedure to 

easily adapt the optimal solutions based on projected data to near-optimal real-time 

solutions. 

4 - Scheduling Robotic Cells Served by a Dual-Arm Robot 

Manoj Vanajakumari, Texas A&M University, College Station, TX, 

United States of America, manoj@entc.tamu.edu, Avanti Sethi, 

Chelliah Sriskandarajah, Neil Geismar 

We assess the benefits of implementing a dual-arm robot in a flow shop 

manufacturing cell. The robot has the ability to tend two adjacent machines 

simultaneously. We identify optimal sequences for two and three machine cells and 

also derive structural results for cells with an arbitrary number of machines. For 

cells processing different part-types, we completely analyze two-machine cells. For 

each case we compare the productivity of single-arm and dual-arm robotic cells. 


413 

■ WC12 


Retail Supply Chain Management 


Supply Chain 


Chair: Sandra Transchel, The Penn State University, University Park, State 

College, PA, United States of America, sxt37@psu.edu 

1 - Multi-echelon Inventory System for E-Retailer Fulfillment Centers 

Juan Li, Cornell University, Ithaca, NY, 14853, jl879@cornell.edu, 

John Muckstadt 

The customers of e-retailers have different timeliness requirements on order 

delivery. The demand process has significant impact on designing the fulfillment 

centers. In this paper, a multi-echelon system is designed for this type of system. We 

describe an algorithm to compute the optimal order-up-to level for this system. 

2 - Joint Inventory Pricing and Assortment Decisions for 

Vertically Differentiated 

Mrinmay Deb, Student, Penn State University, 462A Business 

Building, University Park, State College, PA, 16802, 

United States of America, mud166@psu.edu, Susan Xu 

We determine the joint pricing, inventory, and assortment decisions of a retailer 

stocking quality differentiated products. Consumers choose their first choice based 

on the vertical choice model. We consider two cases: inventory is abundant (riskless 

case) and inventory is finite (risky case) and find the optimal prices and optimal 

assortments in each case. 

3 - Joint Pricing and Inventory Decisions Under 

Stockout-Based Substitution 

Sandra Transchel, The Penn State University, University Park, 

State College, PA, United States of America, sxt37@psu.edu, 

Anna-Lena Beutel, Stefan Minner 

We consider a joint inventory and pricing problem for partially substitutable 

products in a given assortment. Both price and inventory decisions have to be made 

under demand uncertainty. Additionally, the retailer has to take into account that 

customers are willing to substitute their first choice product if this is not available. 

We present structural properties of the proposed model and provide managerial 

insights into the interaction of pricing, stocking decision, and product 

cannibalization. 

■ WC13 


Managing Service Systems with Self-interested Actors 




Chair: Eren Cil, University of Oregon, 1208 University of Oregon, 

Lindquist College of Business, Eugene, OR, 97403-1208, 

United States of America, erencil@uoregon.edu 

1 - A Service Marketplace with Multiple Classes and Multiple 

Skilled Agents 

Eren Cil, University of Oregon, 1208 University of Oregon, Lindquist 

College of Business, Eugene, OR, 97403-1208, United States of 

America, erencil@uoregon.edu, Gad Allon, Achal Bassamboo 

We consider a service marketplace in which customers have different service needs. 

Each customer can be served by either a group of agents specialized to cater 

particular needs of that customer or a general pool of agents who can handle any 

requests but not as competently as the specialized agents. We analyze how the selfinterested 

customers choose between the dedicated and the general service, and 

characterize the response of the agents in the general pool to the customers’ 

choices. 

2 - Price Competition Under Multinomial Logit Demand Funcations with 

Random Coefficients 

Margaret Pierson, Harvard Business School, Morgan Hall, Boston, 

MA, United States of America, mpp2002@columbia.edu, 

Awi Federgruen, Gad Allon 

We postulate a general class of price competition models with Mixed MNL demand 

functions under affine cost. By imposing a natural upper bound for the price levels 

of each firm, we characterize the equilibrium behavior of these models in the case 

of single-product firms and then generalize the results to the multi-product case. 

This work provides a justification for the many structural estimation methods which 

require that equilibria correspond with the solutions to this system of FOC 

equations.

WC14 

3 - Coordinating Diagnosis and Service in a Service Supply Chain 

Mehmet Fazil Pac, PhD Candidate, Wharton School of Business, 

University of Pennsylvania, 3730 Walnut Street, Jon M. Huntsman 

Hall, Office 527.6, Philadelphia, PA, 19104, United States of America, 

mpac@wharton.upenn.edu 

We consider a service consisting of two phases; diagnosis and service. Each phase is 

provided by a self-interested agent. Diagnostic accuracy depends on the effort 

exerted by the diagnosis agent. Investing more time in diagnosis leads to higher 

valuation of the actual service by customers, however it also leads to lower 

throughput and more congestion for the system. Using a queuing framework, we 

analyze the agents’ diagnosis and pricing decisions under the presence of strategic 

customers. 

4 - Service Control of a Queue Under Different Delay 

Information Structures 

Brent Dooley, Wharton School, 3730 Walnut St., 500 Jon M. 

Huntsman Hall, Philadelphia, PA, 19104, United States of America, 

dooleyb@wharton.upenn.edu, Noah Gans, Omar Besbes 

We analyze the problem of profit-maximization for an M/M/1 queue with dynamic 

service rate control when customers have access to varying degrees of state/policy 

information. In particular, we examine customer equilibrium behavior and its 

implications on optimal service rate policies. 

■ WC14 


Supply Chain Management X 


Chair: Jie Yang, Associate Professor, University of Houston-Victoria, 

14000 University Blvd, Sugar Land, TX, 77479, United States of America, 

jieuhv@gmail.com 

1 - Mid-term Planning Coordination Between Maker and Retailer by 

Capacity Reservation 

Seung-Jin Ryu, Research Associate, Waseda University, 51-14-07, 3- 

4-1 Okubo, Shinjuku, Tokyo, 1698555, Japan, sjryu@aoni.waseda.jp, 

Hisashi Onari 

Collaborative SCM is the collaborated initiatives among SC members to deal with 

various problems over SC. It is the challenging issue, in high-tech industry, 

especially experiencing severe market demand change, keeping the stable and 

profitable operation level. We propose the mid-term planning coordination method 

between maker and retailer by capacity reservation. Also, we verify strengths and 

weaknesses of the proposed method by dynamic simulation with various business 

environments. 

2 - Transportation Pricing of a Truckload Carrier 

Aysegul Toptal, Assistant Professor, Bilkent University, Department of 

Industrial Engineering, Ankara, Turkey, toptal@bilkent.edu.tr, 

Safa Bingol 

In this study, we investigate the transportation pricing problem of a truckload 

carrier in a setting that consists of a retailer, a truckload carrier and a less than 

truckload carrier. Numerical evidence shows that the truckload carrier may increase 

his/her gainings significantly through better pricing and there is further opportunity 

of savings if the truckload carrier and the retailer coordinate their decisions. 

3 - Does Agility Matter in Improving Supply Chain Performance? 

Evidence From a Transition Economy 

Jie Yang, Associate Professor, University of Houston-Victoria, 14000 

University Blvd, Sugar Land, TX, 77479, United States of America, 

jieuhv@gmail.com 

This study develops and empirically tests a conceptual framework to investigate the 

antecedents of manufacturers’ supply chain agility and the connection of their 

agility with performance in an emerging economy. Drawing upon the information 

theory, this study argues that technical (IT capability) and relational factors 

(information sharing and trust, and operational collaboration) are the antecedents 

of a manufacturer’s supply chain agility. 


414 

■ WC15 


Organization Theory I 


Chair: Theresa Edgington, Baylor University, One Bear Place, #98005, 

Waco, 76798, United States of America, theresa_edgington@baylor.edu 

1 - A Clustering Approach For Multi-Facility Location Problem 

Cem Iyigun, Department of Industrial Engineering, Middle East 

Technical University, Ankara, Turkey, iyigun@ie.metu.edu.tr, 

Adi Ben-Isreal 

A new clustering approach is proposed. The method is a generalization of Weiszfeld 

method and applied for solving K facilities location problem. The problem is relaxed 

using probabilistic assignments, and is decomposed into K single facility location 

problems, that are coupled by the probabilities, and can be solved in parallel. 

2 - Considering the Relationship Between Empowerment and 

Resistance to Change 

Nathan Culmer, The University of Iowa, 2800 UCC, Iowa City, IA, 

52242, United States of America, nathan-culmer@uiowa.edu 

Both psychological and team level empowerment have been shown to have 

desirable effects in organizations while employee resistance to change can have 

undesirable effects on organizational or work group progress. Interestingly, these 

characteristics share some common theoretical elements. This paper considers 

possible theoretical relationships between empowerment as a motivational construct 

and the influence that empowerment can exert on employee resistance to change. 

3 - Institutional Barriers to Application of Continuous Improvement 

Processes From Safety to Energy Efficiency 

Rodney Lacey, Lecturer, Univ of California Davis, Graduate School of 

Management, One Shields Avenue, Davis, CA, 95616, 

United States of America, rolacey@ucdavis.edu 

Historic institutional differences between the fields of safety and energy efficiency 

could limit transfers of continuous improvement programs. Three differences of 

commensuration (Espeland and Stevens, 1998) between lives and units of energy 

are potential barriers: being efficient is commendable, rather than inefficiency 

deplorable; moral value of efficiency gains varies across industries; and there are no 

moral objections to making tradeoffs between efficiency and other goals. 

4 - To Build or Break Away? Exploring the Antecedents of Category 

Spanning Nanotechnology Innovation 

Tyler Wry, U of Alberta, 3-23 Business Building, Edmonton, AB, T6G 

2R6, Canada, twry@ualberta.ca 

Category spanning has implications for innovation and institutional change. Still, 

studies have focused primarily on its detrimental effects, eliding consideration of 

causal antecedents. Examining the nanotube technology field, I argue that patterns 

of linkage amongst categories enable spanning. Competing hazard rate models are 

supportive. Further, I find these linkages are shaped endogenously by prominent 

actors and condition the effects category richness, social networks, and social 

influence. 

5 - Evidentiary Analysis of Organizational Coordination and Questions 

of Group Fragmentation 

Theresa Edgington, Baylor University, One Bear Place, #98005, 

Waco, TX 76798, United States of America, 

theresa_edgington@baylor.edu 

Analysis processes contribute to organizations by increasing their knowledge, which 

is necessary to maximize decision-making and resolution outcomes. We investigate 

whether a core group provides exemplary coordination by investigating its database. 

Utilizing quantitative analysis, the evidence refutes this claim, contributing to the 

literature by identifying eleven coordination process indicators. 

■ WC16 


Flexible Manufacturing Systems 


Chair: Mabel Chou, Associate Professor, National University of Singapore, 

BIZ 1 Mochtar Riady Building, #8-66, 15 Kent Ridge Drive, Singapore, 

119245, Singapore, mabelchou@nus.edu.sg 

1 - Value of Flexibility 

Sanjeev Bordoloi, Associate Professor, University of St Thomas, 1000 

LaSalle Ave, TMH 434, Opus College of Business, Minneapolis, MN, 

55403, United States of America, sbordoloi@stthomas.edu 

Economic justification of investments in flexibility has become increasingly 

important. This paper tries to provide possible measures for flexibility so that 

managerial decision making for flexibility can be validated. We offer a model that 

minimized production planning costs and then use some of its characteristics for 

flexibility measurements.

2 - Joint Maintenance and Operations Decision Making in Flexible 

Manufacturing Systems 

Merve Celen, University of Texas at Austin, 1 University Station 

C2200, Austin, TX, 78712-0292, United States of America, 

merve@mail.utexas.edu, Dragan Djurdjanovic 

In semiconductor manufacturing, the dynamic interactions among operation types, 

chamber degradations and wafer yields necessitate a joint decision making of 

maintenance scheduling and product dispatching. To address this problem, we 

devise an integrated decision making policy with the objective of maximizing an 

adaptive profit function with respect to operation-dependent degradation models 

and production target by using a metaheuristic method based on the results of 

discrete-event simulations. 

3 - Dynamic Control of Closed Flexible Queueing Network with 

Application to Shipbuilding 

Fang Dong, Ph.D Candidate, University of Michigan, Industrial & 

Operations Engineering, 1205 Beal Ave., Ann Arbor, MI, 48109- 

2117, United States of America, ppfang@umich.edu, 

Mark Van Oyen, David Singer 

The U.S. Shipbuilding Industry is facing the challenge of building ships on-time and 

at budgeted cost. We introduce operational flexibility to ship production to mitigate 

the issues like high variability in production workload and low facility utilization. 

We formulate this problem as a flexible, controlled closed queueing network with 

CONWIP release policy. Under an effective control policy, the flexible system can 

significantly reduce ship completion time, and improve workshop utilization. 

4 - A Graph-theoretic Methodology for Deadlock Resolution in 

Automated Manufacturing Cells 

Venkatesh Angirasa, Research Scientist & Manager, Group for 

Decision Technology Solutions, SETLabs, Infosys Technologies, 

Electronics City, Hosur Road, Bangalore, 560100, India, 

venkatesh_angirasa@infosys.com 

We present a deadlock detection and resolution strategy for manufacturing cells 

with alternate part routing. We distinguish between cells with centralized buffers 

and those with dedicated I/O buffers for individual machines. A unified bipartite 

graph of the part-machine relationship enables the detection and resolution scheme 

in both cases. Resolution policies are developed to accommodate varying 

manufacturing systems requirements. An efficient algorithm is proposed to manage 

deadlocks. 

5 - Value of the Third Chain: Effect of Partial Production Postponement 

on Process Flexibility 

Mabel Chou, Associate Professor, National University of Singapore, 

BIZ 1 Mochtar Riady Building, #8-66, 15 Kent Ridge Drive, 

Singapore, 119245, Singapore, mabelchou@nus.edu.sg, 

Geoffrey A. Chua, Chung Piaw Teo 

Using a multi-item newsvendor model with second-stage supply and partial capacity 

sharing, we discover that the flexibility loss of the 2-chain is no longer negligible 

under partial postponement. For small systems, this loss can be as high as 20% to 

30%. However, we find that by adding another layer of flexibility, a third chain, the 

flexibility loss can be restored to the same level as 2-chain with full postponement. 

We show that the value of the third chain extends even to very large systems. 

■ WC17 


Supply Chain, Practice and Empirics 


Chair: Orrin Cooper, University of Pittburgh, 233 Mervis Hall, Pittsburgh, 

PA, 15260, United States of America, orc1@pitt.edu 

1 - Key Metrics and Current Industry Practices in Supply 

Chain Measurement 

Ramesh Bollapragada, Associate Professor, College of Business, San 

Francisco State University, 1600 Holloway Avenue, San Francisco, 

CA, 94132, United States of America, rameshb@sfsu.edu, Calvin Lee, 

Tuna Cencki 

The paper identifies current business practices and key measurements in supply 

chain performance management. Our survey based research indicates that supply 

chain professionals give high importance to quality and reliability factors, increasing 

company’s competitiveness, and not just to cost factors, while creating effective 

supply chains. The research underlines that more steps need to be taken to connect 

supply chain partners, and create models for the quantification of intangible metrics. 


415 

2 - Information Transmission and the Bullwhip Effect 

Robert Bray, PhD Student, Graduate School of Business, Stanford 

University, 311 Sheridan Ave, Palo Alto, CA, 94305, 

United States of America, rlbray@stanford.edu, Haim Mendelson 

We study the bullwhip effect in a sample of 4,689 public U.S. companies over 1974- 

2008. The traditional bullwhip is negative, but the “demand-uncertainty” bullwhip 

is positive. We decompose the total bullwhip into deterministic and stochastic 

components, and further decompose the latter by information transmission lead 

times. Bullwhips come in several flavors—-firms can anticipate much, but not all, of 

the bullwhip. 

3 - Business Relationship Functions and Supply Chain Relationship 

Quality: Evidence From China 

Yongtao Song, School of Management, PO Box 2341, Xi’an Jiaotong 

University, Xi’an, 710049, China, xjtusyt@gmail.com, Qin Su 

This paper adopts a functional view to analyze the value that buyers attain from 

buyer-seller relationships and investigates the links between business relationship 

functions (BRF), supply chain relationship quality (SCRQ) and buyer’s performance. 

The results indicate that BRF have a direct and an indirect effect on buyer’s 

performance through the mediating effect of SCRQ. Moreover, the availability of 

alternative suppliers has a moderating influence on the relationship between BRF 

and SCRQ. 

4 - 3rd Party Logistics Provider Selection with Performance 

Metrics and ANP 

Orrin Cooper, University of Pittburgh, 233 Mervis Hall, Pittsburgh, 

PA, 15260, United States of America, orc1@pitt.edu, Jennifer Shang, 

Pandu Tadikamalla 

Selecting a third party logistics provider with the Analytical Network Process (ANP) 

allows one to measure the interrelated influences of performance metrics in the 

supply chain. The criteria in the decision matrix are organized according to the 

temporal stages or flow of a product through the supply chain. A general model is 

presented and applied to case data and tested for robustness with detailed sensitivity 

analysis. 

■ WC18 


Industry Applications 


Chair: Fubin Qian, PhD Candidate, Molde University College, 

Britvegen 2, N-6411, Fannestrandveien 76, 6416, Molde, Norway, 

fubin.qian@himolde.no 

1 - Dynamic Capacity Planning for Short Life Cycle Products 

Saman Alaniazar, PhD Candidate, Wayne State University, 4815 

Fourth St., Detroit, MI, 48202, United States of America, 

saman.alaniazar@wayne.edu, Alper Murat, Ratna Babu Chinnam 

We study capacity planning for short life cycle products. Given the non-stationary 

and region-based nature of the demand in this type of products, our model employs 

a hybrid-strategy of short-term demand forecasting and long-term demand 

modeling to monitor amount and time of expansions. Moreover, we consider and 

apply a risk model in the process of capacity planning. We report experimental 

results based on an implementation of the model in an agent based system for 

Tamagotchi Case. 

2 - Optimizing Credit Lines 

Lisa Kart, Director, Analytics, FICO, Austin, TX, 78704, 

United States of America, lisakart@fico.com 

Over the past several years, FICO has helped banks optimize credit line decisions 

using predictive and decision modeling in a framework of constrained optimization. 

Hear about the approach, how the analytics were implemented, and the results 

achieved for banks across the globe. 

3 - The Industry Emergence Funnel: Towards a Conceptual Framework 

for Public Sector Coordination, Investment Prioritization and 

Strategy Development 

Eoin O’Sullivan, University of Cambridge, Cambridge, 

United Kingdom,eo252@cam.ac.uk 

This paper introduces a conceptual framework for analyzing the emergence of novel 

technology-based industries. We make the case for an industry-level analogue of 

established, firm-level “funnel” models of innovation. The features of the proposed 

framework are based on observed patterns of historical emergence of technologybased 

industries analyzed using roadmapping techniques.

WC19 

4 - A Tabu Search Heuristic for Offshore Helicopter Routing Problem 

with Focus on Passenger Safety 

Fubin Qian, PhD candidate, Molde University College, Britvegen 2, 

N-6411, Fannestrandveien 76, 6416, Molde, Norway, 

fubin.qian@himolde.no, Irina Gribkovskaia, Gilbert Laporte, ÿyvind 

Halskau 

A mathematical model is proposed to improve passenger transportation safety by 

minimizing the expected number of fatalities in offshore helicopter transportation. 

Tabu search heuristics are developed. Both the mathematical model and heuristics 

are capable of producing general solutions, namely solutions allowing a second visit 

to installations. Computational results show that safety performance can be 

significantly improved by introducing general solution strategy to helicopter routing 

problem. 

■ WC19 


Dynamic Optimization in Energy Pricing 



Chair: Canan Uckun, University of Chicago, 5807 South Woodlawn 

Avenue, Chicago, IL, 60637, United States of America, 

cuckun@chicagobooth.edu 

1 - Market Power Analysis in Electricity Markets with 

Time-of-use Pricing 

Emre Celebi, University of Waterloo, Department of Management 

Sciences, 200 University Ave. West, Waterloo, ON, N2L 3G1, Canada, 

ecelebi@engmail.uwaterloo.ca, David Fuller 

We propose variational inequality models for electricity markets with time-of-use 

(TOU) pricing. The demand response is dynamic in the model through a 

dependence on the lagged demand. Different market structures are examined 

within this context. With an illustrative example, the welfare gains/losses are 

analyzed after an implementation of TOU pricing scheme over the single pricing 

scheme. Also, break-up of a large supplier into smaller parts is investigated. 

2 - Enabling Price-Responsive Demand Management in 

Electricity Markets 

Hung-po Chao, ISO New England, hchao@iso-ne.com 

The traditional approach to demand response suffers from the missing property right 

problem that could undermine electricity market efficiency. A multi-settlement 

retail tariff solves the problem by allowing each customer to establish a contractbased 

baseline through demand subscription before joining a demand response 

program. A two-settlement system with demand subscription and dynamic default 

rate facilitates price-responsive demand for general consumer benefits. 

3 - An ADP Approach to Decomposing Smart Grid Pricing Problems 

Canan Uckun, University of Chicago, 5807 South Woodlawn 

Avenue, Chicago, IL, 60637, United States of America, 

cuckun@chicagobooth.edu, Dan Adelman 

In the electricity smart grid, millions of homes will receive price signals 

simultaneously. We can formulate the problem of optimizing prices through time as 

a dynamic programming problem, containing state information for each home. 

However, the problem has such large scale that it is intractable using standard 

solution methods. We propose a new decomposition approach to solving this 

dynamic program approximately. 

■ WC20 


Pricing and Revenue Management II 


Chair: Oben Ceryan, Ross School of Business, University of Michigan, 

701 Tappan St, Ann Arbor, MI, 48109, United States of America, 

oceryan@umich.edu 

1 - Dynamic Revenue Management with Nonlinear Pricing 

Wei Wei, Case Western Reserve University, Dept of Operations, 

Weatherhead School, 10900 Euclid Ave, Cleveland, OH, 44106-7235, 

United States of America, wei.wei@case.edu, Matthew J. Sobel 

We compare linear pricing with multi-part tariffs in a class of dynamic revenue 

management models with stochastic iso-elastic demand. Each period a firm sets 

multiple prices corresponding to a multi-part tariff and decides how much inventory 

to hold back for sale later. There is a nearly explicit myopic optimum that can be 

computed easily. The results exploit the homogeneity of an associated Markov 

decision process. We illustrate the results with a numerical example. 


416 

2 - Pricing and Timing of New Version Releases in the Presence of 

Strategic Consumers 

Shubin Xu, Lundquist College of Business, University of Oregon, 

Eugene, OR, 97403, United States of America, sxu@uoregon.edu, 

Michael Pangburn 

We study a firm offering successive product versions. Due to ongoing R&D or 

improving technology, each new version implies an opportunity cost associated with 

continued use of the prior product. The firm decides the time between successive 

introductions, and price. In turn, consumers strategically choose whether to 

purchase or wait for a later version. We consider the firm’s profit maximizing policy 

assuming a homogeneous market and then extend the analysis to address consumer 

heterogeneity. 

3 - Integrating Capacity Control Concepts into the Locate-to-Order- 

Systems of Automotive Manufacturers 

Thomas Volling, Technische Universität Braunschweig, Institute of 

Automotive Management and Industrial Production, 

Katharinenstrasse 3, Braunschweig, 38106, Germany, 

t.volling@tu-bs.de, Thomas S. Spengler 

Lead times in the automotive industry exceed customers’ expectations for the 

delivery of new cars. As a consequence, most manufacturers have established hybrid 

order fulfillment strategies combining elements from build-to-order and build-tostock 

(BTS) production. The focus of the contribution is on the BTS share of 

vehicles. A capacity control is developed to support the allocation of preconfigured 

cars to customer requests. Assignments are evaluated based on product specific 

opportunity costs. 

4 - Dynamic Optimal Design for Sequential Online Auctions 

Xi Chen, University of Washington, Box 352650, Industrial 

Engineering, Seattle, 98195, United States of America, 

chenxi07@uw.edu, Archis Ghate, Arvind Tripathi 

Retailers often conduct a sequence of online auctions to sell identical items as a 

revenue generation and inventory management tool. We show that under a second 

order condition on the single- auction expected revenue function, a threshold policy 

is optimal for inventory scrapping and a monotone staircase with unit jump policy is 

optimal for lot sizing. This condition is met in all common auction mechanisms. We 

investigate an extension where the minimum bid is also optimized. 

5 - Product Upgrades and Pricing with Strategic Consumers 

Oben Ceryan, Ross School of Business, University of Michigan, 

701 Tappan St, Ann Arbor, MI, 48109, United States of America, 

oceryan@umich.edu, Ozge Sahin, Izak Duenyas 

We consider a firm that allows upgrading of customers purchasing a lower quality 

product to a higher quality product if there is excess demand for the former and 

excess capacity for the latter. We investigate the optimal pricing and capacity 

decisions in the presence of strategic consumers that consider product prices as well 

as upgrade possibilities when choosing which product to purchase. 

■ WC21 


Service Contract and Incentive Design 



Chair: Hermann Jahnke, Professor, Bielefeld University, 

Universitaetsstrasse 25, Bielefeld, 33615, Germany, hjahnke@wiwi.unibielefeld.de 

1 - After-sales Contract Analysis for Service Supply Chains 

Dong Li, Rotterdam School of Management, Erasmus University, P.O. 

Box 1738, Rotterdam, Netherlands, dli@rsm.nl, Yugang Yu, Nishant 

Mishra, Xinguo Ming 

The multiple types of contract of the service supply chain may cause conflicting 

incentives. We use a game-theoretic frame to model the behavior of a service supply 

chain, and compare the different type of contracts and show that how the 

parameters affect the optimal solutions. We also find out that the inefficiency can be 

coordinated with a revenue-sharing performance based contract. 

2 - Incentive Design in Industrial Product Service Systems: 

A Simulation Study 

Partha Datta, Assistant Professor, IIM Calcutta, Diamond Harbour 

Road, Joka, Calcutta, 700104, India, parthapriya.datta@gmail.com 

Powerful incentives and risks are normally used in industrial service contracts to 

transfer risks to measure compliance with performance measures. This paper studies 

the uncertainty in service delivery driven by the agreed contract type and incentive 

mechanism using agent based discrete event simulation model under multiple 

scenarios.

3 - Lower Price Limits for Flat-fee Service Contracts Under Risk 

Hermann Jahnke, Professor, Bielefeld University, Universitaetsstrasse 

25, Bielefeld, 33615, Germany, hjahnke@wiwi.uni-bielefeld.de, 

Jan Thomas Martini 

Many manufacturers of capital equipment offer services under flat-fee service 

contracts. We address the determination of lower price limits for such contracts. 

Under these contracts, the service providers assume part of the customer’s risk. We 

focus on the impact this risk has on price limits. Our modeling tool, almost 

stochastic dominance, allows us to examine decision making under risk without 

precisely knowing the decision makers’ risk preferences as well as a multi-person 

decision context. 

4 - Towards a Theory of Service Improvisation 

Enrico Secchi, Clemson University, 128 Cochran Rd. Apt. 1, 

Clemson, SC, 29631, United States of America, 

esecchi@clemson.edu, Aleda Roth 

This paper examines the role of improvisation in the context of service delivery 

systems. Drawing from organizational improvisation literature and service 

operations and marketing, we develop antecedents and consequences of service 

improvisation. First, we define the concept of service improvisation. Second, we 

highlight the importance of the interplay between planning and execution. Finally, 

we develop a theoretical link between service delivery and the emergence of service 

innovations. 

■ WC22 


Service System Design and Effectiveness 



Chair: Adelina Gnanlet, Assistant Professor, California State University, 

Fullerton, 800 N. State College Blvd, Dept of Mgmt, Fullerton, CA, 

92832, United States of America, agnanlet@fullerton.edu 

1 - Impact of Labor and Capacity Flexibilities on Quality and a 

Financial Performance of Hospitals 

Adelina Gnanlet, Assistant Professor, California State University, 

Fullerton, 800 N. State College Blvd, Dept of Mgmt, Fullerton, CA, 

92832, United States of America, agnanlet@fullerton.edu, 

Muge Yayla-Kullu, Chris McDermott 

To reduce costs and meet variable demand, service firms frequently cross-train 

employees and use flexible capacity in capital intensive service firms. Higher crosstraining 

is cost-effective but may not provide adequate quality of service due to 

learning effects. Flexible capacity may not be conducive to provide highest level of 

quality for certain demand segments. We determine the effects of cross-training and 

flexible capacity on quality and financial performance of hospitals. 

2 - Impact of Task Complexity on Productivity in Professional Services 

Anil Akpinar, IE Business School, C/ Maria de Molina 12, Madrid, 

28006, Spain, aakpinar.PhD2010@alumno.ie.edu, Fabrizio Salvador 

In this paper we explore the effect of task complexity on the flexibility efficiency 

trade-off in knowledge worker productivity. Using a longitudinal data from one of 

the largest multinational technology and consulting firm, we provide empirical 

evidence that while specialization and variety jointly drives productivity, their 

effects are quite distinct for varying levels of task complexity. 

3 - More with Less - Service Resource Scheduling by Time 

Capacitated Splits 

Pasi Porkka, Assistant Professor, Aalto University School of 

Economics, P.O. Box 21220, Helsinki, Fin-00076, Finland, 

porkka@hse.fi 

The balancing of resource time used for production or services and for capacity 

consuming set-ups is critical for the realistic planning of high capacity utilization. 

We combine the allocation of shared resources, the time-based splitting of tasks and 

variable set-ups in mobile service operations. The potential for substantial capacity 

time savings is demonstrated. Extensions and solution approaches for realistic 

applications are discussed. 

4 - The Impact of Service Quality Variation on Service Quality, 

Operational Efficiency, and Performances 

Hong-il Kim, PhD Candidate, Korea University Business School, 

Anam-dong, Seongbuk-gu, Seoul, 136-701, Korea, Republic of, 

itlime@korea.ac.kr, Hosun Rhim, Shijin Yoo, Daeki Kim 

We investigate how service quality variation affects service quality, operational 

efficiency, and business performances. Data of branch operation in a retail bank is 

collected. Perceived service quality of customers is surveyed with SERVERPF 

questionnaire. HLM (Hierarchical Linear Modeling) and DEA (Data Envelopment 

Analysis) are used. 


417 

■ WC23 


Service Management and Virtual Enterprise 



Chair: Munish Goyal, Research Staff Member, IBM Research, AC2 L1, 

ISB, GACHIBOWLI, Hyderabad, AP, 500032, India, 

mungoyal@in.ibm.com 

1 - Value Based Dynamic Resource Allocation in a Service Cloud 

Munish Goyal, Research Staff Member, IBM Research, AC2 L1, ISB, 

GACHIBOWLI, Hyderabad, AP, 500032, India, 

mungoyal@in.ibm.com, M Rammohan Rao 

Cloud computing is a pool of virtualized computer resources which can be 

dynamically added or removed in response to changing business demands while 

meeting service level agreements at the minimal energy or operational cost. In this 

work, we develop relative value based dynamic resources allocation strategies where 

a unit of resource is allocated to a customer request with the highest value above 

the energy value threshold at any time. Algorithm is supported with numerical 

results. 

2 - Negotiation Based Completion Risk Management for 

Virtual Enterprise 

Min Huang, Professor, Northeastern University, Box 135#, 

Northeastern University, Shenyang, 110004, China, 

mhuang@mail.neu.edu.cn, Hongyu Jiang, W.H. Ip, Qing Wang, 

Xingwei Wang 

In the view of the distribution feature of decision-making in a virtual enterprise, a 

novel decision-making framework based-on negotiation is proposed for the 

completion risk management of VE. Under this framework, according to the 

characteristics of the problem, the evaluation mechanism of the owner is designed 

based on PERT, and then the concession tactic is proposed. The example analysis 

shows that this framework can achieve effective risk management. 

3 - Service Parts Inventory Control Under Obsolescence 

Cerag Pince, PhD Candidate, Erasmus University, Burg. Oudlaan 50, 

Rotterdam, 3000 DR, Netherlands, pince@few.eur.nl, 

Rommert Dekker, Hans Frenk 

We consider a single location inventory system of a slow moving item where 

Poisson demand rate drops to a lower level at a known future time. Under the 

assumptions of full backordering and fixed lead time, we incorporate obsolescence 

into a one-for-one policy with the option to reduce the base stock level in advance. 

We show that when obsolescence can be foreseen, early adaptation of base stock 

levels leads to important savings. 

■ WC24 


Decision Making for Wildfire Response and Evacuation 

Sponsor: Public Programs, Service and Needs 


Chair: Nada Petrovic, Graduate Student, Physics Department, 

Broida Hall, UC Santa Barbara, Santa Barbara, CA, 93106-9530, 

United States of America, petrovic@physics.ucsb.edu 

1 - Dynamic Resource Allocation in Wildfire Suppression 

Nada Petrovic, Graduate Student, Physics Department, Broida Hall, 

UC Santa Barbara, Santa Barbara, CA, 93106-9530, United States of 

America, petrovic@physics.ucsb.edu, David Alderson, Jean Carlson 

Wildfire response demands dynamic decision tools because fires and suppression 

evolve simultaneously. Time delays can lead to larger fires and thus greater demand 

for resources. We capture this tension using a queuing model that treats fire 

progression as a birth and death process, with rates that incorporate intrinsic fire 

dynamics and suppression. Using this framework we explore trade-offs in 

effectiveness and time delay of response. 

2 - A Space-Time Flow Optimization Model for 

Neighborhood Evacuation 

David Alderson, Assistant Professor, Naval Postgraduate School, 

Naval Postgraduate School, 1411 Cunningham Road, Monterey, CA, 

93943, United States of America, dlalders@nps.edu, 

William Langford 

We model the evacuation of vehicles in a residential neighborhood using a spacetime 

network flow representation. Our model solves for “best case” evacuation 

routes and clearing times, as could be identified and implemented by a central 

authority. Our models are large but can be solved efficiently and quickly. We apply 

this model to the Mission Canyon neighborhood near Santa Barbara, California, and 

contrast our results to a previous simulation-based study.

WC25 

3 - Influence of Information Networks on Collective 

Evacuation Dynamics 

Danielle Bassett, Postdoctoral Research Associate, University of 

California Santa Barbara, 6213 Broida Hall, Santa Barbara, CA, 

93106, United States of America, dbassett@physics.ucsb.edu, Jean 

Carlson, David Alderson 

The collective behavior of humans during an evacuation is a poorly understood, 

complex phenomenon which we model as a combined centralized-decentralized 

consensus problem, influenced by information flow over layers of technological, 

social, and geographic networks. An individual’s belief regarding disaster severity is 

constantly updated until reaching a decision threshold. We describe differential 

dynamics over these layers, indicating sensitivity of human decision-making to 

information origin. 

4 - Making Emergency Evacuation Decisions with 

Uncertain Information 

Emily Craparo, Naval Postgraduate School, Glasgow Hall Room 226, 

Monterey, CA, United States of America, emcrapar@nps.edu, 

David Alderson, Jean Carlson 

In emergency situations, time-critical decisions must be made based on uncertain 

information. Situational awareness is improved through additional observation of 

the threat; however, observation delays action. We consider an individual decisionmaker 

who faces an uncertain threat and who must decide when (and whether) to 

perform a costly evacuation. We model this evacuation decision problem using 

dynamic programming and establish optimal evacuation policies under a variety of 

cost models. 

■ WC25 


Transportation, Planning II 


Chair: Chi Xie, Research Fellow, The University of Texas at Austin, 


chi.xie@mail.utexas.edu 

1 - Composite Variable Formulation for Truckload Relay 

Network Design 

Hector Vergara, Graduate Research Assistant, University of Arkansas, 

4207 Bell Engineering Center, Fayetteville, AR, 72701, 

United States of America, hvergara@uark.edu, Sarah Root 

Full truckload (TL) trucking usually considers a Point-to-Point dispatching method. 

Alternatively, multi-zone dispatching under a network configuration of relay points 

can be used to improve driver retention. We propose a new mathematical 

formulation for strategic relay network design that places relay points and 

determines driver routes. Our model minimizes total costs while considering 

operational constraints such as driver tour length and load circuity within the 

variable definition. 

2 - An Optimal Cycle Length Model for Feeder Transit Services 

Shailesh Chandra, Student, Department of Civil Engineering, Texas 

A&M University, 3136 TAMU, College Station, 77843-3136, United 

States of America, chandrashailesh@gmail.com, Chung-Wei Shen, 

Luca Quadrifoglio 

A simulation based generic model has been derived for estimating optimal cycle 

lengths of a demand responsive transit “feeder” services. For input parameters such 

as the shape size and daily demand density of an area, the success of this model has 

been validated by a case study. 

3 - Transportation Planning for Squatter Developments 

Hani Al-Naghi, PhD Candidate, American University of Beirut, 

P.O.Box 11-0236, Riad El Solh, Beirut 11, Beirut, Lebanon, 

haa31@aub.edu.lb, Nabil Nehme 

This paper addresses urban transportation planning issues related to squatter 

developments and their integration with the surrounding areas. A general 

framework encompassing all criteria related to social, economics, land use and 

political aspects, is formulated and applied to selected case studies in Lebanon. 

4 - A Primal-dual Mathematical Programming Framework for Traffic 

Assignment Problems 

Chi Xie, Research Fellow, The University of Texas at Austin, 


chi.xie@mail.utexas.edu, Travis Waller 

Traffic assignment problems have been formulated as mathematical programs, 

variational inequalities, complementarity systems, and fixed-point models. It is well 

known that the last three modeling techniques all provide a common functional 

form for modeling different traffic assignment problems. This talk presents a 

uniform primal-dual mathematical programming framework, which can be used to 

accommodate a variety of traffic assignment problems. 


418 

■ WC26 


Interface Between Efficient Data Collection and 

Flexible Data Modeling 

Sponsor: Data Mining 


Chair: Peter Qian, University of Wisconsin-Madison, 1300 University Ave, 

Madison, WI, 53706, United States of America, peterq@stat.wisc.edu 

1 - Improvement on Cross-Validation via Sliced Statistical Design 

Xinwei Deng, Visiting Assistant Professor, University of Wisconsin- 

Madison, Department of Statistics, 1300 University Ave., Madison, 

WI, 53706, United States of America, xdeng@cs.wisc.edu, Peter Qian 

A training data is often used to construct models to predict the future response. The 

cross-validation is a traditional method to compare different models and assess their 

prediction performance. It can provide a nearly unbiased estimate of prediction 

error, but can be with high variability. In this work, we proposed a novel sliced 

statistical design strategy to improve the performance of cross-validation, 

substantially outperforms the usual method in various classification problems. 

2 - Regularized REML for Estimation and Selection of Fixed and 

Random Effects in Linear Mixed-Effects Mo 

Sijian Wang, Assistant Professor, University of Wisconsin, Madison, 

1300 University Ave., Madison, WI, 53706, United States of America, 

wangs@stat.wisc.edu 

In the practice of LMM, inference on the structure of random effects component is 

of great importance not only to yield proper interpretation of subject specific effects 

but also to draw valid statistical conclusions. In this paper, we propose a novel 

method of regularized restricted maximum likelihood to select fixed and random 

effects simultaneously in the LMM. We also investigate large sample properties for 

the proposed estimation, including the oracle property. 

3 - Consistent Selection of the Number of Clusters via 

Clustering Stability 

Junhui Wang, University of Illinois at Chicago, 851 S Morgan St, 

Chicago, IL, 60607, United States of America, jwang@math.uic.edu 

In cluster analysis, one major challenge is to estimate the number of clusters. In this 

talk, I will present a novel selection criterion that is applicable to all kinds of 

clustering algorithms. The key idea is to select the number of clusters such that the 

resulting clustering algorithm has the smallest instability, which measures its 

robustness against the sampling randomness. Numerical examples and asymptotic 

selection consistency will be discussed. 

4 - Optimal Supersaturated Design for Variable Selection via Lasso 

Dadi Xing, PhD Student, Purdue University, 224-7 Arnold Drive, 


hwan@purdue.edu, Hong Wan, Yu Zhu 

In the supersaturated design(SSD) study, most existing criteria for constructing 

optimal SSD are motivated and further justified from the estimation perspective. We 

will propose a number of optimality criteria for the construction of SSD from the 

perspective of variable selection with Lasso. The properties of these criteria will be 

discussed. A computing algorithm will be used to construct such optimal SSD, and 

examples of simulation and real applications will also be presented. 

■ WC27 


Network Optimization II 


Chair: Nan Jiang, Student, University of Texas, Austin, 1 University 

Station C1761, ECJ 6.2, Austin, TX, 78712, United States of America, 

njiang@mail.utexas.edu 

1 - A Reformulation-Linearization Technique for the Two-level Facility 

Location Problem 

Youngho Lee, Korea University, Sungbuk Ku Anam Dong 5-1, Seoul, 

Korea, Seoul, Korea, Republic of, shadowpp@korea.ac.kr, 

Gigyoung Park 

In this paper, we deal with an two-level facility location problem based on the tree 

topology. In particular, we develop a model for this problem and apply the 

reformulation-linearization technique (RLT) to construct various enhanced 

tightened versions of the proposed model. And we derive necessary and sufficient 

conditions for a family of some inequalities to be facet-defining.

2 - Continous Network Design Problem with Emission Constraint 

Nan Jiang, Student, University of Texas, Austin, 1 University Station 

C1761, ECJ 6.2, Austin, TX, 78712, United States of America, 

njiang@mail.utexas.edu, ManWo Ng, Travis Waller 

In this paper, a traffic network design model with emission constraint and its 

solution method are presented. The resulting solutions are a set of capacity 

improvements to a given network, for a given demand, subject to user-specified 

budget constraints and emission constraint and resulting in minimal system total 

cost. Application results prove this model decreases system emission and provide 

information useful for planning road network improvements under air quality 

constraints. 

■ WC28 


Manufacturing Process Optimization 



Chair: Roshan Vengazhiyil, Coca-Cola Associate Professor, Georgia 

Institute of Technology, Industrial and Systems Engineering, Atlanta, GA, 

30332, United States of America, roshan@isye.gatech.edu 

1 - An Interactive Method to Multiresponse Surface Optimization Based 

on Pairwise Comparisons 

Dong-Hee Lee, Pohang University of Science and Technology, 

Department of Industrial and Management, Pohang, Korea, Republic 

of, princeps@postech.ac.kr, Kwang-Jae Kim, Murat Köksalan 

In multiresponse surface optimization, responses are often in conflict. To obtain a 

satisfactory compromise, the preference information of a decision maker (DM) on 

the tradeoffs among the responses should be incorporated into the problem. We 

propose an interactive method where the DM provides preference information in 

the form of pairwise comparisons. The results of pairwise comparisons are used to 

estimate the preference parameter values in an interactive manner. The method is 

effective in that a highly satisfactory solution can be obtained. 

2 - Analysis of Computer Experiments with Functional Response 

Ying Hung, Rutgers, Camden, NJ, United States of America, 

yhung@stat.rutgers.edu, Roshan Vengazhiyil, Shreyes Melkote 

We develop an efficient implementation of kriging for analyzing functional 

responses. The main contribution of this paper is to develop a two-stage model 

building procedure and a general framework which can be used irrespective of the 

data structure. The methodology is illustrated using a computer experiment 

conducted for optimizing residual stresses in machined parts. 

3 - New Variable Selection Methods Under Engineering 

Inequality Constraints 

Hin Kyeol Woo, Georgia Institute of Technology, 765 Ferst Drive, 

Atlanta, GA, 30329, United States of America, hinkyeol@gatech.edu, 

Andres Hernandez, Jye Chyi Lu, Martha Grover 

This presentation discusses new variable selection methods under engineering 

inequality constraints. A case study of nanoparticle synthesis with a solubility 

constraint motivates and illustrates the research. Simulation experiments explore 

properties of the proposed method. 

■ WC29 


Maintenance Management and Service Logistics 



Chair: Alaa Elwany, Assistant Professor, Eindhoven Unviersity of 

Technology, Eindhoven, Netherlands, elwany@tue.nl 

Co-Chair: Nagi Gebraeel, Associate Professor, Georgia Tech, 765 Ferst Dr., 

Atlanta, United States of America, nagi.gebraeel@isye.gatech.edu 

1 - Modified Block Replacement Can Outperform Age Replacement in 

Terms of Spare Parts Ordering 

Rommert Dekker, Erasmus University Rotterdam, Burg Oudlaan 50, 

Rotterdam, 3062 PA, Netherlands, rdekker@ese.eur.nl 

It is well known that age replacement outperforms block replacement as better 

information on the likelihood of failures is used. Yet block replacement allows a 

more appropriate ordering of spare parts as replacements can be planned in time. In 

this presentation we present a study which shows that the better spare parts 

planning can offset the advantages of age replacment. The study also reveals what 

kind of inventory policies are useful in this respect. 


419 

2 - Upgrading Policy After Redesign of a Component for 

Reliability Improvement 

Kurtulus Oner, Assistant Professor, Eindhoven University of 


k.b.oner@tue.nl, Geert-Jan Van Houtum, Gudrun P. Kiesmüller 

We introduce a model for studying the following two upgrading policies that an 

OEM may implement after the redesign of a component: (i) Upgrade all 

preventively at time 0, (ii) Upgrade one-by-one correctively. We develop a problem 

formulation for the comparison of the two policies and perform exact analysis. We 

conduct a numerical study and derive insights on the optimality of the policies. 

3 - Component Reliability Criticality or Importance Metrics for Systems 

with Degrading Components 

David Coit, Associate Professor, Rutgers University, Industrial & 

Systems Engineering, 96 Frelinghuysen Rd., Piscataway, NJ, 08854, 

United States of America, coit@rutgers.edu, Hao Peng, Qianmei Feng 

This paper proposes new importance measures (IMs) for systems with either 

independent or dependent degrading components. As functions of time, the 

proposed IMs can provide timely feedback on the critical components based on the 

observed degradation. The correlation between components and the dependency of 

failure thresholds are considered through multivariate distributions. Numerical 

examples show that the proposed IMs are effective in assessing criticality of 

degrading components. 

4 - Maintenance Policy Based on MultiCriteria Decision Aiding 

Cristiano Cavalcante, Federal University of Pernambuco, Caixa 

Postal, 5125, Cep 52070960, Recife, PE, 52070960, Brazil, 

cristianogesm@gmail.com, Adiel Almeida 

Maintenance planning is very sensitive to characteristics of the system, the context 

and the objectives of the decision maker. Regarding this problem, the most common 

analyses consist of evaluating the cost rate function. But, in some specific contexts 

the consequence of failures has distinct dimensions, which are difficult to represent 

by only one criterion (monetization). Thus, we discuss models based on MCDA 

(MultiCriteria Decision Aiding) approach, in order to support maintenance 

planning. 

■ WC30 


Statistics/Quality Control II 


Chair: Xuan Huang, Assistant Professor, University of Alabama at 

Birmingham, 345 Lincoln Ave., Apt 115, Amherst, MA, 01002, United 

States of America, xuan@som.umass.edu 

1 - Comparison of Ozone Levels Among Five Cities using 

Control Charts 

Gautam Eapi, PhD Student, University of Texas at Arlington, Civil 

Engineering, Box 19308, Arlington, TX, 76019, United States of 

America, gautam.raghavendra@gmail.com, Melanie Sattler, 

Mostafa Ghandehari 

Ozone is one of the six criteria pollutants, as specified by the USEPA. Statistical 

quality control can be helpful in improving quality. In this paper, data for ozone 

levels of five different cities across the U.S. is analyzed by using control 

charts.Control charts are used to compare the cities. In addition, out of control data 

points are discussed. 

2 - A Bayesian Approach to Estimating Market Implied Risk 

Neutral Densities 

James Delaney, Assistant Professor, Temple University, Department 

of Statistics (006-12), 1810 N 13th Street, Philadelphia, PA, 19122- 

6012, United States of America, james.delaney@gatech.edu, 

Marc Sobel 

Contemporaneous prices of financial derivatives provide much information about 

the so-called “market implied risk neutral distribution” (MIRND) of the security that 

underlies those derivatives. Here we propose a Bayesian model to provide the 

regularity necessary for estimating a MIRND that corresponds to a set of options’ 

prices. We provide details on a much more generally useful technique for simulating 

from the very complex posterior distribution of this model’s parameters. 

3 - Orthogonal Polynomial and Saddle Point Approximations for Sums 

of Non-identical Binomial Random Variables 

Aysun Taseli, Research Assistant, Northeastern University, 

360 Huntington Avenue, 334 Snell Engineering, Boston, 02115, 

United States of America, aysunt_qpl@yahoo.com, James Benneyan 

We compare performance and discuss relative advantages of saddle point 

approximations (SPA) and cumulant based orthogonal polynomial expansions for 

estimating the convolution of non-identical binomial distributions. Some important 

healthcare and service applications of this distribution involve methods that require 

repetitive computation of probabilities. Both SPA and a normalized Gram-Charlier 

expansion are shown to be accurate and fast compared to the exact PDF and Monte 

Carlo estimation.

WC31 

4 - Beta Model-based Control Chart for Fraction Monitoring with 

Correlated Process Variables 

Michel Anzanello, Professor, Federal University of Rio Grande do 

Sul, Av. Osvaldo Aranha, 99, Porto Alegre, 90.035-190, Brazil, 

michel.anzanello@gmail.com, Angelo Sant’Anna, Carla ten Caten 

Although widely used to monitor processes where quality characteristics vary with 

adjustments in control variables, model-based control charts’ efficiency is 

jeopardized by high correlated control variables. Our method integrates Principal 

Components Analysis to Beta model-based control charts to overcome such 

limitation. Sensitivity Analysis using Monte Carlo simulation validates the method. 

5 - Dimension Reduction of Multivariate Autocorrelated Processes 

Xuan Huang, Assistant Professor, University of Alabama at 

Birmingham, 345 Lincoln Ave., Apt 115, Amherst, MA, 01002, 

United States of America, xuan@som.umass.edu 

In traditional multivariate literature, Principal Components Analysis (PCA) is the 

standard tool for dimension reduction. For autocorrelated processes, however, PCA 

fails to take into account the time structure information. It is arguable that PCA is 

still the best choice. In this presentation I propose an enhanced dimension reduction 

method which by design takes into account both cross-correlation and 

autocorrelation information. I demonstrate it through case studies and simulations. 

■ WC31 


Health Care, Therapy and Treatment 


Chair: Chunhua Men, University of California, San Diego, 3855 Health 

Sciences Dr. #0843, La Jolla, CA, 92093, United States of America, 

cmen@ucsd.edu 

1 - Identifying and Quantifying Protein Values for Obtaining 

Cancer Biomarkers 

Chaitra Gopalappa, PhD Student, University of South Florida, 4202 E 

Fowler Ave, Tampa, United States of America, chaitrag@gmail.com 

Identifying proteins that are only produced in the cancerous state of cells (cancer 

biomarkers) can lead to its use as a diagnostic tool for early detection of cancer. The 

task prior to detecting biomarkers (proteins that distinguish cases from controls), 

i.e., identifying and quantifying the amount of all proteins, is an analytically 

challenging task that requires mathematical models. We present the analytical 

challenges and the mathematical models. 

2 - Retrofitting Tissue and Cell Banking: Best Practices and Emerging 

Business Models 

Katrina Nordstrom, Professor, Aalto University School of Science and 

Technology, Department Biotechnology and Chemical technol, 

Kemistintie 1A 16100 Aalto, Espoo, Finland, 

katrina.nordstrom@tkk.fi, Marko Narhi, Petri Lehenkari, 

Ari P.J. Vepsalainen, Olli Natri, Mika Pietila 

The study explores “retrofitting” of blood banking by development of business 

models based on best practices for collection, processing, storage and delivery of 

cells and tissues. Plausible business models examine production of bone marrow, 

umbilical cord blood stem cells and hematopoietic stem cells. A framework is also 

specified for classifying and evaluating the capabilities and hurdles of supply chains 

for living products for safe and traceable future products and therapies. 

3 - Turkey Disposal and Recycling Network Design Model for 

Drug Industry 

Ayse Gunes, Assistant Specialist, Industrial Engineer, Scientific and 

Technical Research Council of Turkey (TUBITAK), 06100 

Kavaklidere, Ankara, Turkey, ayse.gunes@tubitak.gov.tr, 

Bahar Ozyoruk 

In Turkey, to determine the tendency of people, we first have made a survey related 

to drug use, recycling, etc. with military personnel. We have developed a model that 

provides waste disposal and recycling of paper in the minimum cost. In the 

developed model, human and environmental health, hazardous waste disposal and 

recycling for reuse are discussed and the model are solved with GAMS program. 

This study includes case study, it is important for healthcare logistics and global 

health topics. 

4 - Linearity Effects in Brachytherapy Treatment Planning 

Asa Holm, PhD Student, Linköpings Universitet, Matematiska 

institutionen, Linköpings universitet, Linköping, 58183, Sweden, 

asa.holm@liu.se, Torbjörn Larsson, Asa Carlsson Tedgren 

Modern optimization techniques for inverse planning of HDR brachytherapy makes 

it possible to efficiently calculate dose plans. On of the tenets of such techniques is 

the use of linear penalty functions. Plans generated with these techniques tend to 

have a few dwell positions that dominate the solution, however physicians prefer 

homogeneous plans. In this talk we show that one reason for the long dwell times is 

the linear penalties and introduce a solution that reduces the effects of linearity. 


420 

5 - Optimization Models for Online Adaptive Radiotherapy 

Chunhua Men, University of California, San Diego, 3855 Health 

Sciences Dr. #0843, La Jolla, CA, 92093, United States of America, 

cmen@ucsd.edu, Steve Jiang 

Traditional treatment plan optimization models based on a snapshot of the patient’s 

anatomy prior to treatment are not suitable for online adaptive radiotherapy (ART) 

which allows real-time treatment adaptations based on the current patient anatomy. 

In this work, we develop and evaluate various optimization models for online ART. 

To obtain real-time treatment plans, we implement the algorithms on GPU. Tests on 

clinical cancer cases showed the effectiveness and the efficiency of these models. 

■ WC32 


Computational Stochastic Programming 



Chair: Shabbir Ahmed, Associate Professor, Georgia Institute of Tech, 

765 Ferst Drive NW, Atlanta, GA, United States of America, 

sahmed@isye.gatech.edu 

1 - A Preconditioning Technique for Schur Complement Systems 

Arising in Stochastic Optimization 

Mihai Anitescu, Computational Mathematician, Argonne National 

Laboratory, Math and Computer Science Division, 9700 S Cass Ave, 

Argonne, IL, 60439, United States of America, anitescu@mcs.anl.gov, 

Cosmin Petra 

We discuss a parallel interior-point method for stochastic programming that uses a 

Schur complement mechanism. We propose a stochastic preconditioner to improve 

scalability. The spectral analysis of the preconditioned matrix indicates an 

exponential clustering of the eigenvalues around 1. The numerical experiments 

performed on the relaxation of a unit commitment problem show good 

performance, in terms of both the accuracy of the solution and the execution time. 

2 - Models and Formulations for Optimization with Multivariate 

Stochastic Dominance Constraints 

James Luedtke, University of Wisconsin-Madison, 1513 University 

Av., Madison, WI, United States of America, jrluedt1@wisc.edu, 

Benjamin Armbruster 

Multivariate stochastic dominance constraints provide an interesting modeling tool 

for problems having multiple stochastic objectives. Recently proposed models use 

extensions of the notion of positive linear stochastic dominance, but appear 

computationally challenging to use. We propose to use a different notion of 

dominance, based on expected utility theory, and present linear and integer 

programming formulations for the corresponding problem. 

3 - Learning Price Functions in Cournot Games 

Uday Shanbhag, Asst. Professor, University of Illinois at Urbana 

Champaign, Urbana, Il, United States of America, 

udaybag@illinois.edu, Sean Meyn, Hao Jiang 

We consider a regime where firms compete in a Nash-Cournot game without the 

knowledge of the precise parameters of the price functions. We show that the 

resulting trajectory can be characterized. Convergence of the learning update 

scheme is examined in deterministic and stochastic regimes. 

■ WC33 


Optimization Strategies for Real-World Applications 



Chair: Eva Lee, Professor & Director, Georgia Institute of Technology, 

Center for Operations Research in Medici, Industrial & Systems 

Engineeriing, Altanta, GA, 30332-0205, United States of America, 

eva.lee@gatech.edu 

1 - On the Simultaneity of Row and Column Generation 

Jon Petersen, PhD Student, Georgia Institute of Technology, 

765 Ferst Drive NW, Atlanta, GA, 30309, United States of America, 

Petersen@gatech.edu, Ellis Johnson 

While both constraint generation and column generation have been successfully 

employed in solving practical real-world problems, they are often thought of as 

being mutually exclusive. The efficacy of solving large-scale models can be improved 

by using these two principles concurrently. We propose a new method for doing so, 

and present computational results to validate our approach.

2 - Audience Space Aggregation for Ad Planning 

John Turner, University of California - Irvine, The Paul Merage 

School of Business, Irvine, CA, 92697, United States of America, 

john.turner@uci.edu 

Whether serving ads on web pages or in newer media such as video games, 

targeting constraints lead to a combinatorial explosion in the number of audience 

segments. Using an aggregation heuristic for large transportation problems, we 

allocate impressions to ad campaigns at an appropriate granularity. Computational 

results show that a little bit of disaggregation goes a long way: Near-optimal 

solutions are achieved despite a high degree of aggregation. 

3 - Constraint Optimal Selection Techniques (COSTs) for 

Linear Programming 

H. W. Corley, IMSE Department, The University of Texas at 

Arlington, Arlington, TX, 76019, United States of America, 

corley@uta.edu, Jay Rosenberger, Goh Saito 

We present a Constraint Optimal Selection Technique (COST) for efficiently solving 

large-scale nonnegative linear programming problems. We provide a geometric 

interpretation of the COST and computational comparisons with the CPLEX primal 

simplex, dual simplex, and barrier algorithms. 

■ WC34 


Optimization on Graphs 



Chair: Doug Altner, Assistant Professor, United States Naval Academy, 

United States of America, altner@usna.edu 

1 - Integer Programming Techniques for Matroid Circuit Problems 

John Arellano, PhD Student, Rice University, 6100 Main St. MS 134, 

Houston, TX, 77030, United States of America, jda2@rice.edu, 

Illya Hicks 

Although some combinatorial optimization problems associated with matroids can 

be solved in polynomial time, finding particular circuits in matroids is an NP-hard 

problem. It is related to compressive sensing and finding the degree of redundancy 

of sensor networks. In this talk, we attempt to solve these types of problems to 

optimality using integer programming techniques and present computational results. 

2 - Over Restriction of Network Expansion Big M Constraints 

Kael Stilp, Georgia Institute of Technology, 765 Ferst Drive, NW, 

Atlanta, GA, United States of America, mstilp3@isye.gatech.edu, 

Ozlem Ergun, Pinar Keskinocak 

We discuss a multi period network expansion model where flow capacities are big- 

M constrained and nodes have supply and demand of commodities. The objective is 

to minimize unsatisfied demand over all of the periods. We show computational 

results behind restricting the M value to infeasible values as a means of speeding up 

computation and achieving better solutions. To further understand the results we 

discuss theoretical reasonings for the occurrence. 

3 - Coverings and Matchings in r-Partite Hypergraphs 

Doug Altner, Assistant Professor, United States Naval Academy, 

United States of America, altner@usna.edu, Paul Brooks 

We present a few results regarding matchings and coverings in r-partite 

hypergraphs. First, we present an alternate proof showing the integrality gap of the 

standard BILP for r-dimensional matching is at least r-k where k is the smallest 

positive integer such that r-k is a prime power. Second, we prove r-dimensional 

covering is NP-hard for intersecting hypergraphs. Third, we prove a few upper 

bounds on the covering number of a special class of intersecting hypergraphs that 

are not balanced. 


421 

■ WC35 


Advances in Anomalous Diffusion III 



Chair: Mark Meerschaert, Michigan State University, Department of 

Statistics and Probability, East Lansing, MI, United States of America, 

mcubed@stt.msu.edu 

1 - Approximation of Tempered Operator Stable Processes 

Boris Baeumer, University of Otago, Department of Mathematics and 

Statistics, Dunedin, New Zealand, bbaeumer@maths.otago.ac.nz, 

Mihaly Kovacs 

Operator stable processes are characterised by the scaling matrix H and mixing 

measure M. By randomly choosing a direction and then generating a onedimensional 

jump distance, the resulting process lies in the domain of attraction of 

the operator stable process. All tempered processes lie in the domain of attraction of 

a Gaussian. We show that for some tempered op-stable processes the speed of 

convergence of the approximant to the tempered process is faster than the 

convergence to the Gaussian. 

2 - Cauchy Problems Solved by Running Subordinate Processes 

Erkan Nane, Assitant Professor, Auburn University, 221 Parker Hall, 

Auburn, AL, 36849, United States of America, ezn0001@auburn.edu 

Subordinated Markov processes will be studied. These are obtained by taking 

Markov processes and replacing the time parameter with other processes such as 

Brownian motion, symmetric stable process, an inverse of a stable subordinator, or 

local time of an stable process of index between 1 and 2. We obtain frational 

Cauchy problems or Cauchy problems involving the powers of the generator of the 

Markov Process by running these subordinated Markov processes. 

3 - Space-time Duality for Fractional Diffusion 

Mark Meerschaert, Michigan State University, Department of 

Statistics and Probability, East Lansing, MI, United States of America, 

mcubed@stt.msu.edu, Boris Baeumer, Erkan Nane 

Fractional diffusion equations govern scaling limits of random walk models. The 

limit process is a stable Levy motion that models the jumps, subordinated to an 

inverse stable process that models the waiting times. Using Zolotarev duality, we 

relate the density of a spectrally negative stable process with index $1

WC37 

2 - A Simple Management Tool to Increase High School Track 

Team Participation 

Sambhavi Lakshminarayanan, Assistant Professor, Medgar Evers 

College - CUNY, S-Building, Department of Business Admin, 1637 

Bedford Avenue, Brooklyn, NY, 11225, United States of America, 

sLakshminarayanan@mec.cuny.edu, Ashwin Acharya 

Convincing busy high school students to make a significant time commitment to 

participate in athletics is a challenge. However, schools and athletic coaches consider 

it highly desirable, if not necessary, for students to participate in athletic activities. 

This paper discusses an approach developed by the senior members of the track 

team at an academically focused high school to increase team numbers and 

retention. 

3 - Teaching Business Statistics using Participatory Learning Methods 

in a Multicultural Setting 

Mark Ferris, Saint Louis University, 3674 Lindell Blvd, Room 457, 

Saint Louis, MO, 63104, United States of America, ferrisme@slu.edu, 

Reuven Levary 

Teaching pedagogy in business schools emphasize participatory learning methods 

such as class discussion, group problem solving and case studies. The predominant 

teaching methods for international students may not include a participatory 

component, rather it is a system that emphasizes the teacher as the prime source of 

knowledge. A strategy for ensuring the contribution of international students in a 

participatory learning environment for statistics was developed, implemented and 

evaluated. 

4 - Undergraduate and Graduate Students’ Perception about 

Web-Enhanced and Online Courses 

Hiral Shah, Assistant Professor, St Cloud State University, 720 Fourth 

Ave S - ECC 101, St Cloud, MN, 56301, United States of America, 

hashah@stcloudstate.edu, Devang Mehta 

The purpose of this study was to compare the perceptions of students about 

completely online courses against web-enhanced courses. Data were collected from 

both graduate and undergraduate students enrolled at two different higher 

education institutions. The results of this study will enable instructors to modify 

their teaching styles using either of these methods of teaching. 

5 - What is Academic Quality? A Comparison of Traditional and Adult 

Students’ Perceptions 

Helene Caudill, Associate Professor of Management, St. Edward’s 


United States of America, helenec@stedwards.edu 

There are numerous definitions of academic quality, but the one that I believe 

affects student satisfaction and faculty reputations most often is the perception of 

what students believe to be “quality in the classroom.” With responses from over 

250 students, the results indicate that students focus on three main areas: the 

credentials of the faculty member, the usefulness of the materials and assignments, 

and the communication skills of the faculty member. 

■ WC37 


Applied Probability 


Chair: Raja Jayaraman, Postdoctoral Research Fellow, University of 

Arkansas, Industrial Engineering, Bell 4207, Fayetteville, AR, 72701, 

United States of America, rjayaram@uark.edu 

1 - The Action Gambler and Equal-sized Wagering 

David Hartvigsen, Professor, University of Notre Dame, 354 Mendoza 

College of Business, Notre Dame, IN, 46556-5646, United States of 

America, Hartvigsen.1@nd.edu 

A gambler, with a bankroll B, faces a sequence of n identical, independent, win-lose 

bets. When the total amount wagered (the total action) must be at least B, we show 

that wagering B/n on each bet maximizes the expected utility of the final bankroll 

iff the probability of winning a single bet is at most some p* (which is an explicit 

function of B, n, and the utility function). 

2 - Optimal Sequential Selection of a Unimodal Subsequence From a 

Random Sample 

Alessandro Arlotto, University of Pennsylvania, 3730 Walnut Street, 

500 Jon M. Hunstman Hall, Philadelphia, PA, 19104, United States 

of America, alear@wharton.upenn.edu, J. Michael Steele 

The length of the longest unimodal subsequence in a random sample of size $n$ is 

known to be asymptotic to $2\sqrt{2n}$. We study the sequential version of the 

same problem in which, at every decision time $k\in\{1,...,n\}$, a decision-maker 

has to select or reject the current observation in order to form a unimodal 

subsequence of maximal expected length. We show that this expected length is 

asymptotic to $2\sqrt{n}$. 


422 

3 - On Pro-rata Pricing Strategy for Extended Product Warranties 

Raja Jayaraman, Postdoctoral Research Fellow, University of 

Arkansas, Industrial Engineering, Bell 4207, Fayetteville, AR, 72701, 

United States of America, rjayaram@uark.edu, Timothy Matis 

Product warranties play a challenging and decisive role in today’s dynamic business 

environment. Several strategies and models have been proposed assuming fixed cost 

associated with various repair actions available to the manufacturer towards 

rectifying product failures. In this presentation we shall address pro-rata pricing and 

its overall effect towards minimizing expected cost for products carrying extended 

warranties. 

■ WC38 


Theory and Applications in Copositive Programming 




Chair: Samuel Burer, Associate Professor, University of Iowa, S346 

Pappajohn Business Building, Iowa City, IA, 52242-1994, 

United States of America, samuel-burer@uiowa.edu 

1 - New Approximations for Copositive Matrices 

Samuel Burer, Associate Professor, University of Iowa, S346 

Pappajohn Business Building, Iowa City, IA, 52242-1994, 

United States of America, samuel-burer@uiowa.edu, Hongbo Dong 

We introduce a new hierarchy of inner approximations of the copositive matrices. 

The distinguishing feature of the hierarchy is its recursive nature, which deals with 

smaller (but more) matrices as the recursive depth increases. For fixed depth, the 

resulting inner approximation is a polynomially sized linear-semidefinite program. 

2 - Mixed Zero-One Linear Programs Under Objective Uncertainty: 

A Completely Positive Representation 

Chung Piaw Teo, Professor, National University of Singapore, NUS 

Business School, BIZ1 8-72, NUS, 15 Kent Ridge Drive, Singapore, 

119245, Singapore, bizteocp@nus.edu.sg, Zhichao Zheng, 

Karthik Natarajan 

We analyze mixed 0-1 linear programs under objective uncertainty using a moment 

based approach, assuming descriptive statistics of the objective coefficients are 

known, but not the exact form of the distribution. Our main result shows that 

computing the supremum of the expected optimal objective value of such problem 

is a completely positive program. The result is extended to objective coefficients over 

Euclidean space, uncertain moments and more complicated objective functions. 

3 - Separating Doubly Nonnegative and Completely Positive Matrices 

Kurt Anstreicher, Professor, University of Iowa, Department of 

Management Sciences, Iowa City, IA, 52242, United States of 

America, kurt-anstreicher@uiowa.edu, Hongbo Dong 

Completely Positive (CP) matrices can be used to formulate a variety of NP-Hard 

problems. A natural issue in the optimization setting is to separate a given Doubly 

Nonnegative (DNN) but non-CP matrix from the CP cone. We describe 

constructions for such a separation that apply to 5x5 DNN but non-CP matrices, as 

well as to larger matrices with block structure. Computational results illustrate the 

ability of these procedures to generate improved bounds on difficult problems. 

■ WC39 


Column Generation in Integer Programming 

Sponsor: Optimization/Integer Programming 


Chair: Wilbert Wilhelm, Professor, Texas A&M University, Department of 

Industrial and Systems Eng, TAMUS 3131, College Station, 77843-3131, 

United States of America, wilhelm@tamu.edu 

1 - A Stabilized Dynamic Constraint Aggregation/Column Generation 

Method for the MDVSP 

Guy Desaulniers, Ecole Polytechnique de Montréal and GERAD, CP. 

6079, Succ. Centre-ville, Montréal, Canada, 

Guy.Desaulniers@polymtl.ca, Pascal Benchimol, Jacques Desrosiers 

In a column generation context, dynamic constraint aggregation that reduces the 

number of constraints in the master problem was recently introduced to reduce 

degeneracy. Dual variable stabilization is also effective at reducing the number of 

column generation iterations. In this talk, we present a method that combines both 

techniques and we report computational results for the multi-depot vehicle 

scheduling problem.

2 - Partial Path Column Generation for the Vehicle Routing Problem 

Mads Jepsen, PhD., Technical University of Denmark, 

Produktionstorvetbygn. 424, Kgs. Lyngby, 2800, Denmark, 

makj@man.dtu.dk, David Pisinger, Björn Petersen 

This paper presents a column generation algorithm for the Capacitated Vehicle 

Routing Problem (CVRP). The Set Partitioning model with elementary routes, have 

shown superior results. However, algorithms for solving the pricing problems do not 

scale well. We suggest to relax the constraint that ‘each column is a route’ into 

‘each column is a part of the giant tour’. This way, the length of the partial path can 

be bounded and a better control of the size of the solution space is obtained. 

3 - Unified Branch-Cut-and-Price for Routing and Scheduling 

Marcus Poggi de Aragao, PUC-Rio, R. M. S. Vicente 225, Rio de 

Janeiro, Brazil, poggi@inf.puc-rio.br, Eduardo Uchoa, Artur Pessoa 

We extend the 1978 Picard-Queyranne approach to routing and sheduling 

problems. The resulting Branch-Cut-and-Price adds a number of families of valid 

inequalities described on variables from the original as well as on the ones from 

extended formulations. Several variants of these problems are tackled with minor 

adaptations. We address stabilization and performance issues. A wide range of 

experimental results are presented. Finally, the generality and the evolution of the 

approach is discussed. 

4 - A Branch-and-cut Equivalent to Branch and Price 

Wilbert Wilhelm, Professor, Texas A&M University, Department of 

Industrial and Systems Eng, TAMUS 3131, College Station, 77843- 

3131, United States of America, wilhelm@tamu.edu, Deepak Warrier 

Branch and price is a leading approach to solve integer programs but suffers from 

serious shortcomings, including converging slowly, failing to indicate it if prescribes 

tighter bounds than the linear relaxation of the original problem, and failing to 

readily incorporate cutting planes to tighten bounds. This paper describes a branchand-cut 

equivalent that prescribes the same bounds but overcomes these 

shortcomings. Computational tests compare the effectiveness of the two approaches. 

■ WC40 


Decision Analysis IV 


Chair: Kash Barker, Lecturer, University of Oklahoma, 202 W. Boyd, 



1 - Optimal Pricing, Modularity Level and Consumer Return for MC 

Products using Uncertainty 

Na Liu, The Institute of Textiles and Clothing, The Hong Kong 

Polytechnic University, Hunghom, Kowloon, Hong Kong, Hong 

Kong, Hong Kong - PRC, 08900900r@polyu.edu.hk, Jason Choi 

Mass customization (MC) is a pertinent industrial practice. MC retailers can gain 

substantial advantages if return is considered. We study the optimal decisions under 

a mean-variance analytical formulation. Structural properties are revealed and the 

closed-form optimal solutions are derived. Sensitivity analysis is subsequently 

conducted to explore how the risk sensitivity and other parameters affect the 

optimal decisions. Counter-intuitive findings are obtained and insights are 

generated. 

2 - Determining the Objective-based Feasibility of Installing a Sewage 

Treatment Plant at a University 

Mario Chew, Full Time Teacher, Technological Institute of Superior 

Studies of Coacalco, Av. 16 de septiembre No. 54, Cabecera, 

Municipal, Coacalco, Edo. de Mexico, México D.F., 55700, Mexico, 

mchew@tesco.edu.mx, Verónica Velàzquez 

A typical feasibility analysis of a Waste Water Treatment Plant (WWTP) calculates 

economical or technical metrics, which are used to decide if the WWTP is to be 

installed. If the fundamental objectives of the decision maker aren’t technical or 

economical, such an analysis is not satisfactory; this is the case when a University 

contemplates acquiring a WWTP. Here, we use Keeney’s Value-Focused Thinking to 

incorporate the fundamental objectives of a University to the feasibility analysis of a 

WWTP. 

3 - Using Decision Analysis to Model the Triune Relationship of Supply, 

Demand, and Price 

Jeff Stonebraker, Assistant Professor, North Carolina State University, 

College of Management, Raleigh, United States of America, 

jeff_stonebraker@ncsu.edu 

Bayer was deciding whether to develop the third generation product used in the 

treatment of hemophilia A. Bayer also wanted to right size its production capacity to 

meet the demand while maximizing profitability. We use decision analysis to model 

Bayer’s supply-demand-price conundrum. 


423 

4 - Telling Value Stories with Value Diagrams 

Somik Raha, Student, Stanford University, 1329 Park Drive #12, 

Mountain View, CA, 94040, United States of America, 

somik.raha@gmail.com 

We will examine how decision diagrams currently tell value stories, and propose 

“Value Diagrams” to tell richer value stories that help clarify the value frame, create 

mutual understanding on value and inform our value thinking in decision diagrams. 

We will show through case studies how this can help us make decisions that are 

aligned with our values. 

5 - A Simple Interval-Valued Decision Tree 

Kash Barker, Lecturer, University of Oklahoma, 202 W. Boyd, 



Important to decision making is recognition of what today’s decisions have on 

future options. An oft-used tool to aid in this problem is the decision tree. To 

address situations when uncertainty arises in the metrics associated with different 

decision paths, a simple decision tree is developed for uncertain parameters where 

only bounds, not distributions, are known. Single- and multiobjective trees are 

discussed. 

■ WC41 


Metaheuristics I 


Chair: Heidi Taboada, Assistant Professor in Industrial, Manufacturing and 

Systems Engineering, Univeristy of Texas at El Paso, 500 W. University 

Ave., El Paso, TX, United States of America, hataboada@utep.edu 

1 - A New Evolutionary Algorithm Based on Adaptive Echolocation 

Heidi Taboada, Assistant Professor in Industrial, Manufacturing and 

Systems Engineering, Univeristy of Texas at El Paso, 500 W. 

University Ave., El Paso, TX, United States of America, 

hataboada@utep.edu, Karla Gutierrez 

A new evolutionary algorithm that is based on the principle of echolocation, also 

called biosonar is presented. This principle is active in numerous animals such as 

bats. These animals use it as radar in order to find food, obstacles or locate objects. 

The algorithm developed uses the radar method in order to explore the search space 

to obtain optimal solutions. The new method is tested on the well-known single 

objective redundancy allocation problem. 

2 - A Solution Method for the Constrained Level of Repair 

Analysis Problem 

Jose Espiritu, Assistant Professor in Industrial Engineering, 

Univeristy of Texas at El Paso, 500 West University Avenue, El Paso, 

TX, 79902, United States of America, jfespiritu@utep.edu, 

Carlos Ituarte-Villareal 

A Level of Repair Analysis model, determines the most cost-effective 

maintenance/replacement policy for each component within a system. In the 

present research we develop a heuristic approach to solve the Level of Repair 

Analysis considering budget constraints to indicate the optimal maintenance levels 

at which items will be removed, repaired and replaced to meet operational 

standards in a least optimal cost. 

3 - Tabu Search with Strategic Oscillation for a Maximum Dispersion 

Territory Design Problem 

Jabneel R. Maldonado-Flores, Universidad Autonoma de Nuevo 

Leon, CIDET-FIME, AP111-F, Cd. Universitaria, San Nicolàs de los 

Garza, NL, 66450, Mexico, jabneelmf@gmail.com, 

Roger Z. Rìos-Mercado, José Luis Gonzàlez Velarde 

We address a districting problem motivated by the application of the WEEE 

recycling directive in the European Union. In contrast to classical territory design, 

maximum territory dispersion is sought for avoiding the creation of monopolies 

forbidden by law in some countries. A tabu search with strategic oscillation is 

proposed and evaluated over a wide range of instances with very promising 

empirical results. 

4 - Bicriteria Optimization of Energy Efficient Placement and Routing in 

Heterogenous Sensor Networks 

Mustafa Baydogan, Research Assistant, Arizona State University, 

1802 E Randall Dr. Apt 3, Tempe, AZ, 85281, United States of 

America, mbaydoga@asu.edu, Nur Evin Ozdemirel 

We locate different type of sensors and route data generated to a base station under 

two conflicting objectives: minimization of network cost and maximization of 

network lifetime by satisfying connectivity and coverage requirements as well as 

sensor node and link capacity constraints. We propose formulations and use an 

exact solution approach to find Pareto solutions and develop a multiobjective GA to 

approximate the efficient frontier, as the exact solution requires long computation 

times.

WC42 

5 - An Evolutionary Approach Based on Viral Replication for Solving 

Combinatorial Optimization Problems 

Claudia Valles, MS student, The University of Texas at El Paso, 500 

W. University Av., El Paso, TX, United States of America, 

cevalles@miners.utep.edu, Heidi Taboada 

A new algorithm that mimics the performance of viruses is presented. The 

replication mechanism as well as the hosts’ infection processes are used to develop 

this new metaheuristic.The problem presented to show the performance of the 

proposed algorithm is the multiple objective redundancy allocation problem. The 

solution to this multiobjective problem is a set of Pareto-optimal solutions. 

■ WC42 


Optimization and Scheduling for Supercomputers 

Sponsor: Optimization/Computational Optimization and Software (Joint 

Cluster ICS) 


Chair: Xueping Li, Assistant Professor, University of Tennessee, 408 East 

Stadium Hall, Knoxville, TN, 37996, United States of America, 

Xueping.Li@utk.edu 

1 - Making Software Maintenance More Efficient on Kraken, the 1st 

Academic Petaflop Computer 

Mark Fahey, Scientific Computing Group Leader, National Institute 

for Computational Sciences, 1 Bethel Valley Road, P.O. Box 2008 MS 

6173, Oak Ridge, TN, 37831, United States of America, 

mfahey@utk.edu 

The National Institute for Computational Sciences is the newest NSF High 

Performance Computer center delivering 600 millions compute hours yearly to the 

TeraGrid and is the 3rd fastest machine on the Top500. I will describe Kraken’s 

capabilities and then talk about some infrastructure to (1) improve software 

installation and maintenance and (2) track library usage by all the users; both with 

the ultimate goal of improving the user experience while simultaneously make 

NICS more efficient. 

2 - Maintenance and Spare Part Inventory Control of Supercomputers 

Xiaoyan Zhu, Assistant Professor, University of Tennessee, 408 East 


xzhu5@utk.edu, Haitao Liao 

To ensure the normal operation of a supercomputer, maintenance must be 

performed effectively with the availability of spare parts to replace degraded or 

failed units. This work provides a preliminary model that simultaneously 

manipulates maintenance schedules and the number of spare parts to be held in 

inventory to minimize the overall operating cost (downtime cost, and procurement 

and holding costs for spare parts) while ensuring a high fill rate of spare parts. 

3 - User Strategies for Super Computer Scheduling 

Joe Wilck, The University of Tennessee - Knoxville, 411 East 

Stadium, Knoxville, TN, United States of America, jwilck@utk.edu, 

Jonathan Celso, Xueping Li, Mark Fahey 

We present strategies exhibited by users when using a super computer that has 

priority scheduling. The schedule is prioritized to run jobs that utilize the most 

nodes for the longest amount of time, and then backfills smaller jobs for the 

remaining capacity. Due to maintenance schedules, the users know that the queue 

must be cleared weekly. 

4 - Supply Chain Management Models and Heuristics for Data 

Cache Management 

Zhe Zhang, Research Staff Member, Oak Ridge National Laboratory, 

ORNL, 1 Bethel Valley Road, P.O. Box 2008 MS6008, Oak Ridge, TN, 

37831-6008, United States of America, zhezhang@ornl.gov, 

Xiaoyan Zhu, Rui Xu, Galen Shipman, Xiaosong Ma, Xueping Li 

For better application I/O performance, modern operating systems place certain data 

blocks in faster storage devices based on future access predictions. In Computer 

Science this method is named caching. While enhancing the system costeffectiveness, 

caching also creates challenges in data management, including data 

placement and transfer. In this work we exploit the similarities between the caching 

and supply chain management problems and apply SCM models and heuristics in 

data cache management. 


424 

■ WC44 


Cost Effectiveness Models in Health Care 



Chair: Greg Zaric, Associate Professor, Ivey Business School, 

1151 Richmond St., London, Canada, gzaric@ivey.uwo.ca 

Co-Chair: David Hutton, Stanford University, Palo Alto, CA, United States 

of America, billdave@stanford.edu 

1 - Cost Effectiveness of a Safe Consumption Site in Toronto, Canada 


94035, United States of America, evaenns@stanford.edu, Greg Zaric, 

Jennifer Jairam, Ahmed Bayoumi 

The establishment of a safe consumption site (SCS) in Toronto, Canada has been 

proposed to reduce the spread of disease through the sharing of drug use 

equipment. We developed a compartmental model of the spread of HIV and 

hepatitis C through populations of drug users and non-drug users matching those of 

the Toronto area. We account for geographically disparate drug user groups and 

mixing patterns between them, as well the influence of a centralized SCS on drug 

user mixing behavior. 

2 - Cost-Effectiveness of Stockpiling Masks and Respirators for the 

Next Influenza Pandemic 

David Hutton, Stanford University, Palo Alto, CA, 

United States of America, billdave@stanford.edu, Nayer Khazeni 

Surgical masks were used Mexico during the initial outbreak of the 2009 pandemic 

influenza (H1N1). Several countries are stockpiling masks to use in future 

pandemics. We use mathematical models of influenza disease spread to determine 

the minimal necessary effectiveness and compliance to make stockpiling of masks 

effective and cost-effective in an influenza pandemic. We compare these results with 

other mitigation strategies involving pharmaceuticals. 

3 - Cost-Effectiveness of a 21-gene Recurrence Score Assay in 

Patients with Early Stage Breast Cancer 

Malek Hannouf, University of Western Ontario, 

University of Western Ontario, London, ON, Canada, 

Malek.Bassam@schulich.uwo.ca, Bin Xie, Muriel Brackstone, 

Greg Zaric 

We developed a Markov model to evaluate the cost effectiveness of a 21-gene 

recurrence score assay versus current Canadian clinical guidelines in women with 

early stage breast cancer. The model was parameterized using 5 and 10 year follow 

up data from the Manitoba Cancer Registry and cost data from Manitoba Health. 

■ WC45 


Radiation Therapy Optimization 



Chair: Timothy Chan, Assistant Professor, University of Toronto, 5 King’s 

College Road, Toronto, ON, M5S 3G8, Canada, tcychan@mie.utoronto.ca 

1 - Quantifying the Trade-off Between Beam-on-time and Treatment 

Plan Quality 

Ehsan Salari, PhD Student, University of Florida, ISE Department, 

P.O. Box 116595, Gainesville, FL, 32611-659, United States of 

America, esalari@ufl.edu, Edwin Romeijn 

Beam-on-time is an important aspect of IMRT treatment efficiency, but its 

optimization is traditionally postponed until the leaf sequencing phase of treatment 

planning. However, there is a trade-off between treatment plan quality and beamon-time. 

The aim of this study is to incorporate the beam-on-time into a direct 

aperture optimization model. We formulate a bi-criteria optimization model and 

develop a solution method that efficiently obtains the entire set of Pareto-optimal 

treatment plans. 

2 - A Beam Angle Optimization Approach for Intensity Modulated 

Proton Therapy Treatment Planning 

Gino Lim, Associate Professor PhD, University of Houston, E211, 

Engineering Building 2, Houston, TX, 77204, United States of 

America, ginolim@uh.edu, Wenhua Cao 

We present an LP based local neighborhood search algorithm for solving the BAO 

problem in IMPT treatment planning. Three prostate cancer cases at M. D. Anderson 

were tested. Optimized angle sets demonstrated evident advantages comparing with 

the lateral opposed angles currently used at M. D. Anderson. Furthermore, we 

applied a worst case robust analysis that accounts for range uncertainties and setup 

errors on the optimized angles in order to validate the robustness of plan quality.

3 - An Inverse Optimization Approach to Determine Objective Function 

Weights in Radiation Therapy 

Taewoo Lee, PhD student, University of Toronto, 5 King’s College 

Road, Toronto, ON, M5S 3G8, Canada, taewoo.lee@utoronto.ca, 

Timothy Chan, Michael Sharpe, Tim Craig 

In a multi-objective optimization model for radiation therapy treatment planning, 

the determination of weights for different organ-specific objective functions is based 

on subjective beliefs and manual iterative loops. We present a linear inverse 

optimization model that objectively and efficiently determines the weights using 

historical treatment data from prostate cancer patients. 

4 - Dynamic Robust IMRT Optimization 

Timothy Chan, Assistant Professor, University of Toronto, 

5 King’s College Road, Toronto, ON, M5S 3G8, Canada, 

tcychan@mie.utoronto.ca, Velibor Misic 

Previous robust IMRT optimization studies used an uncertainty set to model data 

uncertainty, solved a single treatment planning problem and delivered the solution 

in all fractions. In this talk, we present a dynamic robust optimization methodology 

where prior data observations are used to update the uncertainty set and guide 

treatment re-optimization. We present results for a lung case demonstrating 

simultaneous improvement in tumor coverage and lung sparing over non-dynamic 

robust treatments. 

■ WC46 


Game-Theoretic Applications in Healthcare II 



Chair: Murat Kurt, PhD Student, University of Pittsburgh, Department of 

Industrial Engineering, 3700 O’Hara Street, 1048 Benedum Hall, 

Pittsburgh, PA, 15217, United States of America, muk7@pitt.edu 

Co-Chair: Reza Yaesoubi, Post-Doctoral Research Fellow, Harvard Medical 

School, 641 Huntingtone Ave., Boston, MA, 02115, United States of 

America, reza.yaesoubi@gmail.com 

1 - Influenza Vaccine Supply Chain Coordination with Certain Delivery 

and Asymmetric Information 

Javad Nasiry, HKUST, Clear Water Bay, Kowloon, Hong Kong - PRC, 

Javad.NASIRY@insead.edu, Stephen E. Chick, Sameer Hasija 

We develop a model to investigate the effects of information asymmetry between a 

manufacturer and a single buyer on flu vaccine supply chain coordination. We 

design an optimal menu of output-based screening contracts and develop a simple 

alternative menu that can achieve screening with no loss of efficiency under some 

conditions. The policy implications of the model are further discussed. 

2 - Imaging Room and Beyond: The Underlying Economics Behind 

Physicians’ Test-Ordering Behavior 

Tinglong Dai, PhD Student, Carnegie Mellon University, Tepper 

School of Business, Pittsburgh, PA, 15213, United States of America, 

dai@cmu.edu, Sridhar Tayur, Mustafa Akan 

Excessive diagnostic tests have long been viewed as one major aspect of health care 

inefficiency and are often attributed to the fee-for-service payment model. In this 

study we investigate the underlying operational and economic drives behind 

physicians’ test-prescribing behavior in the outpatient setting, motivated by a 

collaborative study with a major outpatient clinic. Our work also provides insights 

into the effects of the increasingly popular “comparison shopping” of health 

services. 

3 - Stochastic Dynamic Allocation of Deceased-donor Kidneys 

M Gisela Bardossy, University of Maryland, R.H.Smith School of 

Business, and Institute for Systems Research, College Park, MD, 

United States of America, bardossy@umd.edu, Inbal Yahav 

Deceased-donor kidneys are currently allocated to candidates through a priority 

point system that combines waiting times and human tissue matches. In this paper 

we evaluate a stochastic dynamic programming approach in search of rules of 

allocation that maximize a multicriteria objective to balance efficiency with equity. 

We test our approach on the current candidates’ kidney waiting list and find 

promising results. 


425 

■ WC47 


Diversification in Projects 



Chair: Karolina Glowacka, Assistant Professor, Stevens Institute of 

Technology, Howe School of Technology Management, Hoboken, NJ, 

07030, United States of America, kglowack@stevens.edu 

1 - Impacts of Change Orders on Project Performance 

Young Hoon Kwak, Associate Professor, The George Washington 

University, Department of Decision Sciences, 2201 G Street, NW, 

Suite 415, Washington, DC, 20052, United States of America, 

kwak@gwu.edu, Kunhee Choi, Jane Park 

Change orders are inevitable in infrastructure projects that often result in project 

disruptions, disputes, and delays. We attempt to quantify the impacts of change 

orders on project performance. Two different benchmarks, the original threshold 

versus the amended threshold, of project cost, schedule, and other variables are 

analyzed to measure the effect of change orders using 1372 infrastructure projects. 

We also apply project portfolio analysis to determine the key effects on change 

order. 

2 - Analytic Contingency Setting: An Algorithm for 

Estimating Contingencies 

Homayoun Khamooshi, Assistant Professor, The George Washington 

University, Decision Sciences,Funger Hall Room 408, 2115 G Street, 

NW, Washington, DC, 20052, United States of America, 

hkh@gwu.edu, Denis Cioffi 

To cover contingencies, usually a fixed percentage of a project budget is set aside. 

Our model uses the binomial probability distribution to estimate the potential 

number of risk realizations at a given confidence level, and ranked risk impacts are 

added over this number of risks, yielding the required contingency funds to provide 

coverage of the accepted risks. The budget found with this method compares 

extremely well with one ascertained from a numerical simulation of the risk 

occurrences. 

3 - Integrating Innovation Projects and Lean Six Sigma Projects in the 

Organizational Portfolio 

Frank T. Anbari, Clinical Professor, Drexel University, Goodwin 

College of Professional Studies, 3001 Market St., Suite 100, 

Philadelphia, PA, 19104, United States of America, 

anbari@drexel.edu 

There are two concurrent, major waves affecting management thinking currently: 

project management/ innovation and Lean Six Sigma/ process improvement. 

Management researchers in each of these two important areas are seldom aware of 

the progress achieved in the other area. We aim to provide a clear vision of the 

integration of the project management office (PMO)/ innovation and Lean Six 

Sigma operations to enhance the organization’s performance through continual 

improvement of its total system. 

4 - OptForceTM: A New Approach to Strategic Workforce Planning 

Jay April, OptTek Systems, Inc., 2241 17th Street, Boulder, CO, 

80302, april@opttek.com 

Companies annually invest billions of dollars in programs to pursue objectives such 

as improving workforce productivity, customer satisfaction, customer contracting 

requirements and legal settlements. OptForceTM provides significant benefits in 

these settings, by evaluating and selecting portfolios of proposed investments in 

programs to generate new decision alternatives that offer improved returns.

WC48 

■ WC48 


Tutorial: The Role and Challenges for Optimization in 

Competitive Electricity Markets 



Chair: Shmuel Oren, Professor, University of California-Berkeley, IEOR 

Department, 4135 Etcheverry Hall, Berkeley, CA, 94720, 

United States of America, oren@ieor.berkeley.edu 

1 - The Role and Challenges for Optimization in Competitive 

Electricity Markets 

Shmuel Oren, Professor, University of California-Berkeley, IEOR 

Department, 4135 Etcheverry Hall, Berkeley, CA, 94720, 

United States of America, oren@ieor.berkeley.edu 

This tutorial will describe key challenges in designing and operating competitive 

electricity markets. It will review the basic components of electricity markets and 

alternative structural approaches adopted in different systems. We will discuss the 

underlying optimization problems being solved in scheduling, operating and 

simulating electricity markets. New computational challenges and research activities 

in this area will also be reviewed. 

■ WC49 


Intellectual Property Issues in Technology 



Chair: Eric Walden, Wetherbe Professor of ISQS, Rawls College of 

Business, Box 42101, Lubbock, TX, 79409, United States of America, 

eric.walden@ttu.edu 

1 - On the Downloading vs. Uploading of Unauthorized Copies of 

Intellectual Property 

Jared Hansen, Assistant Professor, University of North Carolina at 

Charlotte, 9201 University City Boulevard, Charlotte, NC, 28223, 

United States of America, Jared.Hansen@uncc.edu, Eric Walden 

This research examines differences in ethical and legal perceptions regarding the 

downloading versus uploading of music files through unauthorized file sharing. Two 

explanations are tested (to explain willingness to participate): (1) restrictedness of 

use, (2) word of mouth (WOM). Survey results indicate different effects for 

uploading and downloading behavior. In short, consumers use different heuristics in 

deciding whether to participate in downloading versus uploading. 

2 - Digital Piracy: Trends in the Perception of Consequences 

Abbe Forman, Assistant Professor Teaching/Instruction, Temple 

University, 435 Alter Hall, 1805 N. Broad St, Philadelphia, PA, 

19122, United States of America, abbe.forman@temple.edu 

Digital piracy continues to be an alarming problem. Attempts to slow or stop the 

theft of intellectual property have been nearly futile. Many people do not believe 

that digital piracy is a crime or that they will “get caught”. This study represents an 

analysis of the literature regarding perceived consequences of digital piracy between 

2001 and 2009. Additionally, the results of an open ended questionnaire asking 

specifically about perceived consequences will be presented. 

3 - Designing Content Licensing Arrangements for Boundary 

Management in Hybrid Business Models 

Karl Lang, Professor of Information Systems, Baruch College, 

Zicklin School of Business, City University of New York, 

Karl.Lang@baruch.cuny.edu, Sirkka Jarvenpaa 

Using data from economic experiments and case studies from the music industry we 

find that boundary management, i.e. managing the organizational tension that the 

overlapping of commercial and sharing economies create, is a key success/failure 

factor for hybrid business models that combine elements of the private investment 

and collective action models. We argue that licensing design specifying the content 

access and reuse rights for consumers is a critical for effective boundary 

management. 

4 - Running in the Pack vs. Running Alone: When Does It Make Sense 

to Jointly Develop a Technology? 

Nitin Aggarwal, Assistant Professor, San Jose State University, 

1 Washington Square, San Jose, CA, 95192-0244, 

United States of America, aggarwal_n@cob.sjsu.edu 

We collect data from 436 experts involved in technology development to confirm 

that technology is made up of bundles of intellectual property, mostly owned by 

different entities. The data shows that there are significant transaction costs - search, 

coordination, and opportunism - involved in assembling the bundle. Initially, these 

costs are lower for hierarchies, but they gradually increase with complexity, nonsubstitutability, 

and IP distribution patterns, thereby making networks economical. 


426 

■ WC50 


Information Systems II 


Chair: Narges Kasiri, Assistant Professor, SUNY Oneonta, 2004 E 

Matthews, stillwater, OK, 74075, United States of America, 

kasiri@okstate.edu 

1 - Differential Privacy in the Context of Masking Numerical Data 

Krish Muralidhar, Professor, University of Kentucky, School of 

Management, Lexington, KY, 40506-0034, United States of America, 

krishm@uky.edu, Rathin Sarathy 

Recently, researchers at Microsoft have proposed a new privacy standard called 

Differential Privacy. In this study, we evaluate the efficacy of using differential 

privacy when numerical confidential data is masked. 

2 - Analysis of Spamming Behavior at Different Aggregation Levels and 

Implications for IT Security 

Serpil Sayin, Dr., Koç University, Rumeli Feneri Yolu, Sariyer, 

Istanbul, 34450, Turkey, ssayin@ku.edu.tr, John Quarterman, 

Manoj Parameswaran, Andrew Whinston 

Using e-mail spam data from different blocklist sources, we analyze the observed 

spamming behavior at the single IP and Autonomous System (AS) levels. Our single 

IP level study reveals the variability in the spamming tactics employed by different 

botnets. Our AS level analysis indicates that the distribution of spam contribution is 

highly skewed across organizations. We discuss the role of a reputation system in 

alternative incentive mechanisms to address IT security concerns. 

3 - When Should Software Firms Commercialize New Products via 

Freemium Business Models 

Marius Florin Niculescu, George Institute of Technology, College of 

Management, 800 West Peachtree Street NW, Atlanta, GA, 30308, 

United States of America, marius.niculescu@mgt.gatech.edu, 

D. J. Wu 

Freemium approach, whereby a firm gives away for free a certain level or type of 

consumption while making money on premium consumption, is spreading fast in 

the software industry. We advance a multiperiod framework with network 

externalities in order to describe several freemium business models. We solve firm’s 

optimal pricing strategy under each model, derive conditions when freemium 

models are superior to conventional for-fee and seeding models, and discuss 

managerial and policy implications. 

4 - Valuing RFID Investment in the Retail Sector: A Real Options Model 

Narges Kasiri, Assistant professor, SUNY Oneonta, 2004 E Matthews, 

stillwater, OK, 74075, United States of America, kasiri@okstate.edu, 

Ramesh Sharda 

The investment in Item-level RFID in the retail sector similar to any other IT 

investments is associated with high uncertainties in its potential benefits and costs. 

A real options approach is appropriate to analyze the cost and benefits of the 

investment problem. We apply a system dynamics simulation technique to generate 

the parameters of the real options model and discuss various options and scenarios 

retailers have in implementing item-level RFID in retail operations management. 

■ WC51 


Analysis at US Army Aviation and Missile 

Command (AMCOM) 



Chair: Wayne Bruno, Director, Command Analysis, U.S. Army Aviation & 

Missile Command, BLDG 5308, Restone Arsenal, Al, 35898, 

United States of America, wayne.bruno@us.army.mil 

1 - A Method for Determining the Mismatch Between Supply and 

Demand in the Army Supply Chain 

David Berkowitz, Professor of Marketing, University of Alabama - 

Huntsville, 301 Sparkman Drive, College of Business Administration, 

Huntsville, AL, 35899, United States of America, berkowd@uah.edu, 

Lucas Neidert, Al Wilhite, Fan Tseng 

Our paper develops a method for determining the mismatch between inventory 

control point supply and the field level demand for a Class IX item. We identify the 

characteristics of items that are likely to be in short supply. We determine the 

probability structure for supply. Then we run a regression based model to help 

predict demand. Finally we match the results to project a supply on hand at the end 

of a period.

2 - Demonstrating the Business Case for Condition-Based 

Maintenance (CBM) in Army Aviation 

Josh Kennedy, Chief, Ops Analysis Branch, Command Analysis 

Directorate, US Army Aviation & Missile Command, Bldg 5308, 

Redstone Arsenal, AL, 35898, United States of America, 

josh-kennedy@us.army.mil 

The Army’s Aviation & Missile Command is pursuing a large-scale CBM program for 

its helicopter fleet. DoD has directed CBM as a reliability and sustainability driver in 

life cycle systems management. The objectives for AMCOM’s CBM program are: 

decrease Soldier maintenance burden, increase availability, enhance safety, and 

reduce costs. However, the investment required for a CBM program is substantial. 

This presentation outlines how AMCOM made a clear business case for this 

investment. 

3 - Application of Cost-Benefit Analysis (CBA) to Respond to Army 

Aviation Maintenance Issues 

Tina Theiss, Operations Research Analyst, Command Analysis 

Directorate, US Army Aviation & Missile Command, Bldg 5308, 

Redstone Arsenal, AL, 35898, United States of America, 

tina.theiss@us.army.mil 

Currently, there is no cyclic Army Aviation sustainment program to mitigate risk 

associated with the long term effects of airframe aging and use. At the same time, 

the Army has initiated a thorough CBA process to accompany unfunded 

requirements, the net result of which should be a strong value proposition stating 

that the benefits of a new requirement more than justify the costs. This presentation 

discusses how we applied the CBA process to respond to shortfalls with helicopter 

maintenance. 

■ WC52 


Homeland Security 


Chair: Hiram Moya, PhD Candidate, Texas A&M University, 

2807 Henry Ct., College Station, TX, 77845, United States of America, 

Hiram@tamu.edu 

1 - Infrastructure Security via Game Theory 

Zhe Duan, Rutgers University, 900 Davidson Road, Apt. 94, 

Piscataway, NJ, 08854, United States of America, 

duanzhe@gmail.com, Melike Baykal-Gursoy 

Recent attacks in Mumbai, Russia, and the attempted attack in New York city have 

forced governments to devote significant time and resources to secure 

infrastructures. In this paper, we study how to secure infrastructures via game 

theory. The uncertainty of the timing and the location of an attack is alleviated by 

considering the occupancy level of each location. Zero-sum and non-zero sum 

games are solved and equilibrium policies are obtained. 

2 - Protecting Electric Power Systems From Terrorist Attacks 

Vanlapha Santithammarak, PhD Student, Texas Tech University, 

2619 19th St. Apt 10, Lubbock, 79410, United States of America, 

vanlapha.santithammarak@ttu.edu, Milton Smith 

Terrorist attacks designed to weaken the economy are increasing concern after 9/11. 

Electrical power systems should be regarded as likely targets of terrorists due to the 

most effects of economic activities. The objective of this study is to investigate the 

methods for protecting from physical attacks and reducing vulnerability in power 

system networks. 

3 - Safe Path Optimization with Local Protection Required 

Ruben Dario Yie Pinedo, University at Buffalo (SUNY), Department 

of Industrial & Systems Engg, 308 Bell Hall, Buffalo, NY, 14260, 

United States of America, rubenyie@buffalo.edu, Rajan Batta, 

Moises Sudit 

The edges in a network can be exposed to certain events that compromise the safety 

of the vehicle. To reduce the likelihood of an event protection could be used. This 

talk will discuss the problem of routing several vehicles from multiple origins to 

multiple destinations. The network contains zones in which local protection is 

required. Local escorts are not allowed to leave their jurisdiction zones. The main 

objective is to minimize the total treat level to which the vehicles are exposed. 

4 - A Risk Based Sensor Allocation Problem 

Amir Ghafoori, PhD Student, Rutgers University, Department of 

Industrial and Systems Eng, 96 Frelinghuysen Road, Piscataway, NJ, 

08854, United States of America, ghafoori@eden.rutgers.edu 

We develop a risk based model for underwater sensor allocation to detect divers. 

The model investigates various aspects of sensor placement problems for underwater 

application and attempts to put sensors in the set of candidate grid points. Each 

sensor type brings its specifications to the model and poses a level of complexity. 

Finally an optimal scheme of sensor positioning is proposed in order to minimize 

the average total risk in the field. 


427 

5 - Transient Analysis of the Border Crossing Process using Congestion 

Based Policies 

Hiram Moya, PhD Candidate, Texas A&M University, 

2807 Henry Ct., College Station, TX, 77845, 

United States of America, Hiram@tamu.edu 

Trade is the U.S. depends on an efficient flow of inspected containers in and out of 

the border ports of entry (POE), while focusing on security, and being cost effective. 

This research focuses on all commercial traffic at a southern border POE, where 

there is a non-steady state, terminating system. Using transient analysis, we present 

analytical and experimental results of congestion based policies with a fixed number 

of servers, by implementing a primary inspection station service switch. 

■ WC53 


Operations/Marketing Interface II 


Chair: Kunpeng Li, Sam Houston State University, Avenue I, Huntsville, 

TX, United States of America, kli@shsu.edu 

1 - The Effect of Traffic on Retail Store Conversion Rate and Sales 

Olga Perdikaki, Assistant Professor, Texas A&M, Mays Business 

School, College Station, 77843-4217, United States of America, 

operdikaki@tamu.edu, Jayashankar Swaminathan, 

Saravanan Kesavan 

In this paper, we use proprietary data pertaining to a retailer to study the 

relationship between traffic and store sales performance, measured as the number of 

transactions and sales volume. We find that store sales performance exhibits 

diminishing returns to scale with respect to store traffic and increases in traffic 

variabilities are associated with declines in store sales performance. Finally, we find 

that store labor moderates the impact of traffic on store sales performance. 

2 - Product Line Decision with Incomplete Information 

Michael Lim, Asst Professor, University of Illinois, 350 Wohlers Hall, 

1206 S. 6th Street, Champaign, 61820, United States of America, 

mlim@illinois.edu 

We study manufacturer’s product line length decision under information asymmetry 

where retailer has a private information regarding local consumers’ preference. We 

use mechanism design to construct a contract that elicits retailer’s information rent. 

Our analysis identifies the condition in which the manufacturer have more (or less) 

incentive to extend the product line compared to full information case. We will also 

discuss some insights and policy implications from the model. 

3 - Return Policies and Informational Tools in Experience 

Goods Markets 

Eylem Koca, University of Maryland, College Park, Smith School of 

Business, VMH 3330, College Park, MD, 20740, 

United States of America, ekoca@umd.edu, Gilvan Souza 

We investigate the role of return policies and informational tools provided by the 

seller in consumer purchasing behavior and on the overall market outcome. We 

build a model of consumer learning and analyze a seller’s decision process in a twoperiod 

setting. We attain significant analytical findings with no distributional 

assumptions, and fully study the joint optimization problem under uniform 

valuations. Finally, we study competition in a duopoly setting, and look at some 

extensions. 

4 - Shelf Space Competition Between Store and National Brands 

Shu-Jung Sunny Yang, Lecturer in Management, The University of 

Melbourne, Level 10, 198 Berkeley Street, Melbourne, 3010, 

Australia, sunnyy@unimelb.edu.au, Chia-wei Kuo, Pei-Ju Lu 

Shelf space allocation is one of the retailer’s most challenging operational decisions. 

We propose a game-theoretic model in which one retailer, acting as a leader by 

deciding the total shelf space available and selling both national and store brands, 

maximizes her category profit, and one national-brand manufacturer, acting as a 

follower, maximizes his own profit. Our analysis suggests that the allocation of the 

shelf space depends on two thresholds of total shelf space for both brands. 

5 - Optimal Price Reduction Strategy in Product Sales 

Kunpeng Li, Sam Houston State University, Avenue I, Huntsville, TX, 

United States of America, kli@shsu.edu, Chongqi Wu 

Consumers with high valuation are willing to pay a higher price than those with 

low valuation. Therefore, firms should make sure that consumers with high 

valuation buy first. Later on, firms reduce the product price and sell to consumers 

with low valuation. Under such a setting, our paper investigates many important 

questions in the decisions of product design and introduction.

WC54 

■ WC54 


Innovation Management for Sustainability 



Chair: Hsueh-Ming Wang, Associate Professor, University of Alaska 

Anchorage, ESM Department 3211 Providence Dr., Anchorage, AK, 

99508, United States of America, afhsw1@uaa.alaska.edu 

1 - Curriculum Development for the Innovation Management Program 

for Sustainability 

Seong Dae Kim, Assistant Professor, University of Alaska Anchorage, 

PM Department at University Center, Room 155, 3901 Old Seward 

Highway, Anchorage, AK, 99503, United States of America, 

afsdk1@uaa.alaska.edu, Hsueh-Ming Wang 

The curriculum is focusing on the sustainability through the innovation of quality of 

life. This program will help students for building industry careers as system 

designers, architects, project managers, developers, and entrepreneurs and providing 

students with broad understanding of green engineering, energy saving lighting, and 

innovation of quality of life as well as to stimulating the development and 

application of energy efficient lighting, green technologies with quality innovation. 

2 - The Critical Factors for Decision-making of the Technology 

Policies of the Industrialized Countries for Municipal Solid 

Waste Management 

Leslie Simmons, PhD Student, University of Alaska, 3211 Providence 

Dr, Anchorage, AK, 99508, United States of America, 

aflfs@uaa.alaska.edu, Hsueh-Ming Wang 

In the United States of America, landfill disposal is a major method to handle 

municipal solid waste. It may generate pollutants and carbon emissions in the 

environment. Some industrialized countries, such as Germany, Denmark, and 

Japan, are using technologies to reduce, reuse, recycle, or combust to limit amounts 

of land disposed municipal solid waste. Technology policies from governments 

generate different outcomes of solid waste management. Many factors may affect 

policy-making process. This research evaluates critical factors that impact technology 

policies by surveying and comparing results in industrialized countries. Critical 

impact factors for decision-making priorities differ between the USA and other 

industrialized countries. These different technology policies may affect 

environmental protection and carbon emissions in the future. 

3 - Cognitive Oriented Integration for Innovation Management 

Hsueh-Ming Wang, Associate Professor, University of Alaska 

Anchorage, ESM Department 3211 Providence Dr., Anchorage, AK, 

99508, United States of America, afhsw1@uaa.alaska.edu, 

Muchiu Chang 

The acceptance of a new product depends on the market needs of the innovation 

and service compliance through the product life. We develop an innovation 

management process by cognitive based consumer behaviors integration. It includes: 

new ideas to lead the customer needs, adopting new technology in design, patent 

mapping market anaylsis, and life cycle management. Modeling simulation is a 

robust solution to virtually validate and verify the life cycle behaviors of various 

design concepts. 

4 - Agility Management for the Project Sustainability 

Peter Lang, Graduate Student, University of Alaska Anchorage, ESM 

Department, 3211 Providence Dr., Anchorage, AK, 99508, United 

States of America, peter@alaskadatatech.com, Hsueh-Ming Wang 

The projectized organization surrounding culturalvalues addresses concerns of 

balancing versus optimizing various value propositioning factors. Information 

Technology for project portfolios selection processes may profit by looking across a 

variety of fields ranging from project management to the cognitive sciences to 

increase project agility. The solutions may obtain by meta-modeling balancing, 

aligning, and integrating a proposed adaptive risk management approach. 


428 

■ WC55 


Innovation and the Economy 



Chair: Erica Fuchs, PhD, Carnegie Mellon University, 5000 Forbes 

Avenue, Baker Hall 129, Pittsburgh, PA, 15217, United States of America, 

erhf@andrew.cmu.edu 

1 - Patents, Materials Transfer Agreements (MTAs), and the Flow of 

Scientific Knowledge 

David Mowery, Professor, University of California Berkeley, HAAS, 

mowery@haas.berkeley.edu, Neil Thompson, Arvids Ziedonis 

How does university involvement in academic patenting affect academic science? 

This paper extends the work of Murray and Stern (2007) to cover a broader sample 

of published scientific papers in the biomedical and other disciplines, and examines 

the effects of both patenting and material transfer agreements (MTAs) on citations 

to scientific papers. 

2 - Economic Downturns, Inventor Careers, and Technology 

Trajectories in the U.S 

Wunmi Akinsanmi, PhD Student, Carnegie Mellon University, 

5000 Forbes Avenue, Baker Hall 129, Pittsburgh, PA, 15217, 

United States of America, eyidearie@gmail.com, Erica Fuchs 

This research explores the relationship between the telecommunications bubble 

burst and the quantity, direction and locus of U.S. innovation. We focus on 

optoelectronics, a general purpose technology with applications in energy, 

biomedical, telecommunications, computing and military. Leveraging USPTO patents 

and inventor CVs, we analyze how inventors’ pre-bubble productivity, mobility, 

social capital and degree of specialization influence these same post-bubble and 

thereby the national trend. 

3 - Intellectual Property, Prior Knowledge & New Firm Survival 

Sonali Shah, Buerk Fellow and Assistant Professor, University of 

Washington, Box 353200, Seattle, WA, 98195, United States of 

America, skshah@u.washington.edu, Sheryl Winston Smith 

We examine the joint effects of founders’ prior knowledge and intellectual property 

protection on the survival of young firms. We examine three different sources of a 

founder’s prior knowledge: prior industry experience, prior entrepreneurial 

experience, and prior entrepreneurial experience in the same industry. Taken 

together, our findings show the importance of patents as strategic, as well as 

appropriability, tools and are in-line with evolutionary economic theory. 

■ WC56 


Simulation and Optimization I 


Chair: Honggang Wang, Stanford University, Department of Energy 

Resources Engi, Stanford, CA, 94305, United States of America, 

honggang@stanford.edu 

1 - Stochastic Dominance Based Ranking and Selection in Simulation 

Demet Batur, University of Nebraska- Lincoln, 175 Nebraska Hall, 

Lincoln, NE, United States of America, dbatur2@unl.edu, 

Fred Choobineh 

We present a ranking and selection procedure where simulated systems are 

compared based on the stochastic dominance relationship of a performance metric 

of interest. The system which stochastically dominates all other systems is deemed 

as the best system. If there is not a unique best system, but a number of systems 

with crossing distribution functions, then the proposed procedure selects a set of 

nondominated systems with a probability of correct selection guarantee. 

2 - Modeling the Relative Efficiency of Job Assignment in Differing 

Social Structures 

Paul Kerl, Georgia Institute of Technology, 765 Ferst Drive, NW, 

Atlanta, GA, 30332, United States of America, paul.kerl@gatech.edu, 

Joel Sokol 

In the recent and distant past, people have found their job in many different ways, 

from inherited positions to national testing and matching. We use stylized 

assignment models to examine several social structures and measure the relative 

benefit of structural characteristics. These characteristics include job inheritance, 

split social structures, heuristic selection versus optimal selection of jobs, job 

mobility, and the marginal value of information about person-job matching.

3 - Use of Retrospective Optimization for Placement of Oil Wells 

Under Uncertainty 

Honggang Wang, Stanford University, Department of Energy 

Resources Engi, Stanford, CA, 94305, United States of America, 

honggang@stanford.edu, David Echeverri Ciaurri, Louis Durlofsky 

Determining well locations in oil reservoirs under geological uncertainty remains a 

challenging problem in field development. We simulate (with a reservoir simulator) 

multiple model realizations to estimate the expected field performance for a certain 

well placement. The presented RO framework generates a sequence of sample-path 

problems with increasing sample sizes. The numerical results show that the RO 

algorithm finds a solution yielding a 70% increase in the NPV for the problem 

considered. 

4 - Water Quality Monitoring Network Design using Optimization 

via Simulation 

Chuljin Park, PhDStudent, Georgia Technology of Institute, 765 Ferst 

Dr NE, Main Building #124, Atlanta, GA, 30332, United States of 

America, cpark41@mail.gatech.edu, Seong-Hee Kim, Ilker Telci, 

Mustafa Aral 

The problem of designing a water quality monitoring network for river systems is to 

find the optimal location of a finite number of monitoring devices that minimizes 

the expected detection time of a contaminant spill event with good detection 

reliability. We set a stochastic constraint on detection reliability and solve the 

monitoring design problem using optimization via simulation. 

■ WC57 


Risk Analysis 


Chair: John Chachere, Senior Computer Scientist, Stinger Ghaffarian 

Technologies, 1060 Arbor Road, Menlo Park, CA, 94025, 

United States of America, john.m.chachere@nasa.gov 

1 - Why Do Groups Cooperate More Than Individuals to Reduce Risks? 

Min Gong, Columbia University, 1190 Amsterdam Ave, 

406 Schermerhorn Hall - MC 5501, New York, NY, 10027, 

United States of America, mingong@gmail.com, Jonathan Baron, 

Howard Kunreuther 

We find that groups are less cooperative than individuals in a prisoner’s dilemma, 

but are more cooperative than individuals in a stochastic version of the game.Three 

processes are tested: risk concern, cooperation expectation, and social pressure. Data 

shows that guilt aversion and blame avoidance cause groups to be more risk 

concerned than individuals, which drives groups to cooperate to reduce risks. 

Groups also have higher cooperation expectations for their counterpart than 

individuals have. 

2 - Decision Support for Inland Waterways Emergency Response 

Leily Farrokhvar, University of Arkansas, 4207 Bell Engineering, W 

Dickson Street, Fayetteville, AR, 72701, United States of America, 

lfarrokh@uark.edu, Heather Nachtmann, Ed Pohl 

Emergency planning involving transportation resources requires thorough 

contingency planning in case of route destruction due to natural or man-made 

events. Inland waterways can provide emergency transportation to a large area of 

the United States. We explore the potential for communities to benefit from inland 

waterway emergency response through the development of a decision support 

framework to support an inland waterway-based emergency response system. 

3 - Quantitative Method for Analyzing Engineering Defect 

Risks in Projects 


Technologies, 1060 Arbor Road, Menlo Park, CA, 94025, 

United States of America, john.m.chachere@nasa.gov 

Engineering defects often cause complex systems to fail. I provide a quantitative 

model of defects linking causes in an upstream development project to effects on 

downstream operations failure risks. The Model integrates: a simulation of the 

development project, a probabilistic analysis of operations failure risks, and a 

rational model of project decisions (e.g., organizational structure) and product 

decisions (e.g., subsystem redundancies). I evaluate an example project. 


429 

■ WC58 


Financial Engineering I 


Chair: Arun Chockalingam, Visiting Assistant Professor, Purdue 

University, 315 N. Grant Street, West Lafayette, IN, 47907, 

United States of America, achockal@purdue.edu 

1 - Margining Option Portfolios by Offsets with Two, Three and 

Four Legs 

Dmytro Matsypura, Dr, The University of Sydney, 

Merewether Building H04, Sydney, 2006, Australia, 

dmytro.matsypura@sydney.edu.au, Vadim Timkovsky 

As shown in [Rudd and Schroeder, 1982], the problem of margining option portfolios 

where option spreads with two legs are used for offsetting can be solved in 

polynomial time by network flow algorithms. In this paper we extend this result to 

the case where option spreads with three and four legs can also be used for 

offsetting and propose a general method of margining option portfolios. 

2 - Cross Validation and Various Techniques in Utilizing Financial Time 

Series Data for Model Validation 

Wing-Ho Choi, Gradient Analytics, 14614 North Kierland Boulevard, 

Scottsdale, AZ, 85254, United States of America, wc447@cornell.edu, 

Derek Koh 

When modeling financial data using insamples and outsamples, one has to account 

for the information overlap of return information which can create an optimistic 

bias on results. This is typically done by imposing a blackout period between in and 

out samples. The period is dependant on the return horizon where a long horizon 

results in a larger blackout period. This paper modifies the cross-validation 

technique and explores methods that optimizes the use of data available for 

financial modeling. 

3 - New Estimation Techniques for Fractional Brownian Motion with 

Applications to Finance 

Daniel Scansaroli, PhD Candidate, Lehigh University, 200 West 

Packer Ave., Industrial Engineering Department, Bethlehem, PA, 

18015, United States of America, djse@lehigh.edu, Robert H. Storer, 

Vladimir Dobric 

We explore financial applications of fractional Brownian motion and present two 

new methods of estimating the parameters of this process based on ergodic theory. 

We evince that in a fractional Brownian market, the typical assumption of 

independent increments results in an underestimation of market risk and a term 

structure of volatility. We conclude by applying our new technique to the equity 

market to show the behavior of the Hurst index that over the last 35 years. 

4 - Moving-boundary Approaches for American Security Valuation 

Arun Chockalingam, Visiting Assistant Professor, Purdue University, 

315 N. Grant Street, West Lafayette, IN, 47907, United States of 

America, achockal@purdue.edu, Kumar Muthuraman 

Pricing American options gives rise to a free-boundary problem in PDEs (PIDEs if 

the asset price process is discontinuous). We present a computational scheme that 

solves for the option price and optimal-exercise policy by converting the freeboundary 

problem into a sequence of fixed-boundary problems. This scheme is 

general enough to handle a variety of market models. We also discuss error bounds 

for options priced with sub-optimal exercise policies and the implications of these 

bounds. 

■ WC61 


Operations Management VII 


Chair: Jieling Han, PhD Student, Department of ISOM / University of 

Washington, Department of ISOM, Box 353200, University of 

Washington, Seattle, WA, 98115, United States of America, hanjl@uw.edu 

1 - Optimal Pallet Loading Problem with Complex Stacking Constraints 

Amit Garg, Senior Logistics Engineer, Penske Logistics, 

amit.garg@penske.com, Vishwa Ram, Prasad Natarajan, 

Kevin Troyer, Mitch Plesha 

Optimal Pallet Loading Problem is an important problem in the logistics industry. 

We investigate a particular problem for a food manufacturer where several 

categories of items need to be placed on the least number of pallets. Each item 

category primarily differs with other categories in the stacking constraints and 

product charactrestics. There are also specific business rules that dicatate how any 

two or more item categories can be placed on the same pallet. Current solution 

techniques and available commerical software solutions are unable to solve this 

problem because of the pecularities of the business rules and packing preferences in 

near real time for several thousand orders every day. Therefore, we implement a 

heuristc solution that takes various stacking rules and packing preferences and 

minimizes the number of pallets required for each order.

WC62 

2 - Profit- vs. Cost-orientation in the Newsvendor Problem: Insights 

From a Behavioral Study 

Sebastian Schiffels, Technische Universitaet Muenchen, 

TUM School of Management, Muenchen, 80333, Germany, 

sebastian.schiffels@wi.tum.de, Jens Brunner, Rainer Kolisch, 

Andreas Fuegener 

Our research investigates differences in the behavior of individuals in a profit- vs. a 

cost- orientated newsvendor problem. Our hypothesis is that individuals order more 

in the cost orientated than in the profit orientated newsvendor setting. Previous 

studies (e.g. Schweitzer and Cachon 2000) show that individuals deviate from the 

expected optimal decision in the newsvendor game. In fact, they tend to order less 

(more) than the expected profit maximizing order quantity if per unit profit margin 

is high (low). To test our hypothesis we set up a laboratory study which takes into 

account the profit or the cost perspective and three critical ratios. Our results 

confirm our hypothesis as well as the findings of previous studies. In each of the 

defined critical ratios the average order quantity in the cost orientated newsvendor 

game is significantly higher than in the profit orientated problem. The results imply 

that people underlie systematical different biases when facing profit and cost 

orientated situations. 

3 - Dynamic Scheduling Policy in a Make-to-stock System with Two 

Demand Classes of Different Variability 

Jieling Han, PhD Student, Department of ISOM / University of 

Washington, Department of ISOM, Box 353200, University of 

Washington, Seattle, WA, 98115, United States of America, 

hanjl@uw.edu, Apurva Jain 

We consider the dynamic scheduling policy in a make-to-stock queue with a single 

server, exponential processing times, standard holding and backorder cost rates, and 

two demand classes that differ in their variability. We partially characterize the 

optimal scheduling policy and propose heuristics. We then analyze the case where 

the centralized server has information about the arrival process of the more variable 

demand class and show the value of this information. 

■ WC62 


Transportation, Operations 


Chair: Rodrigo Britto, University of Maryland, 3909 Stoconga Drive, 

Betlsville MD 20705, United States of America, rbritto@rhsmith.umd.edu 

1 - Mapping Transportation Waste 

Bernardo Villarreal, Professor, Universidad de Monterrey, I. Morones 

Prieto 4500 Pte, San Pedro Garza Garcia, NL, 62638, Mexico, 

bvillarreal@udem.edu.mx 

Value stream mapping was developed originally to identify and eliminate waste in 

the manufacturing area. This is extended to design improvement programs oriented 

to eliminate waste for transport operations. The definition of several types of waste 

specific to transportation with the goal of improving efficiency as the relevant 

performance measure is suggested. Application to real examples is provided. 

2 - An Integrated Model for Resource Allocation and Scheduling in a 

Transshipment Container Terminal 

Nabil Nehme, PhD Candidate, American University of Beirut, P.O. 

Box 11-0236, Riad El Solh, Beirut 11, Beirut, Lebanon, 

nhn02@aub.edu.lb, Isam Kaysi, Farah Mneimneh, Bacel Maddah 

This paper considers the coordination between quay and yard sides in a 

transshipment process at a container terminal. A model is developed to minimize 

the number of cranes used and to determine the optimal schedule for unloading 

containers for a vessel. Several insights are drawn illustrating the importance of 

coordination. 

3 - A New Packing Heuristic Based Approach for the VRP with Threedimensional 

Loading Constraints 

Yi Tao, PhD Candidate, Sun Yat-sen Business School, Sun Yat-sen 

University, No.135 West Xinggang Road, Guangzhou, 510275, China, 

kenjimore@gmail.com, Fan Wang 

We consider the Three-Dimensional Loading Capacitated Vehicle Routing Problem 

which combines the routing of vehicles and the loading of three dimensional shaped 

goods into the vehicles while minimizing the total cost. We propose a least waste 

packing heuristic based approach for solving the loading subproblem, which is 

iteratively invoked by a simple tabu search algorithm for the routing. Numerical 

experiments on test instances have shown our method outperforms existing ones. 


430 

4 - Comparing Organizational Forms in the Trucking Industry 

Johan Lundin, PhD Student, Lund University, Box 118, Lund, 

22100, Sweden, johan.lundin@tlog.lth.se, Fredrik Eng Larsson 

Studies on organizational forms pertaining to efficiency show that pricing and 

deciding upon the right level of output can vary between forms (Porter and Scully, 

1987). Never before has the trucking been studied from this perspective, which is 

why we focus on exploring the truck investment decision based on economic 

efficiency. The methodology draw on industrial organization and game theory using 

a two-stage supply chain model under deterministic demand expressed as a Cournot 

competition model. 

5 - Assessment of the Impact of Undesirable Outputs on the 

Productivity of US Motor Carriers 

Rodrigo Britto, University of Maryland, 3909 Stoconga Drive, 

Betlsville, MD, 20705, United States of America, 

rbritto@rhsmith.umd.edu, Thomas Corsi 

This research evaluates the impact of undesirable outputs (i.e; crashes and fatalities) 

on the productivity of motor carriers during the years 1999-2003. The nonparametric 

direction output distance function and an additive DEA model are used 

to model both desirable and undesirable outputs. Tobit regression is applied in a 

second stage to analyze the drivers of efficiency. 

■ WC63 


Decision Analysis V 


Chair: Onur Bakir, Visiting Assistant Professor, Bilkent University, Bilkent 

Universitesi, Endüstri Mühendisligi Bˆlümü, Ankara, 06800, Turkey, 

nonur@bilkent.edu.tr 

1 - Deciding on the Decision Frame (The Most Important Decision of 

the Analysis) 

Roberto Ley-Borras, Director, Consultoria en Decisiones, Oriente 13 

A No. 1122, Orizaba, Ver., Mexico, rley@decidir.org 

Identifying the right decision frame is generally regarded as a critical part of a 

decision analysis, but there are few guidelines on how to achieve that. This talk 

presents a simple algorithm and some advice on generating decision frames and 

selecting the one that is best for the decision-maker’s current priorities and 

circumstances. Our proposed approach is considering the selection of the decision 

frame a decision situation by itself and using DA tools to gain clarity on the best 

frame. 

2 - Modified Repertory Grid Approach to Developing Criteria for MCDM 

Eric Johnson, Decision Strategies, Inc., 10260 Westheimer Road, 

Suite 250, Houston, TX, 77079, United States of America, 

ERJohnson@DecisionStrategies.com 

Multi-criterion decision making has been widely used for decisions where 

stakeholders do not agree on a single overriding objective. Assessment of criterion 

weights and scores is well understood. But all too often the criteria are too vague to 

be properly judged, or are understood differently by different judges, leading to 

results that aren’t compelling or don’t even make sense. This talk describes a novel 

protocol for eliciting objectives, and illustrates its benefit in a disguised case. 

3 - A Multiple Criteria Sorting Method Based on Support 

Vector Machines 

Esra Karasakal, Industrial Engineering Department, Middle East 

Technical University, Ankara, 06530, Turkey, esra@ie.metu.edu.tr, 

Asli Duman 

In this study we develop a method based on Support Vector Machines (SVM) for 

multiple criteria sorting problems. We modify SVM models to handle preference 

ordering of classes. We compare the proposed method with SVM models on several 

example problems. 

4 - Risk Aversion and Value of Information Under Various Approaches 

Onur Bakir, Visiting Assistant Professor, Bilkent University, Bilkent 

Universitesi, Endüstri Mühendisligi Bölümü, Ankara, 06800, Turkey, 

nonur@bilkent.edu.tr 

A previous study explored the relationship between the value of information and 

risk aversion using the buying price approach. In this presentation, we discuss 

whether similar conclusions hold for other approaches to evaluate information. The 

information acquisition problem is analyzed in a two-action lottery setting. We 

derive conditions under which there exists a monotonic relationship between the 

decision maker’s risk tolerance and the value of information.

■ WC64 


Health Care, Processes 


Chair: Imran Hasan, Purdue University, 315 N Grant St., West Lafayette, 

United States of America, ihasan@purdue.edu 

1 - Estimation of Service Value Function 

Geun Wan Park, PhD Candidate, Korea University Businsess School, 

Korea Universtiry, Anam-dong, Seongbuk-gu, Seoul, 136-701, 

Korea, Republic of, gw_park@hotmail.com, Kwang-Tae Park 

We consider a dental clinic service as the series of service stages. We want to 

estimate service value function of each service stage. The function is usually linear 

function according to preexisting literatures. However, we think the function can be 

different based on different service stage (for example, quadratic function or cubic 

function, etc.) These functions to be estimated will be useful to determine the 

service value of the dental clinic service appropriately. 

2 - A Survey of the Unintended Consequences of Implementing Health 

Information Technology 

Shinyi Wu, Assistant Professor, University of Southern California, 

3715 McClintock Ave., GER 240C, Los Angeles, CA, 90089, 

United States of America, shinyiwu@usc.edu, Caitlin Hawkins 

Health information technology (HIT) has demonstrated many benefits but 

unintended consequences (UCs) of its implementation cause barriers to realization. 

We analyzed 215 responses to an online survey to discover the types and causes of 

UCs experienced by various settings of healthcare organizations and how they coped 

with them. The results showed that UCs persist over years and the most frequent 

UCs are workflow problems caused by workarounds, software design, and lack of 

stakeholder engagement. 

3 - Utility of Patient Length of Stay Information 

Imran Hasan, Purdue University, 315 N Grant St., West Lafayette, 

United States of America, ihasan@purdue.edu, Yuehwern Yih 

Emergency Department Crowding has become a major problem in ED in the 

US.One of the reasons cited for this problem is the unavailability of beds in the 

intensive care unit.A large number of models for the prediction of length of stay 

have been developed, but there has been no research pertaining to the advantages 

of such predictive models.Our goal is to evaluate the impact of knowing the LOS in 

advance, and how it can help in the planning of admissions and capacity decisions 

in the ICU. 

■ WC65 


Inventory Management VI 


Chair: Ozden Cakici, PhD Student, University of Rochester, Carol G 

Simon Hall 4th Floor, Rochester, NY, 14627, United States of America, 

engin.cakici@simon.rochester.edu 

1 - Managerial Application of Reducing Lead Times on Safety Stock 

Zhi Tao, PhD Student, Kent State University, Dep. of M&IS, Kent, 

44242, United States of America, ztao@kent.edu 

In this paper, I extend Ever’s paper (1999) by applying it to normal distribution of 

lead time and demand and making interactive graphs to show the decision rule of 

reducing safety stock based on the coefficient of variation of demand. Implication of 

the findings on vendor managed inventory is further explored. 

2 - Comparison of Continuous and Periodic Review Inventory Policies 

with Continuous Time Costing 

Ozden Cakici, PhD Student, University of Rochester, Carol G Simon 

Hall 4th Floor, Rochester, NY, 14627, United States of America, 

engin.cakici@simon.rochester.edu, Harry Groenevelt, 

Abraham Seidmann 

We develop a unified analysis characterizing both periodic and continuous 

inventory review policies assuming a general convex inventory related cost function 

and stationary demand with independent increments. The use of continuous costing 

allows us to correctly compare the economic performance of a variety of periodic 

and continuous review policies, while also providing the ability to assess the impact 

of tactical measures like lead time reduction and review policy changes. 


431 

■ WC66 


Miscellaneous Applications of Operations Research 


Chair: Scott Parr, Researcher, Florida Atlantic University, 777 Glades rd., 

Boca Raton, FL, 33431, United States of America, sparr1@fau.edu 

1 - Sequential Testing of K-out-of-n Systems Under 

Precedence Constraints 

Tonguç Unlüyurt, Sabanci University, Tuzla, Istanbul, Turkey, 

tonguc@sabanciUniversityedu, Elif Ozdemir 

We consider the minimum expected cost sequential testing problem for k-out-of-n 

systems under precedence constraints. Mainly the problem is to find a feasible 

strategy that produces the minimum cost binary decision tree that evaluates the kout-of-n 

function at hand when it is costly to learn the values of individual 

variables. We summarize the results from the literature and report our initial results 

of the heuristic algorithms that we develop. 

2 - Transit Signal Priority for Emergency Evacuation 

Scott Parr, Researcher, Florida Atlantic University, 777 Glades rd., 

Boca Raton, FL, 33431, United States of America, sparr1@fau.edu, 

Evangelos Kaisar 

This research answers the question, during an urban evacuation should regional 

planners allow transit units signal priority when police assisted traffic controls are 

not an option. A case study of Washington D.C. shows allowing transit signal 

priority (TSP) during an urban evacuation has little to no effect on evacuation 

clearance time or evacuee travel time. Furthermore, four non-prioritized units are 

required to accomplish the task of three prioritized vehicles. 

3 - An Integration of Genetic Algorithm and GIS For 

Sensor Optimization 

Berna Dengiz, Professor, Baskent University, Baglica Kampusu 

Eskisehir Road 20.km, Etimesgut, Ankara, 06530, Turkey, 

bdengiz@baskent.edu.tr, Derya Oktay, Orhan Dengiz 

This study presents a geographical information system (GIS) based procedure for the 

optimization of sensor locations and sensor location parameters using genetic 

algorithms (GAs). The considered problem which is a special case of the well-known 

setcovering problem is NP-complete. A sensor siting optimization is an integrated 

system that consists of objects used in numerous subjects which can be called as 

sensor, like bare eye, camera, radar, radio, base station terminal and sensor coverage 

area can be defined as viewshed (the total visibility of visible points). The sensor 

operating parameters are as follows: location, height , range, azimuth, vertical 

viewing angles, tilt and pitch values. Each different configuration of these 

parameters results in different coverage area for a sensor. In this work, the 

optimization of sensor siting mainly focus on finding out sensor sites and sensor 

running parameters on terrain at the area of interest which provides maximum 

coverage. 

■ WC67 


Simulation III 


Chair: Javier Faulin Fajardo, Professor, Public University of Navarre, 

Campus Arrosadia, Pamplona, 31006, Spain, javier.faulin@unavarra.es 

1 - Simulation of Contribution Rates to Fund Potential U.S. State and 

Federal Parental Leave Policies 

Beth Neary, PhD Student, Indiana University, 1315 E. 10th St. - 

Room 341, Bloomington, IN, 47405, United States of America, 

bneary@indiana.edu 

The United States is the only developed nation that lacks a national paid parental 

leave program. Several U.S. states, including California, New Jersey, and soon 

Washington, have instituted their own benefit programs. This project seeks to 

inform state and federal policy development by simulating the costs of potential paid 

leave provisions. Specifically, Monte Carlo simulation estimates payroll tax 

contribution rates for varying benefit period lengths, subsidy levels, and weekly 

caps. 

2 - Simulating Team Selection Strategies for Youth Sports Programs 

Stephanie Marhefka, Undergraduate Researcher, University of 

Arkanas, Department of Industrial Engineering, Fayetteville, AR, 

72701, United States of America, smarhefk@uark.edu, 

Scott J. Mason, Ed Pohl 

In this presentation we discuss several strategies for forming youth sports teams. 

The goal is to find a strategy that yeilds a balnced set of teams and accounts for 

pairing of coaches with children as well as skill levels of children. An Excel-Based 

VBA model is used to compare three player allocation heuritic models. Once teams 

are selected a season is simulated and the records of the teams are analyzed for 

balance. One strategy is shown to be superior.

WC68 

3 - Calibration of Computer Models using Stochastic Approximation 

Szu Hui Ng, Department of Industrial and Systems Engineering 

National University of Singapore, ISE Department, NUS, Singapore, 

isensh@nus.edu.sg 

Computer models are widely used to simulate real processes. However, there always 

exist some parameters which are unobservable in the real process but need to be 

specified in the model. The procedure to adjust these unknown parameters to fit the 

model to observed data is known as calibration. Here, we propose an effective and 

efficient algorithm based on the stochastic approximation approach for calibration. 

We demonstrate its feasibility and apply it to a disease microsimulation model. 

4 - Simulation of Multimodal Transport of Goods Between the Regions 

Atlantic and Mediterranean in Spain 

Alejandro Garcia del Valle, Professor, University La Coruna, 

Mendizabal s/n, Ferrol, 15403, Spain, agvalle@udc.es, Diego Crespo 

Pereira, Rosa Rios Prado, David del Rio Vilas, Javier Faulin Fajardo 

The roads congestion and environmental impacts they generate are a significant 

problem in developed countries. Spain has road, rail and sea: it is therefore crucial 

to study the multimodal transport. The sea-road inter-modality presents an 

interesting case of freight transport simulation. We will study how the goods could 

be distributed among the existing infrastructure. The proposed simulation model 

will address the transportation costs and the mechanisms of choice for users. 

■ WC68 


Strategy/Strategic Planning I 


Chair: Dale Amburgey, Director of Enrollment Analysis, Saint Joseph’s 

University, 5600 City Avenue, Philadelphia, PA, 191931, United States of 

America, dale.amburgey@sju.edu 

1 - Temporal Fit, Misfit, and Performance: Testing Pace Entrainment in 

the Movie Production Industry 

Miles Zachary, Texas Tech University, RCOBA Box 42101, 

Lubbock, TX, 79409, United States of America, 

miles.zachary@ttu.edu, Jeremy Short, Tyge Payne 

Entrainment theory is concerned with the effects of pace and/or phase 

synchronizations of two or more activities within a system. This paper utilizes 

entrainment theory to develop hypotheses and longitudinally test a model of 

temporal fit, misfit, and performance in the movie industry in order to examine the 

understudied and elusive role of time and timing’s relationship to organizational 

performance. 

2 - Competitive Dynamics in Buyer-supplier Relationships 

Yan Emma Liu, The University of Melbourne, Level 10, 198 Berkeley 

Street, Parkville, Melbourne, Australia, yanliu@unimelb.edu.au, 

Shu-Jung Sunny Yang 

This paper studies the impact of co-optitive awareness on buyer-supplier 

relationships. We develop a behavioral game-theoretic model, based on prospect 

theory, to investigate the motivation of vertical (dis)integration. Our analysis shows 

that value creation is directly proportional to downstream organization redundancy 

and value appropriation is influenced by upstream organization redundancy. Our 

research highlights the risk of ignoring behavioral drivers in decision making. 

3 - The Stakeholders’ Involvement in the Strategic Planning Process 

Baris Carikci, TUBITAK, Kocaeli, Kocaeli, Turkey, 

bcarikci@yahoo.com 

The tendency of Turkish public organization in favor of strategic management has 

become a apparent fact in Turkey. The aim of the paper is to explain the tools to 

uncover the relation between a successful strategic management process and 

stakeholders’ involvement in the examples of distinguished Turkish public 

institutions. With these basic tools, public sector can improve the tools they use to 

initiate their stratagic management programs. 

4 - Using Business Intelligence in College Admissions: 

A Strategic Approach 

Dale Amburgey, Director of Enrollment Analysis, Saint Joseph’s 

University, 5600 City Avenue, Philadelphia, PA, 191931, 

United States of America, dale.amburgey@sju.edu, John Yi 

Data from first-year enrolling students were analyzed to develop predictive models. 

A decision tree analysis, a neural network analysis, and a multiple regression 

analysis were conducted to predict each student’s GPA at the end of the first year of 

academic study. Overall model performance was evaluated by using the average 

square error. Suggestions for future analysis include expansion of the study to 

include more student-centric variables and to evaluate GPA at other student levels. 


432 

■ WC69 


Reverse Logistics I 






1 - Coordination in Reverse Supply Chains: A System 

Dynamics Approach 

Fereshteh Mafakheri, PhD Candidate, HEC Montreal, Montreal, QC, 

Canada, fereshteh.mafakheri@hec.ca, Fuzhan Nasiri 

A reverse supply chain is dealing with recovering a maximum value from products 

at the end of their life cycle by promoting recycling, re-manufacturing, refurbishing, 

and reusing activities. The objective is to protect the environment while creating 

profit through material or energy recovery. In this paper, we explore a System 

Dynamics approach to investigate the benefits of coordination between various 

parties involved in a reverse supply chain for used printer cartridges. 

2 - Incentives and Reverse Logistics Channels with Remanufacturing 

Chester Xiang, Assistant Professor, Clarkson University, 

8 Clarkson Ave, Potsdam, NY, 13699, United States of America, 

cxiang@clarkson.edu, Dennis Yu 

We study a closed-loop supply chain with remanufacturing. The customer demand 

is divided into two segments, i.e., business and individual customers, which show 

different responses to incentives of returning used products. The reverse logistics 

channels exhibit economies of scale. A third party can be used as a used-product 

collecting agent. We investigate the manufacture’s incentive schemes, reverse 

logistics channel strategies, and resulting financial and social impacts. 

3 - Part Recovery Under Quality and Demand Uncertainty with 

Environmental Costs 

Gonca Yildirim, University of Florida, University of Florida, 

Gainesville, FL, 32611, United States of America, gonca@ufl.edu, 

Elif Akcali, Pelin Bayindir 

We consider acquisition and stocking decisions, motivated by the operations of a 

salvaging facility that collects a particular end-of-life product, performs a series of 

disassembly and recovery operations to reclaim a reusable part and sells the 

recovered part in the used parts market. We model uncertainties in demand and 

quality classification of the parts and include environmental fees that create a nonnegligible 

tradeoff against the operational costs in acquisition and stocking decisions. 

4 - The Impact of Legislation on Product Recovery: Reuse or Recycle? 

Ibrahim Karakayali, Post-doctoral Research Fellow, McGill 

University, Desautels Faculty of Management, 1001 Sherbrooke St. 

West, Montreal, QC, H3A1G5, Canada, 

ibrahim.karakayali@mcgill.ca, Luk Van Wassenhove, Tamer Boyaci, 

Vedat Verter 

In this study, we develop stylized models to assess the effects of material recovery 

targets (induced by legislation such as WEEE) on industry decisions pertaining to 

product reusability and recycling. We conduct a comparative analysis of centralized 

setting where there is cannibalization among the new and the remanufactured 

products. Our analytical framework also incorporates multiple stakeholders 

including the OEM, consumers, regulator, and environmentally conscious groups. 

■ WC70 


Fleet-Level Environmental Evaluation of New Aircraft 

and Technologies 



Chair: William Crossley, Professor, Purdue University, School of 

Aeronautics and Astronautics, 701 W. Stadium Ave, West Lafayette, IN, 

47907-2045, United States of America, crossley@purdue.edu 

1 - Evaluating Future Environmental Impact of US Aviation via System 

Dynamics and Resource Allocation 

William Crossley, Professor, Purdue University, School of Aeronautics 

and Astronautics, 701 W. Stadium Ave, West Lafayette, IN, 47907- 

2045, United States of America, crossley@purdue.edu 

This study integrates system dynamics (SD) and resource allocation (RA) to evaluate 

environmental impact of new aircraft and technologies. SD models observed 

aviation trends and behaviors like order-delivery, price-demand elasticity, and 

environmental policy dynamics. RA determines the fleet mix to maximize airline 

profits subject to constraints. Results illustrate the impact of demand, policies, and 

new aircraft technologies on emissions and noise over the 2005-2040 period.

2 - Revenue-based Allocation Model for Fleet-level Environmental 

Impacts of Airline Operations 

Muharrem Mane, Post-doctoral Researcher, Purdue University, 

School of Aeronautics and Astronautics, 701 W. Stadium Ave, 

West Lafayette, IN, 47907-2045, United States of America, 

mane@purdue.edu, Dan DeLaurentis, William Crossley 

This work uses a resource allocation model of airline operations to estimate the 

impact that new technology and new aircraft have on fleet-level CO2 and NOx 

emissions and airport noise. To approximate airline decision-making, we created a 

simplified revenue model based on flight frequency and trip length. Using this in the 

resource allocation objective function, we can approximate airline operations and 

study the impact of aircraft-specific improvements on fleet-level environmental 

metrics. 

3 - Decomposition Approach for Aircraft Allocation Under 

Environmental Considerations 

Isaac Tetzloff, Graduate Student, Purdue University, School of 

Aeronautics and Astronautics, 701 W. Stadium Ave, West Lafayette, 

IN, 47907-2045, United States of America, isaact@purdue.edu, 

William Crossley 

Previous work to allocate aircraft amongst 257 domestic and international airports 

used one large integer programming problem. Assumptions and abstractions 

simplified the problem to represent a majority of commercial airline operations with 

at least the arrival or destination airport in the US. Decomposing the problem into 

smaller allocation problems maintains runtimes similar to the previous model, and 

now includes multiple airlines, more aircraft models, and ‘one-way’ routes. 

4 - Examining the Potential Environmental Impact of Legacy and 

Budget Carrier Competitive Balance 

Datu Agusdinata, Post-doctoral Researcher, Purdue University, 

School of Aeronautics and Astronautics, 701 W. Stadium Ave, 

West Lafayette, IN, 47907-2045, United States of America, 

bagusdin@purdue.edu 

Legacy and budget carriers in the US operate and compete using different business 

(e.g. cost & fare structure) and operations (e.g. fleet composition, network structure, 

and level of service) models. Based on a logit model to ascertain the market share of 

each carrier, this study investigates how the competition may evolve over time and 

assesses the resulting emissions and noise impact under multiple scenarios of 

technology improvement rate, demand growth, and relative cost advantage. 

■ WC71 


Decision Support Models in DEA 



Chair: Vladimir Krivonozhko, Institute for Systems Analysis, 

KrivonozhkoVE@mail.ru 

1 - DEA Models for Decision Making Support in Negotiation Process 

Vladimir Krivonozhko, Institute for Systems Analysis, 

KrivonozhkoVE@mail.ru, Alexander Piskunov, Andrey Lychev, 

Maria Piskunova 

In this paper, an approach is proposed on evaluation of agreements on transnational 

projects during the negotiation process. One can show that the negotiation process 

can be represented as the behaviour of decision making units (countries) in the 

multidimensional space of economic indicators using DEA models. In this case, the 

goals that can be reached by units as a result of accomplishment of joint project can 

be determined as points in the multidimensional space. Optimal directions toward 

these goals and cones of possible directions can be found with the help of Analytic 

Hierarchy Process (AHP). Our approach is illustrated on the real-life data taken from 

open international sources. 

2 - The Winner is Kobe: Site Selection for the Next-generation 

Supercomputing Center in Japan 

Kaoru Tone, National Graduate Institute for Policy Studies, Japan, 

tone@grips.ac.jp 

The Next Generation Supercomputer R&D Project is an endeavor to create a 10 

Pflop/s system by 2012. It will be considered to be one of the “Key Technologies of 

National Importance of Japan.” Its goals are: (1). Development and installation of 

the most advanced high performance supercomputer system; (2). Development and 

wide use of application software to utilize the supercomputer to the maximum 

extent; (3) Provision of a flexible computing environment by sharing the next 

generation supercomputer through connection with other supercomputers ; (4) 

Establishment of an “Advanced Computational Science and Technology Center.” In 

this talk, I will report how the site selection was decided using AHP and DEA. 


433 

3 - A Cost-constrained Measure of Energy Efficiency 

Kankana Mukherjee, Babson College, kmukherjee@babson.edu, 

Subhash Ray, Lei Chen 

We provide a measure of energy-use efficiency for a firm with cost constraints. 

Using DEA, we investigate the energy efficiency of the US manufacturing sector 

across states. Our results indicate that in many states the actual energy use by the 

typical firm exceeds the optimal use without any increase in cost. We also find that 

an effective tax on energy use in manufacturing would reduce the aggregate energy 

usage while at the same time increasing the aggregate labor employment in this 

sector. 

4 - Eco-efficiency Within Selected US Industries using Data 

Envelopment Analysis 

Paul Rouse, University of Auckland, p.rouse@auckland.ac.nz 

We set out to provide further empirical evidence on the relationship between firm 

environmental performance and economic performance. In contrast to other studies 

which have used either partial productivity measures or purely accounting ratios, 

we use a frontier production model, data envelopment analysis (DEA), using 

multiple inputs and multiple outputs to estimate economic performance. Using data 

from four U.S. industries that are typically viewed as ‘highly environmentally 

sensitive’ for a three year (2006 to 2008) period, we find some evidence that higher 

levels of environmental performance are significantly associated with higher levels 

of economic efficiency. Our empirical evidence is therefore broadly consistent with 

the ‘Porter’ hypothesis; that is greater firm environmental performance is associated 

with higher levels of economic performance. Results, however, are mixed with 

positive and statistically significant coefficients on both environmental strengths and 

weaknesses. We then use the Fare and Grosskopf ‘weak disposability’ framework to 

explain the conflicting results in the regression analyses using the KLD strengths 

and weaknesses. 

■ WC72 


Facility Logistics Interactive Session: Distribution 

Center Operations 



Chair: Russell Meller, Hefley Professor of Logistics and Entrepreneurship, 

University of Arkansas, 4207 Bell Engineering, Fayetteville, AR, 72701, 

United States of America, rmeller@uark.edu 

1 - Decentralized Control of High Density Storage Systems 

Kevin Gue, Associate Professor, Auburn University, Department of 

Industrial & Systems Engin, Auburn, AL, United States of America, 

kevin.gue@auburn.edu, Kai Furmans 

We describe a decentralized control algorithm for a grid-based, high density storage 

system based on the “slide puzzle architecture.” When executed with constant 

work-in-process, the system provides very high throughput, and yet it consumes 

less space than a typical automated storage and retrieval system. 

2 - The Fishbone Triangle Design for Dual-Command Cycles in a 

Unit-Load Warehouse 

Russell Meller, Hefley Professor of Logistics and Entrepreneurship, 

University of Arkansas, 4207 Bell Engineering, Fayetteville, AR, 

72701, United States of America, rmeller@uark.edu, Letitia Pohl, 

Kevin Gue 

The fishbone warehouse design improves upon traditional warehouse designs for 

single-command cycles, but is not as efficient for the travel-between leg of a dualcommand 

cycle. The fishbone triangle design was developed to address this issue. 

We have developed a set of analytical expressions for travel-between in the fishbone 

triangle warehouse, which allows us to optimize this non-traditional warehouse 

design. We will present results on its performance. 

3 - Determining the Optimal Aisle-width for a Semi-automated Picking 

System in a Distribution Center 

Pratik Parikh, Assistant Professor, Wright State University, 

3640 Col Glenn Hwy, 207 Russ Eng Center, Dayton, OH, 45435, 

United States of America, pratik.parikh@wright.edu, Sheena Finney 

We consider the aisle-design problem for a semi-automated picking system that uses 

a person-on-board material handling equipment, such as an order picker truck. 

Using a previously developed model to determine the optimal storage level for such 

a system, we conduct a simulation study to determine the aisle-width (narrow or 

wide) that minimizes total system cost. We also compare our results with those 

obtained for a manual system to generate managerial insights when designing such 

systems.

WC73 

4 - A Network Model for Warehouses with Non-Traditional Aisles 

Omer Ozturkoglu, Auburn University, Department of 

Industrial&Systems Engineering, Auburn, AL, 36849, United States 

of America, ozturom@auburn.edu, Kevin Gue, Russell Meller 

We show how to construct non-traditional aisle designs for unit-load warehouses 

with multiple pickup and deposit (P&D) points, using a network model to represent 

storage locations and P&D points. The model uses a meta-heuristic to search for 

optimal angles of cross aisles and picking aisles. 

■ WC73 


Joint Session TSL/ SPPSN: Methods to Support 

Regional Evacuation 




Chair: Irina Dolinskaya, Assistant Professor, Northwestern University, 

2145 Sheridan Road, M235, Evanston, IL, 60208, United States of 

America, dolira@northwestern.edu 

1 - Design of an Evacuation Network for Hurricane Evacuation 

Xinghua Wang, Texas A&M University, 303K Zachry, TAMU, 

College Station, United States of America, wxh55@tamu.edu, 

Justin Yates, Halit Uster 

We consider large scale evacuations in expected extreme event situations such as 

hurricanes. We present a mixed integer model to minimize total costs under time 

constraints while determining associated facility locations and traffic flows to 

construct evacuation routes. We develop a Benders decomposition based solution 

approach and report numerical results using Texas-based real data facilitated by GIS. 

2 - An Integrated Demand-Supply Framework for 

Evacuation Operations 

Yu Ting Hsu, Purdue University, West Lafayette, IN, 

United States of America, yhsu@purdue.edu, Srinivas Peeta 

Evacuation operations involve perspectives from both the demand and supply sides. 

This study proposes an integrated operational framework to address the interactions 

between evacuation flow management and individual behavior, by using robust 

evacuation behavior models that consider the various factors that influence 

evacuation-related decision-making. 

3 - Large-Scale Evacuations with Arc Routing 

Mark Hickman, Associate Professor, University of Arizona, Civil 

Engineering, 1209 E. Second St., Bldg 72, Tucson, AZ, 85721-0072, 

United States of America, mhickman@email.arizona.edu, 

Moshe Dror 

We consider a routing of public vehicles to evacuate persons during a large-scale 

emergency. To facilitate the evacuation, we consider a routing of vehicles along 

roadways to pick up people at their residences or other locations. We present an arc 

routing model for this case and illustrate the model on a large-scale case study. 

■ WC74 


Hazardous Material Transportation 



Chair: Changhyun Kwon, Assistant Professor, University at Buffalo, 

400 Bell Hall, Buffalo, NY, 14260, United States of America, 

chkwon@buffalo.edu 

1 - Value-at-Risk Model for Hazmat Transport 

Yingying Kang, PhD Candidate, University at Buffalo - SUNY, 

Department of Industrial Engineering, 438 Bell Hall, Buffalo, NY, 

14260, United States of America, ykang4@buffalo.edu, Rajan Batta, 

Changhyun Kwon 

We propose a new measurement of risk in hazmat transportation, namely Value-at- 

Risk (VaR). The VaR measurement aims to gauge the cutoff risks within a certain 

confidence level, instead of minimizing the expected risk or maximal risk of a 

hazmat path directly. This allows the choice of a hazmat route according to decision 

makers’ risk preferences. Furthermore, we display that the route choice with the 

VaR model depends on the level of risk tolerance. 


434 

2 - Environmental Risk Analysis of Railroad Transportation of 

Hazardous Materials 

Rapik Saat, Postdoctoral Research Associate, University of Illinois at 

Urbana-Champaign, 205 N Mathews RM3214, Urbana, IL, 61801, 

United States of America, mohdsaat@illinois.edu, Chris Barkan 

This presentation will describe a quantitative risk analysis of rail transportation of 

hazardous materials using an environmental consequence model in conjunction 

with a geographic information system (GIS) exposure analysis along the U.S. rail 

network. The annual risk estimate incorporates the estimated remediation cost, 

route-specific probability distributions of soil type and depth to groundwater, annual 

traffic volume, railcar accident rate, and tank car safety features. 

3 - Dual Toll Pricing for Hazardous Material Transport with Linear Delay 

Jiashan Wang, University at Buffalo- SUNY, Department of Industrial 

Engineering, Buffalo, NY, United States of America, 

jw282@buffalo.edu, Yingying Kang, Rajan Batta, Changhyun Kwon 

We propose a dual toll pricing method to mitigate risk of hazmat transportation. We 

aim to control both regular and hazmat vehicles to reduce the risk. We incorporate 

a new risk measure to consider duration-population-frequency of hazmat exposure. 

We formulate the model as an MPEC problem and then decompose the formulation 

into first-stage and second-stage problems. For each stage problem, we present 

methods to solve them separately. A numerical example is provided. 

Wednesday, 3:30pm - 5:00pm 

■ WD01 


Joint Session ENRE/ QSR: Renewable Energy 

Integration into Power Systems for Smart Operations 

Sponsor: Energy, Natural Resources and the Environment/ Quality, 

Statistics and Reliability 


Chair: Eunshin Byon, PhD, Postdoctoral Research Associate, Texas A&M 



1 - Optimizing the Acquisition and Operation of On-site 

Electricity Generation 

Kris Pruitt, kpruitt@mymail.mines.edu, Rob Braun, 

Alexandra Newman 

We present a mixed-integer, nonlinear program for designing and operating an onsite 

generation system to supply electricity to a large, commercial building. The 

system design includes renewable and non-renewable generation, net-metering to 

the grid, and on-site storage. The model determines the optimal mix of generators 

and storage units to acquire, along with their operating levels over time, to 

minimize total cost subject to system performance characteristics and the building’s 

demand. 

2 - Simulation and Optimization of Wind Farm Operations Under 

Stochastic Conditions 

Eunshin Byon, PhD, Postdoctoral Research Associate, Texas A&M 



This research aims at developing models and associated solution tools to devise 

optimal maintenance strategies for wind turbines, helping reduce the operation 

costs and enhancing the marketability of wind power. We provide an integrated 

framework including optimization models, and a discrete event-based simulation 

model characterizing the dynamic operations of wind power systems. We highlight 

the benefits of the resulting strategies through a case study. 

3 - Reliability Evaluation of Wind Turbine via Computer Simulation and 

Laboratory Experiments 

Haitao Liao, Assistant Professor, The University of Tennessee, 211 

Pasqua Building, Knoxville, TN, 37996, United States of America, 

hliao4@utk.edu, Seyed Ahmad Niknam, Janet Twomey 

It is important to ensure the long-term reliability of wind generators for secured, 

uninterrupted power extraction from the wind. This research aims at evaluating the 

reliability of wind generators in time-varying environments via computer simulation 

and laboratory experiments.

■ WD02 


Energy II 


Chair: Nicolas Lopez, Research Assistant, The University of Texas at 

El Paso, 500 West University Avenue, El Paso, TX, 79902, 

United States of America, nlopez3@miners.utep.edu 

1 - Incorporating Gas Pricing Into Oil & Gas Asset Development 

Planning - An Example 

Chaiyaporn Wiboonkij-Arphakul, Decision Analyst, Chevron, Tower 

III, SCB Park Plaza, 19 Ratchadapisek Rd, Chatuchak, Bangkok, 

10900, Thailand, cwiboonkij@gmail.com, Ing Jye Tsai 

As gas is sold in the $ per BTU, there is a clear incentive for one to maximize profits 

by accelerating the development of high-BTU wells. However, doing so in some 

environment (e.g., under Gas Sales Agreement) may lead to the loss of low-BTU gas 

reserves and the delivery shortfall penalty. For Chevron Thailand, profit is 

maximized through the ranking of drilling projects’ profitability index while BTU 

forecast is only used to ensure that future production meets contractual 

specifications. 

2 - A Framework to Integrate Renewable Energy to Improve the Energy 

Efficiency: An OR Application 

Haifeng Wang, IBM Research - China, Building19, Zhongguancun 

Software Park, 8 Dongbeiwang West Rd, Haidian District, Beijing, 

China, whf@cn.ibm.com, Wenjun Yin, Jin Dong, Feng Gao 

This paper focuses on developing a framework for smart grid to utilize weather and 

renewable energy production forecasting to help integrate distributed renewable 

energy and energy storage into the electric distribution and transmission system. A 

demand-supply matching model is developed, in which power demand is modeled a 

given stochastic process, and the optimal unit planning policy is researched. 

Moreover, the renewable energy efficiency is evaluated for different feasible policies. 

3 - A Greedy Approach to Scheduling Outage Tasks for Distribution 

Power Network 

Ming Zhao, IBM Research - China, Diamond building, ZGC Software 

Park, Haidian District, Beijing, 100193, China, 

papayazm@gmail.com, Feng Jin, Hairong Lv, Qiming Tian, Jin Dong, 

Wenjun Yin 

Because of the complexity of distribution power network, outage task scheduling 

problem is usually large-scale, which results in unpractical long running time for 

many traditional methods. In this paper, combination weight between different 

tasks was defined. Then, a weight-based greedy algorithm was proposed to facilitate 

the procedure. In this way, computational complexity was reduced from O(t^n) to 

O(tn). Test results on a real distribution system show the efficiency of the approach. 

4 - Hybrid Power Systems Optimization Considering Different 

Renewable Energy Technologies 

Nicolas Lopez, Research Assistant, The University of Texas at El Paso, 

500 West University Avenue, El Paso, TX, 79902, United States of 

America, nlopez3@miners.utep.edu, Jose Espiritu 

Hybrid power systems need to be evaluated based on the requirements of high 

penetration renewable energy technology applications. Additionally, modeling and 

analysis storage systems integration are also necessary to increase the effectiveness 

of hybrid power configurations. In the present talk, a software based simulation to 

understand the Hybrid power systems response considering various renewable 

energy technologies and energy storage options is presented. 

■ WD03 


Interactive Multiple Criteria Decision Making 

Sponsor: INFORMS Section on Multiple Criteria Decision Making 


Chair: Murat Köksalan, Middle East Technical University, ODTU 06531, 

Ankara, Turkey, koksalan@ie.metu.edu.tr 

1 - An Interactive Approach for MCDM using a Hybrid 

Tchebycheff/Linear Utility Function 

Ozgen Ozbey, SUNY Buffalo, 308A Bell Hall, Department of 

Industrial and Systems Eng, Buffalo, 14260-1900, United States of 

America, oozbey2@buffalo.edu, Mark H. Karwan 

We improve our previously developed MILP formulation to estimate a Decision 

Maker’s (DM) utility function used during an interactive method employing 

pairwise comparisons. The utility function is approximated by a Tchebycheff or 

hybrid function with Tchebycheff and linear components. We consider a DM’s 

precision and “strength of preferences” to obtain a most preferred solution among 

implicit alternatives in an effective manner. We present computational results and 

comparisons with other methods. 

INFORMS Austin – 2010 WD04 

435 

2 - Interactive Evolutionary Multi-Criteria Scheduling 

Jon Marquis, Senior Systems Engineer, Raytheon, 1151 East 

Hermans Road, Tucson, AZ, 85756, United States of America, 

jonemailbox1@gmail.com, John Fowler, Esma Gel, Pekka Korhonen, 

Jyrki Wallenius, Murat Köksalan 

We present a new approach to multi-criteria scheduling using an interactive 

evolutionary algorithm with a scheduling heuristic to develop solutions. The 

algorithm sends a subset of the solutions to the decision maker (DM) for evaluation 

and uses the resulting preference information to evaluate new solutions and guide 

the algorithm. We apply the algorithm to the weighted completion time, total 

tardiness, and maximum lateness criteria. 

3 - Solving Multiobjective Mixed Integer Programs using Convex Cones 

Murat Köksalan, Middle East Technical University, ODTU 06531, 

Ankara, Turkey, koksalan@ie.metu.edu.tr, Jyrki Wallenius, 

Banu Lokman, Pekka Korhonen 

We assume that the decision maker’s preferences are consistent with a quasiconcave 

utility function. Based on the convex cones derived from past preferences, we create 

constraints to prevent solutions in the implied inferior regions. We guarantee 

finding the most preferred solution and our computational results show that a 

reasonable number of pairwise comparisons are required. 

■ WD04 


Sustainability II 


Chair: Amrou Awaysheh, Assistant Professor of Operations Management, 

IE Business School, Maria de Molina, 12 - 5 planta, Madrid, 28006, 

Spain, amrou.awaysheh@ie.edu 

1 - The Impact of Green Vehicles on the Market Value of Automakers 

Qindong Liu, University of Connecticut School of Business, 

2100 Hillside Road, Storrs, CT 06269, United States of America, 

qindong.liu@business.uconn.edu, Jan Stallaert 

We conduct an event study to explore how stock markets react to automakers’ fuelefficient 

product strategies. We find automakers’ green technology and product 

innovations have a positive impact on their market valuation, although the 

technology and market segment choices have different implications. The abnormal 

returns could be explained by both internal and external factors. More interesting, 

our analysis indicates that product design variables moderate those relationships. 

2 - A LCA Method Considering Multi-Scenario and System Uncertianty 

Jianzhi Li, Assistant Professor, The University of Texas - Pan 

American, 1201 W University Dr., Edinburg, TX, 78539, 

United States of America, jianzhi.li06@gmail.com, Xiaowei Wang 

Current LCA approaches failed to consider spatial and temporal system uncertainty 

in its calculation. A new LCA method is proposed which considers uncertain future 

multi-scenario. The LCI data are collected as per scenario including pollution 

emissions and character information that are used to generate personal, spatial and 

temporal factors, which reflect the diversity of environment. The impacts is obtained 

by summing up the results of each scenario wighted by the scenario probabilities. 

3 - The Impact of Social Issues Management on Firm Performance - 

An Event Study 

Amrou Awaysheh, Assistant Professor of Operations Management, IE 

Business School, Maria de Molina, 12 - 5 planta, Madrid, 28006, 

Spain, amrou.awaysheh@ie.edu, Robert D. Klassen 

The management of social issues can have a substantial impact on a firm’s financial 

performance. Social issues are complex, and managing them requires a range of 

practices. Examples include establishing workforce policies for safe work practices or 

diversity; however, a firm may also be involved in negative practices, such as the 

use of illegal labor practices. This paper will present the results from an ongoing 

event study that examines the impact of social issues management on a firm’s value.

WD05 

■ WD05 


Dynamic Programming/Control II 


Chair: Suleyman Demirel, PhD Candidate, Ross School of Business, 

University of Michigan, 701 Tappan Ave, Ann Arbor, MI, 48104, 

United States of America, sdemirel@umich.edu 

1 - Ad Valorem Tax and the Cumulated Output of 

Exhaustible Resources 

Runfang Xu, PhD Candidate, Xi’an Jiaotong University, No.28, 

Xianning West Road, Xi’an, 710049, China, runfangxu@yahoo.com, 

Xinmei Liu 

This paper, after modifying the Hotelling model and using the dynamic control 

method, researches the relation between ad valorem tax and cumulated output of 

exhaustible resources under monopoly. It is shown that the cumulated output is 

lower than perfect competition when the ad valorem tax rate equals to zero. And 

the cumulated output is decreased more when the ad valorem tax rate more than 

zero. However, the cumulated output is increased when the ad valorem tax rate less 

than zero. 

2 - Sensitivity-Based Nested Partitions for Solving Markov 

Decision Processes 

Weiwei Chen, University of WIsconsin-Madison, 1513 University 

Avenue, Madison, WI, 53706, United States of America, 

wchen26@wisc.edu, Leyuan Shi, Yanjia Zhao, Xiaohong Guan 

This paper introduces an algorithm to solve finite-horizon total-cost Markov 

decision processes with non-stationary policies. It is based on Nested Partitions (NP) 

global optimization framework, and combines the search power of a local optimizer 

using sensitivity-based analysis. An intelligent partitioning approach is developed to 

determine the new partitions adaptively based on the information from previous 

iterations. The numerical example shows the effectiveness of the proposed 

algorithm. 

3 - Dynamic Airline Asset Allocation Incorporating Econometric 

Demand Models 

Navindran Davendralingam, Purdue University, School of 

Aeronautics and Astronautics, 701 West Stadium Avenue, West 

Lafayette, 47906, United States of America, davendra@purdue.edu, 

William Crossley 

Operational adjustments by airlines translate to latent passenger observations who 

dictate future demand trends based on cost feasibility and ancillary benefits offered 

by various ticket itineraries being published. The current proposed research develops 

a conceptual dynamic framework that builds upon previous paradigms in dynamic 

programming, econometrics and system engineering to maximize revenue for a 

given airline using tactical asset allocation subject to reflexive demand feedback. 

4 - Resource Taxation and the Cumulated Output of Exhaustible 

Resource Under the Externality 

Runfang Xu, PhD candidate, Xi’an Jiaotong University, No.28, 

Xianning West Road, Xi’an, 710049, China, runfangxu@yahoo.com, 

Xinmei Liu 

How to enhance cumulated output and how to lengthen the total lifetime of 

exhaustible resources is a very popular topic. Firstly, this paper constructs a dynamic 

optimal control model under the condition of externality. Secondly, this paper, 

applying simulation, analyses the relationship between resource taxation and 

cumulated output and total lifetime. Finally, the results are confirmed by data from 

China. 

5 - Production and Inventory Control for a Two-Stage Assemble-to- 

Order System with Uncertain Capacities 

Suleyman Demirel, PhD Candidate, Ross School of Business, 

University of Michigan, 701 Tappan Ave, Ann Arbor, MI, 48104, 

United States of America, sdemirel@umich.edu, Roman Kapuscinski, 

Izak Duenyas 

Consider a firm assembling product from subcomponents, where both assembly 

capacity and subassembly capacities are uncertain. The assembly capacity is 

common, while each component has dedicated subassembly resources. The firm 

may hold inventory of subassemblies. We analyze a multi-period inventory model 

for two products, and derive the optimal replenishment policy of the subassemblies 

as well as the optimal priority scheduling of the assembly resource. 


436 

■ WD07 


Supply Chain, Closed-loop II 


Chair: Jo Min, Iowa State University, IMSE Department, 3004 Black, 

Ames, IA, 50011, United States of America, jomin@iastate.edu 

1 - Hybrid Manufacturing/Remanufacturing System in Cascade Reuse 

Yasutaka Kainuma, Tokyo Metropolitan University, 6-6, Asahigaoka, 

Hino, Tokyo, 191-0065, Japan, kainuma@sd.tmu.ac.jp 

In this research, we constructed a cascade reuse hybrid 

manufacturing/remanufacturing model. We proposed the optimal ordering policy 

minimizing manufacturer’s total costs when manufacturing two grades of products. 

In the data examples, comparing the proposal policy with the policy of the actual 

operations of company-A, we could confirm the optimality of the proposal policy. 

2 - Joint Inventory-promotion Decision in Closed-loop Hybrid 

Manufacturing System 

S. Phil Kim, PhD Candidate, Purdue University, 315 N. Grant Street, 


ksphil@purdue.edu, Seokcheon Lee, J. George Shanthikumar 

In this paper, Markov decision process is used for the joint inventory-promotion 

decision problem in a closed-loop hybrid system. The stochastic returns are 

correlated with the volume of circulation. The demands are also stochastic and 

influenced by the promotion decision. The state space can be divided into 

promotion-desired and promotion-not-desired spaces. For a given decision on the 

promotion, we show that the optimal solution structure is completely defined by 

two points in the state space. 

3 - ANP Methodology and Reverse Supply Chain 

Sharon Ordoobadi, University of Massachusetts, 285 Old westport 

Road, Dartmouth, United States of America, sordoobadi@umassd.edu 

The objective is development of a decision tool to help with selection of the third 

party reverse logistics provider. The criteria to be considered in the evaluation 

process are identifgied. The ANP methodology is applied to rank the potential 

providers. The provider with the highest ranking is chosen to perform the reverse 

logistic function. 

4 - Closed-loop Supply Chain with Dynamic Returns and Incentives 

Pietro De Giovanni, NOVA School of Business and Economics, 

Campus Campolide, 1099, Lisbon, Portugal, 

pietro.degiovanni@fe.unl.pt 

In a closed-loop supply chain (CLSC), a single manufacturer and a single retailer 

invest in green activities to enhance the product return. Coordination is evaluated 

by means of a pay-back contract while the return policy is managed in an active 

manner. The return rate is thus modeled as a dynamic equation that evolves over 

time according to the green activities. We investigate the conditions under which 

coordination is Pareto-improving. 

5 - Product Weight Reduction Investment and Collection Rate in a 

Closed Loop Supply Chain 

Jo Min, Iowa State University, IMSE Department, 3004 Black, Ames, 

IA, 50011, United States of America, jomin@iastate.edu, Wenbo Shi, 

Karla Valenzuela 

We investigate a Stackelberg game consisting of a manufacturer/remanufacturer 

who directly sells to customers and a collector of the used products. The collector is 

the follower who determines the collection rate and the manufacturer is the leader 

who determines the price and the product weight reduction investment. As the cost 

saving of remanufacturing increases, the product weight and collection rate both 

increase. i.e., the environmental policies of reduce and reuse may be selfcontradictory.

■ WD08 


Facility Location II 


Chair: Jiamin Wang, Associate Professor, Long Island University, Roth 

Hall 202, Long Island University, 720 Northern Blvd, Brookville, 11548, 

United States of America, jiamin.wang@liu.edu 

1 - A Heuristic Procedure for the Integrated Facility Layout Design and 

Flow Assignment Problem 

Seyed Ali Taghavi, Wayne State University, 4815 Fourth Street, 

detroit, MI, 48202, United States of America, dz3738@wayne.edu, 

Alper Murat 

We study integrated lay-out design problem and product flow assignment problem. 

The lay-out design decisions involve planar location of unequal-area machines with 

duplicates. The product flows are assigned to machines according to the product 

processing routes. We propose a heuristic procedure based on the alternating 

location-assignment heuristic. Then, we apply a perturbation algorithm to escape 

local optima. A sequential location heuristic is also used to speed up the location 

problem. 

2 - Application of Hybrid Analysis of a Discrete Space Location in a 

School Location Problem 

Farshad Majzoubi, PhD Student, University of Louisville, Department 

of Industrial Engineering, JB Speed School of Engineering, 

Louisville, KY, 40292, United States of America, 

f0majz01@louisville.edu, Bulent Erenay, Trivikram Rao 

This research develops a generic Excel model using the Hybrid Analysis method for 

discrete space location problems, and combines it with an Analytical Hierarchy 

Process approach to act as a decision support tool to select the optimal location to 

open a school when both quantitative and qualitative factors are involved. 

Keywords: Location allocation, Hybrid analysis, Analytical Hierarchy Process 

3 - A Minimax Approach to EMS-Helicopter Station and Heliport 

Location Problems 

Takehiro Furuta, Dr., Assistant Professor, Tokyo University of 

Science, 1-3 Kagurazaka, Shinjuku-ku, Tokyo, 162-8601, Japan, 

takef@fw.ipsj.or.jp, Ken-ichi Tanaka 

We propose location models for EMS-helicopter systems. The helicopters require 

their stations and heliports to pick up patients. We formulate the model as an 

integer programming problem which seeks to find locations of both stations and 

heliports to minimize the maximum transportation time. Our model is applied to 

analyzing optimal locations using an idealized city model under various 

assumptions. 

4 - A Median Problem on a Network with Random Travel Speeds and a 

Desirable Travel Time Level 

Jiamin Wang, Associate Professor, Long Island University, Roth Hall 

202, Long Island University, 720 Northern Blvd, Brookville, 11548, 

United States of America, jiamin.wang@liu.edu 

We consider a facility location problem on a network when the travel speeds along 

links are discrete random variables. For each customer, a utility “loss” incurs if 

travel time to reach a facility is beyond a desirable level. The objective is to locate 

facilities so as to maximize the expected total customer utility. A dominant point set 

is identified and solution methods are developed. It is also shown that some classic 

deterministic models are special cases of the problem under study. 

■ WD09 


Service Operations 



Chair: Gad Allon, Northwestern University, 2001 Sheridan Road, 

Evanston, IL, United States of America, gallon@kellogg.northwestern.edu 

1 - The Concert Queuing Arrivals Game: Finite Customer Analysis 

Sandeep Juneja, Associate Professor, Tata Institute, HB Road, Colaba, 

Mumbai, 400005, India, juneja@tifr.res.in, Nahum Shimkin 

We consider a queuing system where a finite number of customers arrive. Each 

customer is free to choose her arrival time (before or after the opening time) and is 

interested in early service completion with minimal wait. We analyze the 

equilibrium behavior of this system and study its convergence to the associated fluid 

limit as the number of customers increases to infinity. 


437 

2 - Bounded Rationality in Service Systems 

Tingliang Huang, Kellogg School of Management, Northwestern 

University, Leverone 529, Jacobs Center, 2001 Sheridan Road, 

Evanston, IL, 60208, United States of America, tinglianghuang@kellogg.northwestern.edu, 

Achal Bassamboo, Gad Allon 

The traditional economics and queueing literature typically assume that customers 

are fully rational. In contrast, in this paper, we study canonical service models with 

boundedly rational customers. We capture bounded rationality using a framework 

in which better decisions are made more often, while the best decision needs not 

always be made. 

3 - Hyperbolic Discounting in a Service System: Implications for Pricing 

& Information Provision 

Erica Plambeck, Professor, Stanford Graduate School of Business, 518 

Memorial Way, Stanford, CA, 94305, United States of America, 

plambeck_erica@GSB.Stanford.Edu, Qiong Wang 

People often lack the self control to undergo an unpleasant service that would be in 

their long-run self interest. This “hyperbolic discounting” has implications for 

optimal pricing, scheduling and whether or not to give a customer real-time 

information about how long he must wait to complete service. 

4 - Advance Selling When Consumers Regret 

Javad Nasiry, HKUST, Clear Water Bay, Kowloon, Hong Kong, Hong 

Kong, Hong Kong - PRC, Javad.NASIRY@insead.edu, Ioana Popescu 

We develop a model to capture the emotional consequences of decision making 

under uncertainty. Negative outcomes trigger regret as consumers reflect ex-post 

what would have been had they decided alternatively. We investigate how regret 

affects consumers’ behavior and how firms can account for consumer regret in 

designing their pricing policies. 

■ WD10 


Pricing Issues in Supply Chain Management 



Chair: Srinagesh Gavirneni, Cornell University, 325 Sage Hall, Ithaca, NY, 

14853, United States of America, nagesh@cornell.edu 

1 - Pricing and Logistics Decisions for a Private-Sector Provider in the 

Cash Supply Chain 

Mili Mehrotra, Assistant Professor, University of Minnesota, 321 

19th Ave. S, Carlson School of Management, Minneapolis, MN, 

United States of America, milim@umn.edu, Vijay Mookerjee, 

Chelliah Sriskandarajah, Milind Dawande 

For secure-logistics providers, the Fed’s cash recirculation policy presents an 

opportunity to offer fit-sorting services to Depository Institutions. We address the 

logistics and joint pricing of the new fit-sorting and the traditional transportation 

services. We characterize the behavior of the optimal prices and quantify the impact 

of the logistics network. 

2 - The Multi-Product Newsvendor Problem with 

Customer-Driven Substitution 

Joonkyum Lee, Cornell University, 301A Sage Hall, Cornell 

University, Ithaca, NY, 14853, United States of America, 

jl883@cornell.edu, Amr Farahat 

We study the stocking problem faced by a newsvendor offering multiple 

substitutable products where a customer’s probability of choosing any given product 

depends on the set of available products at the time of purchase. We present a 

tractable method that is guaranteed to yield an upper bound on the optimal 

expected profit. Numerical tests show that the true expected profits of the solutions 

obtained typically lie within a few percentage points of the upper bound and often 

outperform benchmarks. 

3 - Robust Pricing with Two Substitutable Products 

Zhi-Long Chen, Professor, University of Maryland, 

Robert H. Smith School of Business, College Park, MD, 20742, 

United States of America, ZChen@rhsmith.umd.edu, Ming Chen 

We study a dynamic pricing problem involving two substitutable products. Given 

limited demand information, we use three types of bounds to model demand 

uncertainty. We propose a robust optimization approach for the problem and 

develop a fully-polynomial time approximation scheme based on a DP formulation. 

We report computational results and related managerial insights on how the optimal 

prices change with model parameters.

WD11 

4 - Quality, Inspection, and Pricing Policies in Supply Chains 

Murat Erkoc, Assistant Professor, University of Miami, Miami, FL, 

United States of America, merkoc@miami.edu, Haresh Gurnani 

We consider a decentralized supply chain where market demand depends on the 

supplier’s quality investment and the buyer’s inspection policies. The supplier 

chooses the quality level and the wholesale price, whereas the buyer sets the 

inspection policy and the resale price. Building quality raises costs for the supplier 

and inspection is costly for the buyer. However, they reduce external and internal 

failure costs. We investigate equilibrium quality investment, inspection, and pricing 

policies. 

■ WD11 


Green Supply Chains 



Chair: Feryal Erhun, Stanford University, 380 Panama St, Stanford, CA, 

94305, United States of America, ferhun@stanford.edu 

Co-Chair: Tim Kraft, Stanford University, 920 South California Avenue, 

Palo Alto, CA, 94306, United States of America, tkraft@stanford.edu 

1 - Effect of Carbon Emission Regulations on Transport Mode Selection 

in Supply Chains 

Tarkan Tan, T.Tan@tue.nl, Kristel Hoen, Jan Fransoo, 

Geert-Jan Van Houtum 

We investigate the effect of two regulation mechanisms to drive down carbon 

emissions on the transport mode selection decision: an emission cost and an 

emission constraint. We use an accurate calculation method to determine the 

carbon emissions and incorporate them explicitly in our model. Our results show 

that introducing an emission cost for freight transport, e.g. via a market mechanism 

such as cap-and-trade, will not result in large emission reductions. 

2 - The Carbon-Constrained EOQ 

Saif Benjaafar, Professor, University of Minnesota, 111 Church Street 

SE, Minneapolis, United States of America, saif@umn.edu, Xi Chen 

We examine the impact of carbon emission consideration on the management of 

inventory systems. We do so in the context of the classic economic order quantity 

model (EOQ). We incorporate carbon emission considerations by accounting for 

emissions associated with ordering, purchasing, inventory holding, and sales. We 

examine how different emission control policies (such as strict caps, carbon taxes, 

and carbon trading) affect ordering decisions and the corresponding costs and 

emission levels. 

3 - Replacement Decisions for Potentially Hazardous Substances 

Tim Kraft, Stanford University, 920 South California Avenue, Palo 

Alto, CA, 94306, United States of America, tkraft@stanford.edu, 

Dariush Rafinejad, Feryal Erhun, Robert Carlson 

As public awareness of environmental hazards increases, a growing concern for 

firms is the negative environmental impact of their products. We study the decisions 

of firms and stakeholders when a substance within a product is considered 

potentially hazardous. We find large firms should plan their replacement decisions 

to avoid costs, while small firms should invest to establish niche market positions. In 

addition, NGOs and regulatory bodies should take a pragmatic approach when 

pressuring firms. 

■ WD12 


Different Approaches to Demand Management in Retail 


Supply Chain 


Chair: Gilvan Souza, Associate Professor, Indiana University, Kelley 

School of Business, 1309 E 10th St, Bloomington, IN, 47401, United 

States of America, gsouza@indiana.edu 

1 - Shelf Loathing: Cross Docking at an Online Retailer 

Kyle Cattani, Associate Professor, Indiana University, Kelley School 

of Business, 1309 E 10th St, Bloomington, IN, 47405, United States 

of America, kcattani@indiana.edu, Gilvan Souza, Shengqi Ye 

We analyze basic trade-offs inherent in cross-docking transactions at an online 

retailer. Rather than picking the item from inventory on the warehouse shelves, in 

a cross-docking transaction the item moves directly from the receiving dock to the 

shipping dock. While the cross-docking transaction reduces the shelving and picking 

costs, it potentially increases holding costs and risks changing the customer’s 

expectations for how soon a product will be delivered. 


438 

2 - The Impact of Category Captainship on Retail Assortment 

and Consumers 

Mumin Kurtulus, Assistant Professor, Vanderbilt University, 

Owen Graduate School of Management, 401 21st Avenue South, 

Nashville, TN, 37203, United States of America, 

mumin.kurtulus@owen.vanderbilt.edu, Alper Nakkas 

Category captainship is a practice where a retailer relies on one of the leading 

manufacturers in the category for recommendations on retail assortment. In this 

paper, we investigate the impact of category captainship on the assortment offered 

at the retailer and its impact on the consumers. We identify the conditions under 

which category captainship practices can hurt the consumers. 

3 - Pricing Policy in a Supply Chain: Negotiation or Posted Pricing? 

Goker Aydin, Associate Professor, Indiana University, Kelley School 

of Business, Bloomington, IN, 47405, United States of America, 

ayding@indiana.edu, Hyun-Soo Ahn, Chia-wei Kuo 

This paper examines the choice between posted pricing and negotiation when 

selling to the end customers. We find that the retailer and the manufacturer 

disagree only when the retailer prefers posted pricing, but the manufacturer wishes 

the retailer to use negotiation. Such friction arises when the capacity or the cost of 

negotiation is moderate. Surprisingly, in this region of friction, a decrease in capacity 

or an increase in negotiation costs translates into higher profit for the retailer. 

4 - Motivating Marketing Effort Under Price Delegation: Optimal Retail 

Contract Design 

Shanshan Hu, Asst Professor, Indiana University, Kelley School of 

Business, 1309 E 10th St, Bloomington, IN, 47405, United States of 

America, hush@indiana.edu, Wenbin Wang, Xinxin Hu, 

Robert Jacobs 

Motivated by the interaction between an appliance manufacturer and its regional 

retailers, this paper investigates the contract design problem for the manufacturer. 

We provide answers to three related questions: (a) how to elicit actual demand 

information from the retailer, (b) how to motivate retailer to promote sales, and (c) 

how to construct the contract through relatively simple terms used in practice. 

■ WD15 


Organization Theory II 


Chair: Christopher Rump, Associate Professor, Bowling Green State 

University, Applied Statistics & Operations Research, Bowling Green, OH, 

43403, United States of America, cmrump@bgsu.edu 

1 - Assembly of Successful Teams: Insights From the 

Study of MMORPGs 

Mengxiao Zhu, Northwestern University, 2145 Sheridan Rd, C210, 


mzhu@northwestern.edu, Noshir Contractor, Seyed Iravani 

Teamwork is crucial to accomplish difficult tasks successfully and efficiently. This 

paper investigated the influence of compositional and structural factors on team 

performance. Compositional factors are related to the attributes of team members, 

such as expertise diversity and demographic homophily. Structural factors measure 

the intra/inter team social structures of formal and informal relations. An analysis of 

MMORPG combat teams identifies the impact of these factors on team performance. 

2 - Developing Innovative Capabilities in Biotechnology Firms: 

Internal Building and External Leveraging 

Yuanyuan Wu, Doctoral Student, McGill University, Desautels 

Faculty of Management, 1001 Sherbrooke Street W., Montreal, QC, 

H3A1G5, Canada, yuanyuan.wu@mail.mcgill.ca, Paola Perez-Aleman 

This paper explores the initiation and development of innovative capabilities in 

Montreal-based biotech firms through a multiple-case study design. The existing 

literature separately focuses on internal resources or network leveraging. By 

contrast, this study combines these two aspects, and uncovers different implications 

of the combination on the path and pace of capability development. The results 

extend the role of collaboration in firms with different nature and timing patterns. 

3 - The Evolution of Product Categories: How ‘Spaghetti’ Western 

Impacted American Western Movies 

Moritz Fliescher, PhD Candidate, New York University, Leonard N. 

Stern School of Business, 44 West 4th Street, Room 7-157, 

New York, NY, 10012, United States of America, 

moritz.fliescher@stern.nyu.edu, Gino Cattani 

We add to the category dynamics literature by arguing that established categories 

can evolve. We argue that categories that have an overlapping schema for their 

respective labels have the potential to influence each other. We propose that 

innovations of one category can drive the evolution of another category through 

audience legitimization. We highlight one empirical instance of this by looking at 

the influence of western movies produced in Europe on the meaning of the Western 

genre in the US.

4 - Towards a Systematic Understanding of How Interest-affiliated 

Actors Impact Technology Trajectories 

Theodore Khoury, Oregon State University, Bexell 422B, Corvallis, 

OR, 97331, United States of America, 

ted.khoury@bus.oregonstate.edu, Desiree F. Pacheco 

How do specific actors change technology trajectories? Focusing on actors with 

specific interest-affiliations that are impactful to innovation paths, we propose how 

the strategic actions available to a specific actor’s position can alter the diffusion of 

technological innovations over time. We reinforce our proposed theory with clean 

energy technology illustrations, and consider various influential actor positions in 

both market and non-market roles. 

5 - Constrained Clustering for Departmental Reorganization 

Christopher Rump, Associate Professor, Bowling Green State 

University, Applied Statistics & Operations Research, Bowling Green, 

OH, 43403, United States of America, cmrump@bgsu.edu 

Using data collected from a questionnaire asking how faculty viewed their 

connection to other disciplines in the College of Business, we employed an 

optimization model to reorganize into fewer departmental clusters in order to rectify 

large disparities that have developed over time between departmental faculty sizes. 

We compare these results to those found via traditional hierarchical clustering 

models as well as ad-hoc groupings proposed by members of the college Faculty 

Council. 

■ WD16 


Managing Product Variety 


Chair: Muge Yayla-Kullu, Asst. Prof., RPI, Lally School of Mgmt., 110 8th 

St., Troy, NY, 12180, United States of America, YAYLAH@rpi.edu 

1 - Allocating Capacity Among Quality Differentiated Products: 

Evidence From Airline Industry 

Praowpan Tansitpong, RPI, 110 8th St, Troy, 

United States of America, tansip@rpi.edu, Muge Yayla-Kullu 

This paper explores how the customer perceived quality and the resource 

consumption differences of the products may impact the product line and capacity 

allocation decisions of the firms. We empirically investigate the airline industry in 

three regions of the world; Asia Pacific, EMEA and North America. We find that 

both attributes have a significant impact in all the regions. 

2 - A Unified Framework for Planning a Platform Achieving Both 

Strategic and Operational Benefits 

Changmuk Kang, Ph.D Student, Seoul National University, 599 

Kwanakro, Kwanakgu, Seoul, Korea, Republic of, muk83@snu.ac.kr, 

Yoo S. Hong 

Whereas operational benefit of platform sharing, which is reducing differentiation 

cost, has been well known, its strategic role of establishing commonly preferred 

identity has less been noticed in literature. This study presents a unified framework 

for planning a product platform meeting a firm’s strategic goals of both identity 

establishment and efficient variety offering. A quality function deployment and 

versatility index based approach is applied to identify appropriate platform 

components. 

3 - Assortment Planning with Multiple Quality Levels: A Dynamic 

Programming Approach 

Mark McElreath, Clemson University, 150 Freeman Hall, Clemson, 

SC, 29634, United States of America, mmcelre@clemson.edu, 

Maria Mayorga 

The optimal solution to the assortment planning problem with vertical and 

horizontal differentiation in which consumer preference is described by a locational 

choice model is unknown. We propose a two part solution: a dynamic program to 

find the optimal vertical attributes embedded into a line search to find the optimal 

horizontal attributes. We compare our approach to metaheuristics, explore the 

optimal solution space, and provide insight into the properties of an optimal 

assortment. 

4 - Product Variety, Out-Of-Stock, and Sales 

Xiang Wan, University of Maryland, Van Munching Hall 3354, 

College Park, MD, 20742, United States of America, 

xwan@rhsmith.umd.edu, Martin Dresner, Philip Evers 

Product variety has been analyzed by an extensive body of literature. High product 

variety stimulates sales, while it raises the difficulty of logistics management and 

reduces sales. This paper proposes a framework to investigate the direct and indirect 

(through logistics performance) impacts on sales. The results indicate that the 

influence of product variety on sales depends on variety in both the degree and 

dimensions. 


439 

■ WD17 


Manufacturing III 


Chair: Zulal Gungor, Prof., Gazi University, Maltepe, Ankara, Turkey, 

zulalg@gazi.edu.tr 

1 - Total Productive Maintenance Policy Framework and 

Implementation of 5s Rules 

Bahar Ozyoruk, Assistant Prof.Dr., Gazi University, Faculty of 

Egineering, Department of Industrial Engineering Maltepe, Ankara, 

06570, Turkey, bahar@gazi.edu.tr 

Total Productive Maintenance (TPM) has attracted the attention of industries all 

over the world. Many companies benefit from this policy by increasing the overall 

efficiency of machine is trying to use existing capacity more efficiently. In this study, 

a firm which produces foam in the process of transition to implement TPM and 5S 

implementation of the rules are discussed. The results obtained were presented as 

comparatively. 

2 - Multi-objective Decision Making (MODM) Approach in Optimizing 

Product Design by the Help of House of Quality: A Case Study 

Zulal Gungor, Prof., Gazi University, Maltepe, Ankara, Turkey, 

zulalg@gazi.edu.tr, Elif Kilic 

In practice, the values of design requirements (DRs) having only a few alternatives 

can be discrete in the Quality Function Deployment. We propose a Mixed Integer 

Goal Programming (MIGP) formulation to get the optimum solution from a limited 

number of DRs alternatives. The solution of MIGP model provides decision makers 

with different alternative results by the usage of the lexicographic goal 

programming (LGP) approach. The applicability of the proposed models is 

demonstrated with a problem. 

■ WD18 


Business Applications 


Chair: Alba Bonko, President, Biz Intelligence Solutions, LLC, 9737 NW 

41st Street, Miami, FL, 33178, United States of America, 

alba_n_nunez@hotmail.com 

1 - Work as a Love Object. A New Framework for Analyzing the 

Work-self Relation 

Brad Almond, Assistant Professor of Management, Texas A&M 

University, Division of Business, 1901 S. Clear Creek Rd., Killeen, 

TX, 76549, United States of America, brad.almond@ct.tamus.edu 

Building on popular fascination with the idea of “loving what you do,” this paper 

explores how work functions as a love object by developing a multi-dimensional 

model and testing its structure and relation to key work outcomes using factor 

analysis and structural equation modeling. Explores implications for management 

and organizations. 

2 - Allocation of Bulk Tanks to Customer Sites 

Tejinder Pal Singh, Sr. Research Associate, American Air Liquide, 

12800 W. Little York Rd, Houston, TX, 770941, United States of 

America, tejinder.singh@airliquide.com, Kimberly Ellis 

Bulk tank allocation (BTA) problem consists of allocation of tanks to customer sites 

for distribution of gases. Distribution is based mainly on Vendor Managed Inventory 

(VMI) model. In BTA problem, the goal is to minimize the distribution costs by 

having right tank sizes at customer sites. A right tank size at a customer will ensure 

that the customer doesn’t run out of the product often and at the same time doesn’t 

need deliveries frequently. 

3 - Optimal Contract Problems in Online Advertising with 

Risk Considerations 

Md. Tanveer Ahmed, University at Buffalo, SUNY, 333 Bell Hall, 

Buffalo, NY, 14260, United States of America, 

mtahmed@buffalo.edu, Changhyun Kwon 

In this paper, we study the optimal contract problem for online display 

advertisements with pay-per-view pricing scheme. We first provide and analyze a 

single contract model, which is shown to be equivalent to the newsvendor problem. 

We then consider a stochastic optimization problem with two different contracts and 

show that no mixed contract is optimal. However, we show that a mixed strategy 

may be optimal when we consider the risk attitude of the publisher.

WD19 

4 - Development of a Global Part Sourcing Optimization Model with 

Currency Risk Analysis 

Don Zhang, Cost Optimization Analyst, Ford Motor Company, 2101 

Village Rd., Dearborn, MI, 48121, United States of America, 

xzhang35@ford.com, Mark Everson, Leonardo Vaquero, 

Dawn Gontko, David Shepps 

Ford Motor Company operates globally across six continents. Under the “One Ford” 

initiative, Ford has been leveraging increased economies of scale. One of these 

efforts is to source globally the common parts shared by many vehicle programs. We 

have developed a mixed integer programming optimization model and exchange 

rate risk assessment methodology to optimize sourcing decisions. We will discuss 

our approach for deciding the optimal sourcing decisions and risk assessment for 

various scenarios. 

5 - Continuous Improvement and Efficiency Training: 

Challenges and Opportunities 

Alba Bonko, President, Biz Intelligence Solutions, LLC, 9737 NW 

41st Street, Miami, FL, 33178, United States of America, 

alba_n_nunez@hotmail.com, Rebeca Lergier, Martha Centeno 

We discuss paradigms used in training courses, and highlight some of their strengths 

and weakness. Training courses paradigms range from the “recipe” to the 

“academic”. The former disregards foundations of the methods, whereas the latter 

fails to show practical applications. We propose a scheme that is enterprise-centered, 

so that practicing professionals can effectively use quantitative methods, and with 

management policies, as part of the continuous improvement cycle of the 

enterprise. 

■ WD19 


Practical Applications of Pricing and RM Theory 



Chair: Scot Hornick, Partner, Oliver Wyman, 155 N. Wacker Drive, 

16th Floor, Chicago, IL, 60606, United States of America, 

Scot.Hornick@oliverwyman.com 

1 - Market-Response-Based Inventory Management for Airlines 

and Hotels 

James Rider, Associate Partner, Oliver Wyman, 55 Baker Street, 

London, W1U 8EW, United Kingdom, 

James.Rider@OliverWyman.com, Jessica McLaughlin, Bejugum Rao, 

Daniel Sack 

Most of the prevalent seat and room inventory management methods are no longer 

adequate to serve the needs of today’s dynamic marketplace. In our novel Market 

Response-Based Inventory Management (MRBIM) approach, we generate 

recommendations for pricing and availability decisions by taking into account 

dynamic marketplace data. Demand is managed more effectively by understanding 

customer’s willingness to pay and knowing prevailing competitor prices and 

availability. Revenue is maximized by adjusting own fares/rates and availability, 

possibly multiple times a day. To facilitate MRBIM implementation, we developed 

statistical methods to estimate demand parameters with data that is likely to be 

readily available at a typical airline or hotel company. We developed approaches to 

interface with current inventory allocation models. Piloting of MRBIM approach at 

multiple airlines has revealed consistent revenue benefits. 

2 - Dynamic Pricing for Aftermarket Auto Parts 

Bruce Spear, Associate Partner, Oliver Wyman, 1166 Avenue of the 

Americas, New York, NY, 10036, United States of America, 

Bruce.Spear@OliverWyman.com, Todd Ebe, Bejugum Rao 

We present an application of dynamic pricing at an auto-glass wholesaler. We 

estimate customer’s sensitivity to the wholesaler’s price by customer type, part type 

and geography. We determine optimal prices that maximize the wholesaler’s 

contribution, i.e. revenue minus cost of goods sold, knowing competitors’ prices. We 

show the revenue benefits (>4%) due to optimal prices evaluated in a test-control 

setting. We discuss data and measurement challenges. Our future efforts will include 

extending the approach to retail auto-glass business as well as optimizing part prices 

by considering own and competitor inventory levels. 

3 - Revenue Management Concepts for Freight Railways Practicing 

Dynamic Routing 

Marc Meketon, Oliver Wyman, 212 Carnegie Center, Princeton, NJ, 

08540, United States of America, Marc.Meketon@oliverwyman.com, 

David Lehlbach 

Most railways world-wide operate on a fixed scheduling plan that allows only one 

route per railcar. However, as railways begin to plan their next generation of carscheduling 

systems, they should consider the possibility of dynamic train scheduling 

that changes the operating plan and the rules for routing cars dynamically. As a 

simple example, when faced with capacity limitations in a certain corridor, it may 

make sense to divert some cars to an alternate route. While dynamic planning could 

increase efficiencies, it could also change customer service. Some customers could 


440 

get shorter transit times, others longer. Since service levels change, any move 

towards dynamic planning should involve pricing and revenue management 

considerations. This talk discusses a number of these considerations and gives 

guidance to railways looking to make the jump to the next level of planning. 

■ WD20 


Pricing and Revenue Management III 


Chair: Dincer Konur, PhD Candidate, University of Florida, 

Industrial and Systems Engineering Department, Gainesville, FL, 32611, 

United States of America, dincer@ufl.edu 

1 - A Segmentation Study in Applying Revenue Management to the 

Hospitality Industry 

Murtaza Das, The Rainmaker Group, 5755 North Point Parkway, 

Alpharetta, GA, 30022, United States of America, 

murtazadas@gmail.com, Renaud Menard 

Segmentation is a fundamental step in applying revenue management and pricing 

which directly impacts forecasting accuracy, and thus revenue optimization. While 

forecast accuracy and revenue optimization are paramount to a sound revenue 

management strategy, operational limitations should also be considered in the 

segmentation analysis. We review the current approaches of segmentation in the 

hospitality RM industry, present simulation results, and point out challenges. 

2 - A Statistical Methodology to Find Segment Level Parameters using 

Aggregate Level Data 

Hamed Hasheminia, Sauder School of Business-UBC, 

2053 Main Mall, Vancouver, BC, V6Z2T9, Canada, 

hamed.hasheminia@sauder.ubc.ca, David Gillen 

We develop a novel statistical procedure to estimate segmental demand functions 

from aggregate level data. To achieve our result we combine the invaluable 

information hidden in integer numbers, characteristics of integer numbers, and use 

MLE to estimate parameters at the segment level. The method is applied to estimate 

how demands for different segments of passengers (i.e. single passengers, couples, 

and so forth) are affected by price, time to the flight, etc. 

3 - Modeling Supplier Wholesale Pricing Decisions with Competitive 

Buyers Under Cournot Competition 

Dincer Konur, PhD Candidate, University of Florida, Industrial and 

Systems Engineering Department, Gainesville, FL, 32611, 

United States of America, dincer@ufl.edu, Joseph Geunes 

We model a supplier’s wholesale pricing decision for competitive non-identical 

buyers as a Stackelberg game. To determine a Stackelberg equilibrium, we first solve 

the buyers’ game and analyze the buyers’ quantity decisions under two different 

cooperation levels. We then determine the supplier’s optimal wholesale price, and 

conduct numerical studies to characterize the value of the information to the 

supplier, as well as the effects of buyer and market heterogeneity. 

■ WD22 


Service System Development 



Chair: Ari P.J. Vepsalainen, Professor, Aalto University School of 

Economics, Department of Business Technology, P.O. Box 21220, 

Helsinki, 00076, Finland, ari.vepsalainen@hse.fi 

1 - Functional Selection of Business Processes - Who Works 

for the Market? 

Ari P.J. Vepsalainen, Professor, Aalto University School of Economics, 

Department of Business Technology, P.O. Box 21220, Helsinki, 

00076, Finland, ari.vepsalainen@hse.fi, Mika Raulas, 

Markku Tinnilä, Jukka Kallio 

The analysis of the functional needs of different communities and market processes 

highlights the full market potential of service development. With extreme functional 

specialization, companies and workers will be providing narrower service to 

increasing number and variety of market activities and communities. Our case 

studies illustrate the working conditions of functional selection and the potential coevolution 

of private, public and commercial institutions.

2 - Dynamic Policies in Knowledge-Based Service System with 

Feedback Information 

Qifeng Shao, Northwestern University, 2145 Sheridan RD, IEMS 

C210, Evanston, IL, 60208, United States of America, 

qshao@northwestern.edu, Seyed Iravani 

We propose a modeling framework of a knowledge based service system with an 

information feedback mechanism. The feedback system provides information 

regards correctness of the agent’s decisions. With help from the feedback system, the 

agent adjusts processing strategies to deal with information crises that can change 

the state of arrival customers. Besides the optimal strategy, we examined several 

heuristics and found one effective policy with a simple threshold design. 

3 - A Simulation Based Framework for Service Facility Internal 

Layout Design 

Ming Xie, IBM Research - China, Diamond Building, ZGC Software 

Park, Beijing, China, xieming@cn.ibm.com, Jinyan Shao, Bin Zhang, 

Wenjun Yin, Jin Dong 

Service facilities, including bank branches, supermarkets, etc. are closely related to 

our lives. Whereas, it is a difficult task to make decisions on how to design internal 

layout and configuration for them. In this paper, we propose a service facility 

internal layout optimization framework as well as how to model the system into a 

multi-agent system. Then, a bank branch scenario from real world is adopted to 

demonstrate the methodology and implementation. 

■ WD23 


Retail Managment 


Chair: Mahesh Kumar, Assistant Professor, R.H. Smith School of Business, 

University of Maryland, 4321 Van Munching Hall, College Park, MD, 

20742, United States of America, kumarmahesh@gmail.com 

1 - Assortment Planning of Configurable Products 

Edward Umpfenbach, PhD Student, Wayne State University, 

20011 12 Mile, Roseville, MI, 48066, United States of America, 

as6964@wayne.edu, Alper Murat, Ratna Babu Chinnam 

Considerable work has been done to optimize the assortments of high sales density 

products. We introduce a method to estimate demand and substitution parameters 

of a configurable product given sales data, then solve a joint supply chain planning 

and assortment planning problem. 

2 - The Impact of Execution Errors on Inventory Record Inaccuracy and 

Retail Out-Of-Stock 

Howard Hao-Chun Chuang, Doctoral Student, Mays Business 

School, Texas A&M University, Wehner 301C - TAMU 4217, 


hchuang@mays.tamu.edu, Rogelio Oliva 

We model a continuous (Q,R) inventory system and analyze the impact of 

execution errors on inventory record inaccuracy and retail out of stock. The 

existence of multiple errors could disrupt ordering decisions and reduce on-shelf 

availability. We formalize the argument analytically and develop a system dynamics 

model. We perform Monte-Carlo sensitivity simulation to identify the most costly 

execution errors. Then we provide suggestions to improve operational accuracy 

within a retail store. 

3 - A Pooling Procedure to Improve Sales Forecasts for Retail Fashion 

Goods with Limited Past Sales Data 

Mahesh Kumar, Assistant Professor, R.H. Smith School of Business, 

University of Maryland, 4321 Van Munching Hall, College Park, MD, 

20742, United States of America, kumarmahesh@gmail.com 

Most sales forecasting methods are likely to perform poorly if only a small amount 

of past sales data is available for analysis. The situation is common for new fashion 

items in a store, especially during the early weeks of a selling season. We overcome 

this limitation of a forecasting method by pooling data from groups of similar items 

using a two-step clustering procedure. The new procedure is flexible to use and is 

computationally fast. Its effectiveness is demonstrated on a real-world data. 


441 

■ WD24 


Joint Session SPPSN/ TSL: Decision Support for 

Emergency Response and Public Safety 

Sponsor: Public Programs, Service and Needs/ Transportation 

Science and Logistics Society 


Chair: Alex Savachkin, Assistant Professor, University of South Florida, 

4202 E. Fowler Avenue ENB 118, Tampa, FL, 33620, United States of 

America, alexs@usf.edu 

1 - Non-Pharmaceutical Interventions (NPI) for the Mitigation of 

Pandemic Influenza 

Dayna Martinez, Doctoral student, University of South Florida, 

Tampa, FL, 33620, United States of America, dlmartin@mail.usf.edu, 

Tapas K. Das, Alex Savachkin 

In the event of an influenza pandemic, non-pharmaceutical interventions (NPI), 

such as social distancing, will likely be the only effective containment measure 

available in the early phase of the pandemic. In this research, we examine various 

NPI strategies, such as quarantine of isolated cases, household quarantine, school 

and workplace closures, and study their effect on the infection attack rate and the 

societal and economic cost of the pandemic. 

2 - Developing an Agent-based Model for Poliovirus Transmissions for 

Post-eradication Outbreak Response 

Hazhir Rahmandad, Assistant Professor, Virginia Tech, 7054 Haycock 

Rd., room 430, falls church, va, 22043, United States of America, 

hazhir@vt.edu, Kimberly Thompson, Radboud Duintjer-Tebbens, 

Kun Hu 

Given the possibility of reintroductions of live polioviruses into communities after 

eradication, designing effective and efficient responses to potential outbreaks is 

necessary. In this study we develop an individual-based simulation model of 

poliovirus transmission dynamics in a population and use this model the explore 

alternative response strategies. We also explore the importance of different 

assumptions with regard to human contact network and transmission mechanisms. 

3 - Agent-based Simulation of Mass Egress From Large Public Events: 

Current State and Next Steps 

Douglas A. Samuelson, President and Chief Scientist, InfoLogix, Inc., 

Annandale, VA, United States of America, 

samuelsondoug@yahoo.com 

We review a variety of recent work on agent-based simulation of mass egress, 

especially from sports arenas, noting the capabilities, limitations and design 

compromises in some of the most interesting models. We then discuss next steps, 

including modeling group movement, movement by emergency responders, 

placement of treatment centers, effects of toxic air- and water-borne plumes, and 

integration of planning and training models with real-time crisis management 

information systems. 

4 - Decision Support Systems for Pandemic Influenza (PI) Surveillance 

Alfredo Santana-Reynoso, PhD Candidate, University of South 

Florida, 4202 E. Fowler Avenue ENB118, Tampa, FL, 33620, 

United States of America, asantan2@mail.usf.edu, Diana Prieto, 

Alex Savachkin, Sharad Malavade 

PI contingency plans have concentrated their efforts into a new virus strain of low 

transmissibility and high severity originated in SE Asia. The H1N1 2009 outbreak 

featured higher transmissibility and lower severity than expected, and was 

originated in North America. This presentation analyzes how these scenarios have 

been managed by the current PI surveillance systems. Robust multi-epoch decision 

support systems for PI surveillance adaptable to a more diverse range of situations 

are presented.

WD25 

■ WD25 


Transportation, Planning III 


Chair: Wei Fan, Assistant Professor, The University of Texas at Tyler, 

Department of Civil Engineering, 3900 University Blvd., Tyler, TX, 75799, 

United States of America, wfan@uttyler.edu 

1 - Expansion Planning of Road Networks via 

Agent-Based Optimization 

Alireza Kabirian, University of Alaska - Anchorage, Rasmuson Hall, 

Suite 318D, 3211 Providence Dr, Anchorage, AK, 99508, 

United States of America, a_kabirian@yahoo.com 

We develop optimization models and methodologies for expansion planning of an 

existing network of roads over a long-run horizon. Specifically, we aim at 

determining where and when new roads should be constructed and which and 

when old roads should be reconstructed to minimize public construction costs as 

well as travelers’ expenses. This decision support model helps transportation 

managers in public sector make comprehensive and data-driven decisions about 

expansion of road infrastructures. 

2 - The Changes in Users’ Perceptions on Transportation Problems in 

the Last 10 Years 

Xiaoyu Zhu, PhD Candidate, University of Florida, 365 Weil Hall, 

P.O. Box 116580, Gainesville, FL, 32611, United States of America, 

shuxy03@ufl.edu 

To improve the transportation system, it is important to understand the travelers’ 

perceptions about transportation issues to assess user acceptance of the policies. 

With developments in transportation, the perceptions might change greatly in the 

last decade. To understand the people’s perceptions about the most important 

transportation issue will provide more information for value the project efficiency 

and public acceptance. 

3 - A Sensor-based Vehicle Routing Problem Algorithm 

Chrysafis Vogiatzis, University of Florida, Weil 303, Gainesville, FL, 

32611, United States of America, chvogiat@ufl.edu, Panos Pardalos 

Vehicle routing has always been a vital problem for transportation and traffic 

congestion. Throughout the years, solutions have been proposed but are usually 

hard to implement because of the continuous time nature of the problem, which 

cause alterations in conditions and constraints. We propose an online algorithm 

which searches for an optimal solution based on the input provided by vehicle to 

vehicle sensor communication using augmented lagrange relaxation. 

4 - Optimal Toll Design with Uncertain Demand 

Wei Fan, Assistant Professor, The University of Texas at Tyler, 

Department of Civil Engineering, 3900 University Blvd., Tyler, TX, 

75799, United States of America, wfan@uttyler.edu 

An optimal toll design problem (OTDP) under a second-best link based congestion 

pricing scheme is discussed in this presentation, in which the selection of both toll 

locations and toll levels needs to be optimally determined with uncertain demand. A 

bi-level optimization model is formulated and a simulation-based genetic algorithm 

procedure is proposed to solving the OTDP. Numerical results are presented and 

future research directions are also given. 

■ WD27 


Networks and Graphs 


Chair: Alex Sprintson, Assistant Professor, Texas A&M University, TAMU 

3128, College Station, 77843, United States of America, spalex@tamu.edu 

1 - Screen Sizes of Paths 

Anthony Harrison, Texas State University, 601 University Drive, 

San Marcos, United States of America, ah1411@txstate.edu, 

Nathaniel Dean 

Sphere of infuence graphs (SIGs) are a representation of spatial relationships 

between points. They can be useful in areas related to pattern matching. We 

consider a graph invariant for SIGS called the screen size. This is the smallest integer 

k such that a given SIG can be realized on a k by k grid. We find the screen size for 

paths and then explore integer programming formulations for the problem and 

explain some of the difficulties encountered with this approach. 


442 

2 - Reliable Networks That Can be Modeled by Graphs with Edge 

Connectivity Equal to Two 

Leandro Teixeira, Centro de Anàlises de Sistemas Navais, Praça Barao 

de Ladàrio, AMRJ , Centro, rio de janeiro, RJ, 20091-000, Brazil, 

leandrodteixeira@yahoo.com.br, Nair Abreu, Leonardo Lima 

Deterministic and probabilistic parameters can be used to measure the reliability of a 

graph that models a network. Consider that each node of the network is reliable 

and its failure is related to the probabilities of the edge failures of the graph when 

these probabilities occur randomly and independently. In this work, a class of graphs 

with maximum reliability among all graphs with edge connectivity equal to two is 

presented. 

3 - An Efficient Algorithm for the Distributed Data Retrieval Problem 

Alex Sprintson, Assistant Professor, Texas A&M University, TAMU 

3128, College Station, 77843, United States of America, 

spalex@tamu.edu, Zakia Asad, Mohammad Asad R Chaudhry 

We consider the problem of accessing large volumes of data stored on multiple 

locations across a storage network. We assume that data is stored in either original 

or encoded form on multiple servers across the network. Our goal is to find a set of 

disjoint paths of minimum total cost that connect the client with some of the servers 

across the network such that client is able to retrieve the required data. We present 

an efficient polynomial-time algorithm for this problem. 

4 - An Efficient Heuristic Algorithm for Minimum Labeling Spanning 

Tree Problem 

Kaveh Farokhi Sadabadi, Faculty Research Assistant, University of 

Maryland, ENCE Department, 1173 Glenn L. Martin Hall, College 

Park, MD, 20742, United States of America, kfarokhi@umd.edu, 

Ali Haghani 

We propose a heuristic algorithm to solve Generalized Minimum Labeling Spanning 

Tree problem. This is an NP hard problem with applications in telecommunications 

and network design. Many algorithms are proposed to solve GMLST. Maximum 

Vertex Covering (MVCA) and Genetic Algorithm (GA) are well-known. Global 

optimality of these methods is not guaranteed, nor are they efficient to deal with 

large problems. Reported examples show the efficiency and quality of the proposed 

method compared to others. 

■ WD28 


Computer Models: Prediction and Calibration 



Chair: Lulu Kang, Assistant Professor, Illinois Institute of Technology, 

Engineering 1 Building, Applied Math, 10 West 32nd Street, Chicago, IL, 

60616, United States of America 

1 - Model Calibration with Minimal Adjustments 

Chia-Jung Chang, PhD Student, Georgia Institute of Technology, 

755 Ferst Drive, Atlanta, GA, United States of America, 

cchang43@gatech.edu, Roshan Vengazhiyil 

Model calibration refers to estimating unknown parameters in a physics-based 

model from real data. When model assumption is violated, the estimates become 

inaccurate leading to poor model prediction. Besides, all works ignore the 

potentially important bias that can occur in the observations. In this work, we 

develop a methodology for calibrating the physical model in the presence of both 

model and experimental biases. Two real case studies are presented to demonstrate 

the prediction ability. 

2 - Improving Prediction by Integrating Analytical Models with Finite 

Element Models 

Shan Ba, Georgia Institue of Technology, 765 Ferst Drive, NW, 

Atlanta, GA, 30332, United States of America, shan.ba@gatech.edu, 

Roshan Vengazhiyil, Ramesh Singh 

For the micro grooving process, we develop an innovative two-step design of 

experiments approach to efficiently extract information from both the simple 

analytical force model and the complex finite element model. The simple but less 

accurate analytical model can be adjusted by data from the time-consuming finite 

element simulations, and finally forms an effective surrogate micromachining 

cutting model which is fast, accurate and economically viable to use. 

3 - Kernel Sum Regression and Interpolation 

Lulu Kang, Georgia Institute of Technology, Atlanta, GA, 30332, 

United States of America, lkang@isye.gatech.edu, 

Roshan Vengazhiyil 

In this paper, we propose a new nonparametric regression method, called kernel 

sum regression method. It utilizes iterative kernel regression to achieve better 

prediction accuracy. Meanwhile, if the the number of kernel regression goes to 

infinity, the kernel sum regression becomes an interpolation method. We also 

investigated some properties of the kernel sum regression and interpolation 

methods, and provide algorithm to estimate the bandwidth parameters.

■ WD29 


Quality Improvement in Complex Systems 



Chair: Ran Jin, Georgia Institute of Technology, 755 Ferst Dr. NW, 

Atlanta, United States of America, jinr@gatech.edu 

1 - Challenges of using Feature Selection Methods for 

Waveform Signals 

Nasim Arbabzadeh, Rutgers University, 23667 BPO WAY, Piscataway, 

NJ, 08854, United States of America, nasim@eden.rutgers.edu, 

Susan Albin 

Recent technological advances in data capture and processing result in highdimensional 

waveform signals, making the analysis difficult for the process 

engineers. Feature selection methods can be used to reduce the features to a 

manageable number. In this paper, different feature selection methods have been 

compared on several real datasets, including NIR spectra and vertical depth profile, 

and the impact of different sampling frequencies on their performances has been 

investigated. 

2 - Generalized Selective Assembly 

Matthias H Tan, Student, Jeff Wu/ Department of Industrial & 

Systems Engineering, Georgia Institute of Technology, 301 10th 

Street NW, Apt 208A, Atlanta, GA, 30318, United States of America, 

mtan6@gatech.edu, Jeff Wu 

This paper develops a generalized version of selective assembly, called GSA, for 

improving the quality of assemblies of single units of different component types. 

Two variants are considered: direct selective assembly, which uses measurements on 

component characteristics, and fixed bin selective assembly, which only requires 

sorting components into bins. We formulate the problem of matching N components 

of each type to give N assemblies that minimize quality cost as linear integer 

programs. 

3 - Relationship Between Seamless Tube Quality and Piercing 

Vibration Data 

Weidong Zhang, Vice Director and Associate Professor, National 

Cecter for Materials Service Safety,University of Science and 

Technology Beijing, 30 Xueyuan Road, Haidian District, Beijing, 

100083, China, zwd@ustb.edu.cn, Ran Jin, Jianjun Shi 

We can get vibration data of seamless tube piercing by two cameras. There may be 

some relationship between quality of seamless tube and vibration data. We can find 

the relationship by system informatics method. 

■ WD30 


Statistics/Quality Control III 


Chair: David Han, Assistant Professor, University of Texas at San Antonio, 

One UTSA Circle, San Antonio, TX, 78249, United States of America, 

david.han@utsa.edu 

1 - A Variable Sampling Hotelling T2 Chart for Monitoring Simple Linear 

Quality Profiles 

Galal Abdella, PhD Student, Wayne State University, 4815 4th Street, 

Detroit, MI, 48202, United States of America, bb2941@wayne.edu, 

Kai Yang, Adel Alaeddini 

We design a variable sampling T2 scheme to enhance the detecting speed of offtarget 

conditions while keeping the total number of samples low. We constructed an 

optimization model solved by using the genetic algorithm. The performance of the 

proposed scheme is compared with its fixed sampling counterparts under different 

conditions. 

2 - Beta Model-based Control Chart for Fraction Monitoring 

Carla ten Caten, Profa. PhD., PPGEP/UFRGS, Av. Osvaldo Aranha, 

99 - 5° andar, Porto Alegre, RS, 90.035-190, Brazil, 

tencaten@producao.ufrgs.br, Angelo Sant’Anna, Michel Anzanello 

Modelo-based control charts often use multiple regression models. We propose a 

Beta model-based control chart (BMCC) for monitoring fraction that varies with the 

adjustment of control variables. The BMCC monitors deviances on Beta model’s 

residuals. We use sensitivity analysis to compared BMCC with Hawkins (1991) and 

Haworth’s(1996) methods. 


443 

3 - Analysis Methods for Non-regular Fractional Factorial Designs 

Shilpa Shinde, Arizona State University, 1249 E Spence Avenue 

Apt#343, tempe, AZ, 85281, United States of America, 

scvmadha@asu.edu, Douglas Montgomery 

Non-regular designs are very effective alternatives to regular fractional factorials. 

Their use is restricted due to the lack of robust analysis techniques. We present a 

new analysis and variable selection technique which will help identify significant 

effects in a non-regular design. 

4 - Exact Inference for Progressively Type-I Censored Exponential 

Failure Data 

David Han, Assistant Professor, University of Texas at San Antonio, 

One UTSA Circle, San Antonio, TX, 78249, United States of America, 

david.han@utsa.edu, N. Balakrishnan, G. Iliopoulos 

Progressively Type-I censored life-test is discussed under the assumption of 

exponential distribution. For small sample sizes, a practical modification is proposed 

to guarantee a feasible test under this scheme. We then obtain the exact sampling 

distribution of the MLE of the mean parameter under the condition ensuring its 

existence. Using the exact method as well as the asymptotic and bootstrap methods, 

we then discuss construction of confidence intervals and their performance via 

simulations. 

■ WD31 


Quality Management I 


Chair: Karl Majeske, Associate Professor of Quantitative Methods 

Management, Oakland University, School of Business Administration, 

Rochester, MI, 48309, United States of America, majeske2@oakland.edu 

1 - CUSUM Control Charts for Respiratory Syndromic Surveillance 

Huifen Chen, Professor, Chung-Yuan University, Department of 

Industrial and Systems Eng., Chung-Yuan University, Chungli, 320, 

Taiwan - ROC, huifen@cycu.edu.tw, Chaosian Huang 

This work applies CUSUM charts for detecting outbreaks of the respiratory 

syndrome. The data, daily ambulatory-care visits based on population of size 160 

thousand, are from Taiwan National Health Insurance Research Database. We 

construct a standardized CUSUM chart based on residuals of a fitted regression 

model with an ARMA error term using the 2005 and 2006 data. The CUSUM chart 

seems to be able to detect aberrations of respiratory syndrome when used to 

monitor the 2007 and 2008 data. 

2 - An Integrated Approach for Multivariate Process Control in 

Automobile Manufacturing 

Xiaoyu Ma, PhD Student, Industrial&manufacturing engineering 

department, wayne state university, 4815 4th Street, detroit, MI, 

48202, United States of America, eb1946@wayne.edu, Alper Murat, 

Kai Yang, Adel Alaeddini 

Increasing availability of multivariate process data demands us to develop effective 

analysis techniques. We developed a comprehensive approach that uses statistical 

and optimization techniques. It performs abnormal signal detection, influential 

features selection in optimized fashion during multivariate statistical control 

practice. This approach can help root cause analysis in automobile manufacturing 

process and examples with real data are presented. 

3 - Sustaining Quality Performance Through Collective Mindfulness 

Hung-Chung Su, Student, University of Minnesota, 3-150 321-19th 

Avenue South, Minneapolis, MN, 55455, United States of America, 

suxxx051@umn.edu, Kevin Linderman 

Organizational mindfulness is an emerging construct that has not receive enough 

attention in the quality management. We provide evidence of the five dimensions of 

organizational mindfulness and examine its impact on innovation, improvement 

and the reliability of quality performance. We discuss the implications for sustaining 

quality performance and possible contributions to the organizational learning and 

dynamic capability literature. 

4 - Dynamic Part Fitting to Improve Automotive Body Quality 

Karl Majeske, Associate Professor of Quantitative Methods 

Management, Oakland University, School of Business 

Administration, Rochester, MI, 48309, United States of America, 

majeske2@oakland.edu 

The gaps and flushness between assemblies, such as door-to-fender, represent a 

highly visible automotive body quality characteristic. Traditional design and 

assembly methods attach a door to the body by bolting hinges into established 

mounts. This technique uses vision system data to orient parts during final assembly 

to optimize quality.

WD33 

■ WD33 


Optimization Models in Data Mining with Applications 

in Biomedicine 



Chair: Petros Xanthopoulos, University of Florida, 303 Weil Hall, 

P.O. Box 116595, Gainesville, FL, 32611, United States of America, 

petros.xanthopoulos@gmail.com 

1 - Early Detection of Cardiovascular Disease 

Tsung-Lin Wu, tlwu@isye.gatech.edu, Eva Lee 

Cardiovascular diseases are the No 1 cause of death in US. Coronary heart disease is 

caused by atherosclerosis, the narrowing of the coronary arteries due to fatty build 

ups of plaque, producing angina pectoris, heart attack or both. We present novel 

early detection based on classification of traditional risk factors and novel 

biomarkers. The results can detect risk conditions of arteries and allow clinicians to 

perform early intervention. This work is joint with cardiologists at Emory U. 

2 - Robust Data Mining with Application in Biomedicine 

and Engineering 

Petros Xanthopoulos, University of Florida, 303 Weil Hall, P.O. Box 

116595, Gainesville, FL, 32611, United States of America, 

petros.xanthopoulos@gmail.com, Panos Pardalos, Mario Guarracino 

Supervised and unsupervised learning under uncertainty has become a very 

important problem in data analysis. In this talk we present some algorithms based 

on robust optimization for addressing data uncertainty issues. 

3 - Rule Extraction From Support Vector Machines and Applications to 

Medical Diagnosis 

Sara Nourazari, University of Oklahoma, 100 East Boyd, SEC T 301, 

Norman, OK, 73019, United States of America, sara.n@ou.edu, 

Theodore Trafalis 

Among the classification methods, the propositional if-then rules are very popular 

providing a transparent classification decision. In the case of high-dimensional data, 

SVMs often perform significantly better but suffer from incomprehensibility. In this 

work we evaluate different techniques to extract expressive rules from SVMs and 

apply those in real life cases such as medical diagnosis. 

4 - Network Based Models for Analysis of SNPs 

Zeynep Ertem, Texas A&M University, 1313 Zachry, College Station, 

TX, 77840, United States of America, zeynepsertem@gmail.com, 

Sergiy Butenko 

Constructing associated graph-theoretic models becomes more important as 

communicating vast amount of information accumulated in the laboratories each 

day is getting harder. Single Nucleotide Polymorphisms (SNPs) are of paramount 

importance in DNA related studies due to their role in variation of species. In this 

talk, we first survey metwork based models arising in computational biology and 

then concentrate on applications of cluster-detection algorithms to analyze SNP 

data. 

5 - Parametric Support Vector Machines 

Altannar Chinchuluun, Dr, Centre for Process Systems Engineering, 

Imperial College London, London, United Kingdom, 

a.chinchuluun@imperial.ac.uk, Ashwin Arulselvan, 

Stratos Pistikopoulos 

Support vector machines have been extensively studied and used in practice as a 

data mining tool especially in the field of biomedicine. We study the parametric 

support vector machine, where in we introduce the degree of misclassication as the 

parameter. Multiparametric prorgamming techniques are used for the resulting 

multiparametric mixed integer programs. 


444 

■ WD35 


Ranking and Internet Applications 



Chair: Mariana Olvera, Columbia University, 500 W. 120th Street, 

Mudd Building, Room 306, New York, NY, 10027, 

United States of America, molvera@ieor.columbia.edu 

1 - User Selection Process for More Profitable Display Advertising 

Ana Radovanovic, Research Scientist, Google Research, 

76 Ninth Ave., New York, NY, 10011, United States of America, 

anaradovanovic@google.com 

One of the key performance objectives in display advertising business is maximizing 

the proportion of user clicks out of all of the shown ads by a given advertiser. 

However, the click probability significantly depends on users we are showing a 

certain ad to. We introduce a self-organizing online policy for updating user 

members’ lists and show that this policy allows us to achieve nearly optimal longterm 

proportion of clicks out of all online inventory that is handled by a publisher. 

2 - Implicit Renewal Theory for Ranking Algorithms 

Mariana Olvera, Columbia University, 500 W. 120th Street, 

Mudd Building, Room 306, New York, NY, 10027, 

United States of America, molvera@ieor.columbia.edu, 

Predrag Jelenkovic 

We present a stochastic framework to analyze the qualitative large scale behavior of 

a family of ranking algorithms, in the same spirit of Google’s PageRank algorithm, 

via implicit renewal theory. Our analysis is based on a stochastic recursion 

constructed on a tree, and the techniques we develop can be applied to both linear 

and non-linear recursions. We extend prior work to allow a general correlation 

structure among the different inputs of the algorithm. 

■ WD36 


Quality Management II 


Chair: Scott Dellana, Associate Professor, East Carolina University, Bate 

3102, Dept of Marketing & Supply Chain Mgmt, Greenville, NC, 27858, 

United States of America, dellanas@ecu.edu 

1 - A Robust Technique to Process Qualitative Customer Data and 

Build Satisfaction Models 

Jose Luis Ribeiro, Dr., UFRGS, Av. Osvaldo Aranha 99, 

Porto Alegre, RS, 90035-190, Brazil, ribeiro@producao.ufrgs.br, 

Maria Auxiliado Tinoco 

Customer survey may be used to collect qualitative data and build satisfaction 

models. However, due to respondents’ lack of engagement, customer survey is 

usually contaminated with noise. Following robust regression principles, we 

developed a technique to process qualitative data according to its consistence, 

supporting the construction of robust satisfaction models. 

2 - Supply Chain Quality Management Practices and Perceptions: 

A Preliminary Empirical Study 

Scott Dellana, Associate Professor, East Carolina University, Bate 

3102, Dept of Marketing & Supply Chain Mgmt, Greenville, NC, 

27858, United States of America, dellanas@ecu.edu, John Kros 

Recent global supply quality problems have underscored the need for a better 

understanding of quality across the supply chain. Studies in supply chain quality are 

few, limited, and with mixed results warranting more extensive study. We examine 

quality management practice by industry class and supply chain organizational 

position and compare agreement between perceived supplier quality management 

practices reported by customers and quality practices reported by their supplier 

groups.

■ WD37 


Stochastic Processes 


Chair: Tianke Feng, Industrial and Systems Engineering, University of 

Florida, 303 Weil Hall, P.O. Box 116595, University of Florida, 

Gainesville, FL, 32611, United States of America, fengtk@ufl.edu 

1 - Sharp Bounds for the Distribution of Critical-path Length 

Munevver Mine Subasi, Assistant Professor, Florida Institute of 

Technology, 150 W. University Blvd, Melbourne, FL, 32901, United 

States of America, msubasi@fit.edu, Ersoy Subasi, Andras Prekopa 

In PERT we are frequently concerned with the problem to estimate the values of the 

probability distribution or expectation of the critical-path length. We develop a 

bounding technique to obtain sharp bounds for the values of the distribution of the 

critical-path length under moment information and the assumption that the random 

length of each arc follows a beta distribution. 

2 - Evolutionary Computation-based Statistical Estimation Models for 

Complex System Analysis Based on Count Data 

Emily Zechman, Assistant Professor, Texas A&M University, 3136 

TAMU, College Station, TX, 77843, United States of America, 

ezechman@tamu.edu, Seth Guikema, Royce Francis 

Count data is ubiquitous in engineering practice, and forecasting future counts of 

events based on past data is a critical problem. Examples include forecasting power 

outages and estimating the risk of traffic accidents. This research will explore 

flexible, data-adaptive regression models for count data that preserve insights into 

system behavior while allowing exploration of new combinations of model 

structures. A new evolutionary computation-based method will be developed and 

demonstrated. 

3 - The Replacement Problem in Manufacturing Stochastic Systems 

Eva Selene Hernàndez Gress, PhD, Universidad Autónoma del 

Estado de Hidalgo, Abasolo 600 Col. Centro, Pachuca, 42000, 

Mexico, evah@uaeh.edu.mx 

The Replacement Problem is modeled with Linear and Dynamic Programming. This 

problem can be modeled also as a finite, irreducible, homogeneous Markov Chain. 

The transition probabilities matrix and the optimal basis associated to the Linear 

Programming model are perturbed to find regions of feasibility and optimality. Some 

perturbations bounds of the transition probabilities are explored and a perturbation 

bound for the optimal basis is also proposed. 

4 - Markov Decision Processes with Uncertain Rewards 

Chin Hon Tan, PhD Student, Department of ISE, University of 

Florida, 303 Weil, University of Florida, Gainesville, FL, 32611, 

United States of America, chinhon@ufl.edu, Joseph Hartman 

Sequential decision problems can often be modeled as Markov decision processes. 

Classical solution approaches assume that the parameters of the model are known. 

However, the rewards are often estimated and uncertain in practice. In this paper, 

we look at the marginal change in the value function as a result of the error in the 

reward estimate and identify the region in which a policy remains optimal. We 

illustrate this work with a stochastic lot-sizing problem. 

5 - Sequential Stochastic Assignment Problem with 

Postponing Decisions 

Tianke Feng, Industrial and Systems Engineering, University of 

Florida, 303 Weil Hall, P.O. Box 116595, University of Florida, 

Gainesville, FL, 32611, United States of America, fengtk@ufl.edu, 

Joseph Hartman 

The sequential stochastic assignment problem has wide applications in 

abandonment problems and health care management, and has been well studied. It 

assumes that jobs arrive randomly with random values. Upon arrival, a job’s value is 

known and the decision maker immediately decides whether to accept or reject it. 

In this research, we study the value of postponing decisions by allowing a decision 

maker to hold jobs. We analyze the optimal threshold policies for this version of the 

problem. 


445 

■ WD38 


Algorithms for Real-Time Equilibrium and 

Relocation Problems 




Chair: Miguel F. Anjos, University of Waterloo, 200 University Avenue 

West, Waterloo, ON, Canada, anjos@stanfordalumni.org 

1 - Parametric Complementarity Problems for Online Optimal Control 

Mihai Anitescu, Computational Mathematician, Argonne National 

Laboratory, Math and Computer Science Division, 9700 S Cass Ave, 

Argonne, IL, 60439, United States of America, anitescu@mcs.anl.gov, 

Victor Zavala 

We demonstrate that if points along the solution manifold of a nonlinear model 

predictive control problem are consistently strongly regular, it is possible to track the 

manifold approximately by solving a single linear complementarity problem (LCP) 

at each control step. We derive a fast, augmented Lagrangean tracking algorithm 

and demonstrate the developments through a numerical case study. 

2 - Real-time Optimization and Differential Variational Inequalities 

Victor Zavala, Argonne National Laboratory, Math and Comp. 

Science Div., Bdg 240, 9700 S Cass Ave, Argonne, IL, 60439, 

United States of America, vzavala@mcs.anl.gov, Mihai Anitescu 

We present new insights into how to achieve higher frequencies in real-time 

optimization. The basic idea is that, instead of solving a full NLP problem at each 

sampling time, we solve a single, truncated QP problem. We prove that this 

corresponds to a time-stepping scheme applied to a differential variational 

inequality. We propose a fast and stable scheme using augmented Lagrangian 

regularization and projected Gauss-Seidel to track the solution manifold. 

3 - A Stochastic Optimization Model for Ambulance Relocation 

Joe Naoum-Sawaya, University of Waterloo, 200 University Av W, 

Waterloo, Canada, jnaoumsa@engmail.uwaterloo.ca, Samir Elhedhli 

In this talk, we present our recent work on implementing a real time ambulance 

relocation system for the Region of Waterloo Emergency Medical Services. We 

formulate a stochastic optimization model and devise a heuristic that finds good 

solutions quickly. Results using the Region of Waterloo EMS data are presented. 

■ WD39 


Algorithms and Applications for Integer Programming 

Sponsor: Optimization/Integer Programming 


Chair: Yongpei Guan, University of Florida, 303 Weil Hall, Gainesville, FL 

United States of America, guan@ise.ufl.edu 

1 - Fast Algorithms for Special Cases of the Minimum Cost 

Flow Problem 

Bala Vaidyanathan, Operations Research Advisor, FedEx Express, 

3680 Hacks Cross Road, Memphis, TN, 38125, United States of 

America, bala.vaidyanathan@gmail.com 

We consider special cases of the minimum cost flow problem when the nodes lie on 

a line, a circle, or a tree. We show that the properties of these problems can be 

exploited to develop specialized algorithms that run significantly faster than 

minimum cost flow algorithms for general networks. These problems find 

applications in areas such as transportation, electric power transmission, production 

planning, telecommunications, computational biology, and computational music. 

2 - Polyhedral Results for MIPs with Cardinality Constraints 

Ilksen Icyuz, University of Florida, 303 Weil Hall, Gainesville, 32611, 

United States of America, eceicyuz@ufl.edu, Jean-Philippe Richard 

We consider several mixed integer programs arising from applications that contain 

various forms of cardinality constraints. For these models, we review existing 

polyhedral results and obtain new strong valid inequalities based on disjunctive and 

lifting techniques. We also characterize situations in which these inequalities are 

sufficient to completely define the associated convex hulls. Finally, we present the 

results of a computational study aimed at testing the strength of these new cuts.

WD40 

3 - Strong Formulations for Lot-Sizing Problems with Disruptions 

Zhili Zhou, University of Florida, 303 Weil Hall, Gainesville, FL, 

32611, United States of America, zlzhou@ufl.edu, Yongpei Guan 

Most previous research on lot-sizing problems assumed no disruptions. However, 

disruptions occur in practice, which lead to extra outsourcing or production costs, as 

compared to the normally scheduled case. We formulate this problem as a robust 

integer program, and derive facet-defining inequalities for the problem. Final 

computational experiments show the effectiveness of our proposed approach. 

4 - New Cutting Planes for Cardinality Optimization 

Ismael De Farias, Texas Tech University, Lubbock, TX, United States 

of America, ismael.de-farias@ttu.edu, Ming Zhao, Rajat Gupta, 

Ernee Kozyreff 

We study the cardinality optimization polytope and extensions, particularly the set 

of the minimum 0-norm optimization problem. We present new cutting planes for 

them and computational results on their use in branch-and-cut. 

■ WD40 


Robust Optimization 


Chair: Bacel Maddah, Assistant Professor, American University of Beirut, 

P.O. Box 11-0236 Riad El Solh, Beirut, 1107-2020, Lebanon, 

bm05@aub.edu.lb 

1 - Handling Uncertainty in Supplier Selection using 

Robust Optimization 

Sheela Siddappa, Technical Architect, Infosys Technologies Limited, 

Electronic City, Bangalore, India, sheela_siddappa@infosys.com, 

Paresh Kumar Marwaha 

Selecting the right set of suppliers to procure items can help reduce company’s 

expense and improve customer satisfaction. This research tries to account for 

uncertainty in supplier’s performance, capacity, and cost while selecting the 

suppliers. We develop a two step process 1) select suppliers based on the uncertainty 

in their performance 2) estimate procurement quantity from each supplier based on 

the uncertainty in supplier’s capacity, cost, etc. A Robust Optimization methodology 

is adopted. 

2 - The Relational Algebra of Constraint Sets in Robust Optimization 

G. N. Srinivasa Prasanna, International Institute of Information 

Technology - Bangalore, 26/C Electronics City, Bangalore, India, 

gnsprasanna@iiitb.ac.in 

A relational-algebra for identifying, classifying, & visualizing relationships among 

different constraint sets (input assumptions) in a robust optimization (RO) 

framework is presented. Set-theoretic relations between alternative constraint sets 

are specified in an expression-based query language (composed of set-disjointness, 

union, intersection, etc.) We can compare, visualize & analyze different sets of 

assumptions, facilitating decision-support in general, & RO in particular. 

3 - Loading of Ships at Refineries when Arrival Times are Uncertain - 

A Robust Optimization Approach 

Jens Bengtsson, Norwegian School of Economics and Business 

Administration, Helleveien 30, Bergen, Norway, 

jens.bengtsson@nhh.no, Patrik Flisberg, Mikael Rönnqvist 

Ships that arrive to refinery ports are either served from product or component 

tanks. Arrival times of ships are uncertain and as such the inventory levels in the 

tanks will be uncertain and can in extreme cases be full or empty. Both these 

outcomes may be costly to the refiner. In this paper we take the arrival time 

uncertainty into account and analyze the production planning and loading problem 

by using robust optimization. 

4 - The Newsvendor Problem with Ambiguous Demand Distribution 

Bacel Maddah, Assistant Professor, American University of Beirut, 

P.O. Box 11-0236 Riad El Solh, Beirut, 1107-2020, Lebanon, 

bm05@aub.edu.lb, Ebru Bish, F. Jordan Srour, Kyle Lin 

We study the newsvendor problem when the demand distribution is “ambiguous” in 

terms of being within a given Kullback-Leibler distance from a nominal distribution. 

Our approach is novel as no precise estimate of demand parameters (e.g., moments, 

percentiles, and range) is required. We derive the optimal newsvendor ordering 

policy in this setting. 


446 

■ WD41 


Metaheuristics II 


Chair: Fan Wang, Prof., Sun Yat-sen Business School, Sun Yat-sen 


fanwang@gmail.com 

1 - A Modified Particle Swarm Optimization Algorithm 

Junhyuk Park, Pohang University of Science and Technology, 

Nam-Gu, Hyoja-Dong, POSTECH Eng 4-207, Pohang, Korea, 

Republic of, sacarlee@postech.ac.kr, Byung-In Kim, Jongsung Lee 

In classical PSO algorithms, the global best solution has a strong influence on the 

entire population. As a result, particles may rapidly converge to a premature global 

best solution without exploring enough search space. In this presentation, we 

extend the searching capability of particles by limiting the influence of global best 

solution. We compare the performance of the algorithm with several existing 

algorithms. 

2 - A Genetic Algorithm for the Team Formation Problem 

Paul Rubin, Michigan State University, The Eli Broad Graduate 

School of Managem, Michigan State University, East Lansing, MI, 

48824, United States of America, rubin@msu.edu, Lihui Bai 

We consider the team formation problem that assigns individuals with various 

values of attributes to teams so that the differences across teams with respect to 

each attribute are minimal. The problem is formulated as a mixed integer linear 

program and solved by a genetic algorithm with a constraint programming solver. 

Numerical results on randomly generated as well as real problems are reported. 

3 - Exploration of a Functional Programming Approach for 

Developing a Metaheuristic 

Dennis Drinka, Exploration of a Functional Programming Approach 

for Developing a Metaheuristic, University of Alaska Anchorage, 

3211 Providence Drive, Anchorage, AK, 99508, 

United States of America, afded@uaa.alaska.edu 

This presentation explores the use of a functional programming approach to drive a 

metaheuristic for solving a combinatorial optimization problem. In particular, it 

explores the use of LINQ to Objects to implement a tabu search procedure. It will 

demonstrate the use of this approach on the All Units Quantity Discount problem 

and describe how functional programs can easily be embedded within existing 

imperative programming solution approaches. 

4 - Self Adaptation GA Applied to Routing Problems 

Jaime Mora Vargas, Tecnológico de Monterrey, Research and 

Graduate Division, Campus Estado de México, Mexico, 

jmora@itesm.mx], Nestor Velasco-Bermeo, Miguel Gonzàlez- 

Mendoza 

Mathematical Programming (MP) has proven to be an effective strategy to tackle 

down optimization problems. Traditional methods fail to solve such problems due to 

the “combinatory explosion”, an exponential increase of the solution space 

theorically identified as “NP-Hard” problems, based on such tendency new strategies 

have been proposed; the combination of mathematical programming and hybrid 

metaheuristics, the later having as main characteristic the adaptation ability to the 

environment’s changes without the need to fine tune the parameters or re-run the 

method. GA’s have proven to be one of the most successful methods to solve most 

of the classic optimization problems. Though the benefits of such algorithms 

disappear when situation analyzed changes. Based on such inconvenient a new 

approach is proposed. A hybrid GA with MP and the incorporation of “autoadaptation” 

lets the algorithm not just to evolve solution candidates but its own 

parameters based on the current state (individual’s aptitude according to each 

generation). The problems solved are ; Vehicle Routing Problem With Time Window 

(VRPTW) and Multi-depot Vehicle Routing Problem With Time Window 

(MDVRPTW). 

5 - A Compromised Large-scale Neighborhood Search Heuristic for 

Heterogeneous VRP 

Fan Wang, Prof., Sun Yat-sen Business School, Sun Yat-sen 


fanwang@gmail.com, Yi Tao 

We addressed the Heterogeneous VRP with fixed costs and routing costs in which a 

limited fleet of different vehicles is available for sending goods to customers with 

known demand along the network with the objective of minimizing the total cost. 

We propose a compromised large-scale neighborhood search heuristic where the 

neighborhood aims to relax the subset-disjoint restriction during the process of 

multi-exchange neighborhood search.

■ WD42 


Optimization Algorithms 




Chair: Ilya Safro, Argonne National Laboratory, 9700 S. Cass Avenue, 

Argonne, IL, 60439, United States of America, safro@mcs.anl.gov 

1 - Beyond Fisher’s Linear Dicriminant Function 

Shuichi Shinmura, Professor, Seikei University, 3-3-1 Kichijoujikitamachi, 

Musashino, 180-8633, Japan, shinmura@econ.seikei.ac.jp 

Fisher founded linear discriminant function (LDF). LDF assume two groups are the 

same normal distribution, nevertheless few data satisfy this assumption. If data 

satisfy this, error rate equals to the minimum error rate (MNM). Therefore, LDF 

should be defined by MNM criterion. Revised IPLP-OLDF looks for the estimate of 

MNM using mixture of LP and IP. Evaluation of 13,500 models shows the mean of 

error rates of Revised IPLP-OLDF using LINGO are less than those of LDF & logistic 

regression. 

2 - A Separation Heuristic for Gap Inequalities 

Konstantinos Kaparis, Dr., University of Southampton, 

Southampton, SO17 1BJ, United Kingdom, K.Kaparis@soton.ac.uk, 

Laura Galli, Adam Letchford 

Laurent and Poljak introduced a class of valid inequalities for the max-cut problem, 

called gap inequalities, which include many other known inequalities as special 

cases. Even though they have received limited attention they can make very good 

cutting planes for the Max-Cut problem. We describe a separation heuristic for these 

and we present computational results which illustrate the potentials of the proposed 

scheme. 

3 - Polynomial Complexity Algorithm for Integer 

Transportation Problem 

Vladimir Tsurkov, Prof., Computing Center of Russian Academy of 

Sciences, Vavilov Str., 40, Moscow, 119333, Russia, tsur@ccas.ru, 

Alexander Tizik 

A net optimization large scale problem from telecommunication area is solved by 

means of the known method of column generation. The intermediate problem has 

transportation kind of restrictions. Algorithm of polynomial complexity is proposed. 

Method is based on iterative solutions of two-dimensional knapsack problems. 

4 - Multiscale Approach for the Network Compression-friendly Ordering 

Ilya Safro, Argonne National Laboratory, 9700 S. Cass Avenue, 

Argonne, IL, 60439, United States of America, safro@mcs.anl.gov 

We present a fast multiscale approach for the network minimum logarithmic 

arrangement problem. This type of arrangement plays an important role in a 

network compression. The algorithm is of linear complexity and exhibits good 

scalability which makes it practical and attractive for using on large-scale instances. 

Its effectiveness is demonstrated on a large set of real-life networks. 

■ WD44 


Network Modeling in Healthcare 



Chair: Eva Enns, Stanford University, 117 Encina Commons, Stanford, 

CA, 94035, United States of America, evaenns@stanford.edu 

1 - Modeling HIV Spread in Africa using Sexual Contact Network Data 

Benjamin Armbruster, Northwestern University, 2145 Sheridan Road 

(Tech Bldg), Evanston, IL, 60208-3119, United States of America, 

armbruster@northwestern.edu, Stephane Helleringer 

Using a unique data set from the Likoma Network Study, we examine the structure 

of the sexual contact network, how it changes over time, and how this affects 

models of disease spread. 

2 - The Impact of the Health Delivery Supply Chain on Adherence to 

HIV Treatment 

Jessica McCoy, Stanford University, Stanford, California 94305, 

United States of America,jhmccoy@stanford.edu, Eric Johnson 

Adherence to HIV treatment is a predictor of viral suppression, and transmission 

rates are much lower in adherent populations. The health delivery supply chain in a 

resource-limited region potentially creates or removes barriers for patients seeking 

regular treatment. We develop a model to gain insights about the impact that the 

supply chain has on patients’ ability to be adherent (and hence on HIV prevalence) 

in a community. 


447 

3 - Controlling Epidemics over Multi-level Co-evolving Networks 

Achla Marathe, Asociate Professor, Virginia Tech, CRC XV VBI, 

Blacksburg, 24061, United States of America, amarathe@vbi.vt.edu, 

Stephen Eubank, Bryan Lewis, Jiangzhou Chen, Madhav Marathe 

Human behavior, epidemics and social contact networks are closely intertwined and 

co-evolve. Effective planning and response strategies must take these complicated 

interactions into account. We will describe an interactionist approach for studying 

these co-evolving social systems with the goal of supporting public health 

epidemiology. Individual and collective behavioral adaptation is critical in these 

systems and will be highlighted via illustrative case studies. 

4 - Optimal Link Removal for Epidemic Control Over Networks 


94035, United States of America, evaenns@stanford.edu, Jeff 

Mounzer, Margaret Brandeau 

Control of infectious diseases which spread through close contact often focuses on 

interrupting the network of contacts. We examine the problem of determining 

which links should be removed from a contact network to maximize infections 

averted, given a constraint on the maximum number of links that can be removed. 

We formulate the problem as a non-convex quadratically constrained quadratic 

program. We evaluate the performance of approximate and heuristic solutions. 

■ WD45 


Health Care, Other 


Chair: S.Reza Sajjadi, Post- Doctoral Research Fellow, North Dakota State 

University, NDSU Department2485, 120A CIE, P.O.Box 6050, Fargo, ND, 

58108-6050, United States of America, reza.sajjadi@ndsu.edu 

1 - Feature Level Selection in Disease Clusters 

Saylisse Davila, Arizona State University, 151 E Broadway Rd #210, 

Tempe, AZ, United States of America, saylisse@asu.edu, 

George Runger, Eugene Tuv 

Most public health surveillance methods focus on detecting the presence of disease 

clusters. However, identifying the specific location of disease clusters can be equally 

important. Nowadays, health data is often recorded in large databases. Finding the 

specific location of these disease clusters among a large number of records and 

variables can be an intricate task. We will present a methodology that can be used 

to identify the ranges and/or levels of variables that define the clusters. 

2 - Impact of Medicare Part D on Generic Drug Utilization in Long Term 

Care Facilities 

Changmi Jung, Carnegie Mellon University, 4800 Forbes Ave., 

Rm 242, Pittsburgh, PA, 15213, United States of America, 

changmi@andrew.cmu.edu, Rema Padman, Shamena Anwar 

This study examines the impact of Medicare Part D program on generic drug 

prescription rates in Long Term Care Facilities. We analyze prescription orders from 

a regional online pharmacy to induce a general pattern using the Difference in 

Difference method on four different therapeutic drug classes. 

3 - A Comparison of Drug Price Increases in the 

Pharmaceutical Industry 

Kathleen Martino, Rutgers University, 1 Washington Park, Newark, 

NJ, 07102, United States of America, martinok2@gmail.com, 

Yao Zhao 

Motivated by the current events regarding prescription drug prices, this empirical 

study examines if factors such as therapeutic class, manufacturer, and active 

ingredient levels play a role in the established price of prescription drugs. We 

examine both the wholesale price set by manufacturers and the markup set by 

distributors for both brand and generic drugs. We also quantify the difference 

between brand and generic drug prices as a drug moves downstream from 

manufacturers to pharmacies. 

4 - Optimization of the Dynamic Operational Decisions for Ambulance 

Dispatch: A Reallocation Model 

Sandra Milena Santa, Universidad de los Andes, Carrera 1 N° 18A - 

12, Bogotà, CU, Colombia, sm.santa30@uniandes.edu.co, Andres 

Correa, Ciro Alberto Amaya Guio, Nubia Milena Velasco Rodriguez 

Ambulance dispatch as a response to some emergency medical services (EMS) needs 

to be executed effectively meeting time response and coverage demand 

requirements. Reallocation of ambulances gives the possibility of meeting and 

improving these standards response, but it is important to have in mind that these 

decisions carry transport costs and consume time. Therefore, we present a dynamic 

model as a decision support tool that allows evaluating the necessity to update 

location of ambulances.

WD46 

5 - Patient Flow in a Multiple-Route High-Variability Primary Care Clinic 

S.Reza Sajjadi, Post- Doctoral Research Fellow, North Dakota State 

University, NDSU Department2485, 120A CIE, P.O.Box 6050, Fargo, 

ND, 58108-6050, United States of America, reza.sajjadi@ndsu.edu, 

Jing Shi, Kambiz Farahmand 

Simulation modeling of patient flow in an outpatient primary care clinic is 

considered in this study. In schedule-based clinics where patient arrivals follow the 

schedule given to the patients, inefficient layout and communication system as well 

as variability increase the waiting time of the patients in the clinics and thus result 

in patient dissatisfaction. Considering waiting time and distance traveled, the 

simulation model investigates a number of scenarios to improve the current 

situation. 

■ WD46 


Statistical Modeling and Analysis in Healthcare 

Applications 



Chair: Li Zeng, Assistant Professor, University of Texas at Arlington, 

Industrial & Manuf. Sys. Engr. Department, Arlington, Tx, 76019-0017, 

United States of America, lzeng@uta.edu 

1 - Minimizing Rehospitalization Costs Through Machine Learning 

Mohsen Bayati, Stanford University, 350 Serra Mall, Packard 

Building, room 278, Stanford, CA, 94305, United States of America, 

bayati@stanford.edu, Mark Braverman, Eric Horvitz 

Nearly one in every five patients is rehospitalized within 30 days of their discharge. 

In 2004 the estimated cost of unplanned readmissions to Medicare was $17.4 

billion. In this talk, I will demonstrate how online machine learning and 

optimization tools can be applied to electronic health records to identify patients 

with the highest risk of rehospitalization. We validated these predictions on a major 

hospital’s database and obtained cost-effective policies for minimizing 

rehospitalizations. 

2 - In-N-Out: An Energy Balance Model of Obesity 

Joanna Lankester, PhD candidate, Stanford University, 300 Pasteur 

Lane, Grant S131, Stanford, CA, 94305, United States of America, 

jl3@stanford.edu, Margaret Brandeau, Julie Parsonnet 

Obesity, which is associated with heart disease, diabetes, and other common and 

sometimes life-threatening conditions, has increased in prevalence dramatically over 

the last 50 years. To understand the rise in obesity, we are developing a model of 

body weight in the U.S. population from a perspective of energy balance. The model 

quantifies the changes over time in factors that influence body weight so that we 

can explore variability in obesity between individuals with similar energy intake. 

3 - Reducing Medication Errors in Pediatrics 

Michelle McGaha, PhD Graduate Student, Texas A&M University, 

3131 TAMU, College Station, TX, 77843, United States of America, 

michelle.mcgaha@neo.tamu.edu, Kiavash Kianfar, Lewis Ntaimo, 

Amarnath Banerjee 

Medication errors are common in the complex setting of pediatric medicine. We 

discuss the obstacles in determining the prevalence of pediatric medication errors 

and present a taxonomy chart of the most common pediatric medication errors, as 

well as defining characteristics for each error type. We discuss the challenges and 

highlight prevention strategies and meaningful use of health IT in reducing 

medication errors in the future. 


448 

■ WD47 


Project Management 


Chair: Eduardo G. Hernandez-Martinez, PhD, Tecnologico de Estudios 

Superiores de Coacalco, 16 de Septiembre 54, Coacalco, 55700, Mexico, 

eghm2@yahoo.com.mx 

1 - Analyzing Robustness of Team Communication-performance 

Relationships 

Deanna Kennedy, Assistant Professor, University of Washington, 

Bothell, 1501 Copperfield Pkwy #412, College Station, TX, 77845, 

United States of America, kennedy.deanna@gmail.com, 

Rebecca Perryman, Sara McComb 

Patterns in team communication relate to time and cost performance. We examine 

optimality when tradeoffs are made between the two performance outcomes. 

Further, we analyze the robustness of communication patterns by estimating the 

performance of sampled data from hyperspheres centered at optimal points. 

Implications are discussed. 

2 - Ensuring OR/MS Project Success: A Change 

Management Perspective 

Robert Levasseur, Professor of Management, Walden University, 

2614 Vista Cove Road, St. Augustine, FL, 32084-3069, 

United States of America, robert.levasseur@waldenu.edu 

Research into the reasons for project failure suggests that a high percentage of those 

failures are the result of non-technical problems, such as poor communication and 

resistance to change. The purpose of this presentation is to explore ways in which 

OR/MS project leaders can increase the odds of project success by applying basic 

principles and practices of change management. 

3 - Theory of Latency in integrated Concurrent Engineering 



States of America, john.m.chachere@nasa.gov, John Kunz, 

Ray Levitt 

We observed Stanford and NASA teams using Integrated Concurrent Engineering 

(ICE) to accelerate conceptual design by a factor of ten. We assert that ICE teams 

manage ten enabling factors that decimate information response latency, and thus 

project duration. Latency is a unifying principle and a practical metric that can 

describe, evaluate and manage engineering design collaboration. ICE is the “Just in 

Time” of knowledge work; ICE flows information with minute latency and high 

reliability. 

4 - New Method for Project Crashing using Excel Solver 

Pamela Zelbst, Sam Houston State University, 1256 Avenue I, 

Huntsville, TX, 77341, United States of America, mgt_pjz@shsu.edu, 

Kunpeng Li, Bin Shao 

We developed a new method for project crashing using Excel Solver. The method 

uses AON network. The Excel formulation is much simpler and straightforward 

than the traditional method using AOA network. 

5 - A Discrete-event Approach for the Task Planning of 

PMBOK-based Projects 

Eduardo G. Hernandez-Martinez, PhD, Tecnologico de Estudios 

Superiores de Coacalco, 16 de Septiembre 54, Coacalco, 55700, 

Mexico, eghm2@yahoo.com.mx, Guillermo Torres 

This work presents an approach based on Supervisory Control Theory for the task 

planning of projects based on the PMBOK. The goal is a formal strategy to obtain 

automatically a gant chart considering all possible tasks concurrences according to 

times, resources availability and resources sharing. Unlike the graph methods 

reported on the PMBOK based on the user experience, the approach permits the 

time optimization establishing only the resources list and the restrictions of the tasks 

execution.

■ WD50 


Information Systems III 


Chair: Yun Huang, Northwestern University, 4146 Oakton st #2, Skokie, 

IL, 60076, United States of America, yun@northwestern.edu 

1 - The Value of Information: Evidence of Decreasing Price Dispersion 

in India’s Crop Market 

Chris Parker, PhD Candidate, London Business School, 

Regent’s Park, London, NW14SA, United Kingdom, 

cparker.PhD2007@london.edu, Nicos Savva, Kamalini Ramdas 

This paper aims to measure the effect of a new text message based information 

service on price dispersion empirically. Our dataset contains information about 

several crops, with diverse supply chain characteristics, and information about the 

adoption level of the technology. Having information about multiple crops allows us 

to identify factors impacting different types of crops. Results show that adoption 

levels are associated with lower price dispersion between markets. 

2 - Value From New Electronic Markets: A Diffusion Analysis of Two 

Equity Options Exchanges 

Chris Parker, PhD Candidate, London Business School, 

Regent’s Park, London, NW14SA, United Kingdom, 

cparker.PhD2007@london.edu, Bruce Weber 

Two all-electronic options exchanges opened for trading in 2000 and 2004. The 

exchanges gained trading volumes in competition with four incumbent markets. To 

explain the markets’ diffusion patterns, we model broker order routing decisions. 

The model generates hypotheses, which we test using a panel of quarterly 

disclosures from major brokerage firms. We conclude that firm heterogeneity is 

more influential than network effects in explaining the diffusions of the new 

markets at the broker level. 

3 - An Online Retailer’s Incentive to Become an Online Marketplace 

Kihoon Kim, Assistant Professor, Korea University Business School, 

Anam-Dong, Seongbuk-Gu, Seoul, 136701, Korea, Republic of, 

kihoonk@gmail.com 

When an online retailer develops into an online marketplace, the online retailer can 

invite other online retailers to sell the same product it deals with in the online 

marketplace. This paper investigates under what conditions the online retailer is 

advised to transform itself to the online marketplace, considering the online 

retailer’s trade-off of its referral revenue against the loss of some of its loyal 

customers. We also characterize optimal pricing structures of the online 

marketplace. 

4 - IOS Appropriation and Net Enablement 

Mei Cao, University of Wisconsin-Superior, Department of Business 

and Economics, Superior, United States of America, 

mcao1@uwsuper.edu, Qingyu Zhang 

The objective of the study is to uncover the nature of IOS appropriation and net 

enablement as an antecedent of supply chain collaboration. Reliable and valid 

instruments were developed through rigorous empirical analysis including 

structured interviews, Q-sort, and a large-scale study. Data were collected through a 

Web survey of U.S. manufacturing firms in various industries. Predictive validity is 

evaluated by demonstrating its strong relationship with supply chain collaboration. 

5 - Encounter in Virtual Space 

Yun Huang, Northwestern University, 4146 Oakton st #2, Skokie, IL, 

60076, United States of America, yun@northwestern.edu, 

Roger Chen, Noshir Contractor, Hani Mahmass 

Many virtual worlds such as massively multiplayer online role-playing games 

(MMORPGs) are often designed to mimic the real world along several dimensions, 

including the spatial requirements for engaging in activities. This paper examines 

the impact of virtual “geographical distance” between players in EverQuest II on the 

process of relation building in game. The results show that players are more likely to 

meet and interact with each other if they have more overlapping activity spaces. 


449 

■ WD53 


Operations/Finance Interface 


Chair: Amit Mitra, Auburn University, College of Business, Office of the 

Dean & Department of Manag, Auburn, AL, 36849-5240, 

United States of America, mitraam@auburn.edu 

1 - Vicarious Learning From Operational Failures 

Manpreet Hora, Georgia Institute of Technology, 800 W Peachtree St. 

NW, Atlanta, GA, 30308, United States of America, 

manpreet.hora@mgt.gatech.edu, Robert D. Klassen 

The occurrence of a rare operational failure in a firm (incident firm) provides an 

opportunity for vicarious learning for other firms (knowledge-seekers). Employing a 

field experiment, we examine the firm-level characteristics that enable vicarious 

learning. Results suggest that saliency of the incident firm and complementarities 

between the incident firm and the knowledge-seekers have a positive association 

with vicarious learning. 

2 - Investigating the Impact of Operational Variables on Manufacturing 

Cost by Simulation Optimization 

Wen-Chyuan Chiang, Prof., Collins College of Business, The 

University of Tulsa, The University of Tulsa, Tulsa, OK 74104, Tulsa, 

United States of America, wen-chyuan-chiang@utulsa.edu, 

Rui Zhang 

We focus on the relationship between operational variables (setup time, scrap rate, 

downtime rate) and the manufacturing cost. Keeping such parameters low is 

beneficial for reducing the average variable cost, but the required maintenance will 

incur an additional fixed cost. Therefore, a simulation based optimization approach 

is applied to determine the optimal level of the operational variables for minimizing 

the average cost. Sensitivity analysis and explanations are also provided. 

3 - Managing Operations Risks: An Options Perspective 

Wei Chen, Katz Graduate School of Business, University of 

Pittsburgh, 233 Mervis Hall, Pittsburgh, PA, 15260, 

United States of America, wchen@katz.pitt.edu, Jennifer Shang 

When dealing with demand and supply uncertainties, operations managers must 

also consider risks such as exchange rate variation, distribution cost surge, and 

political upheaval. These uncertainties can significantly influence a firm’s 

profitability. We address these risks from the perspectives of financial markets, 

operations control, and stochastic programming, and show that real options is a 

useful framework for understanding risk hedging in operations. 

4 - An Integrated Model Utilizing R&D Costs for Two-dimensional 

Warranty Policies 

Jay Patankar, Professor, The University of Akron, Department of 

Management, College of Bus. Administration, Akron, OH, 44325- 

4801, United States of America, jgp@uakron.edu, Amit Mitra 

Some consumer durables, such as automobiles, involve warranties involving two 

attributes. These are time elapsed since sale of the product and usage of the product 

at a given point in time. An avenue to impact warrant costs is through research on 

product development. The objective then becomes to determine warranty 

parameters, while constraining the sum of the expected unit warranty costs and 

research and development costs per unit sales, under a limited research and 

development budget. 

■ WD54 


Technology Management 



Chair: Gulru Ozkan, Assistant Professor, Clemson University, Department 

of Management, 101 Sirrine Hall, Clemson, SC, 29634, United States of 

America, gulruo@clemson.edu 

1 - Better Selection or Efficient Contracting?: A Model of Knowledge 

Vendor Selection and Contracting 

Zhijian Cui, PhD Candidate, INSEAD, Constance de Blvd, 

Fontainebleau, France, Zhijian.CUI@insead.edu, Sameer Hasija 

We are studying the vendor selection and contracting issue for knowledge driven 

processes. In particular, we study three vendor selection process: (1)Selection is 

based on initial talks with the vendor and subsequently the contract is negotiated 

with the selected vendor. (2)The vendors are informed about the SLA and asked to 

bid for the contract. (3) The vendors are asked to propose a contract to the client.

WD55 

2 - Innovation in Top-Down and Bottom-Up Strategy Processes 

Fabian Sting, INSEAD, Boulevard de Constance, Fontainebleau, 

France, fabian.sting@insead.edu, Christoph Loch 

We study strategy processes at six German manufacturing organizations using an 

organizational search perspective. While the final decision on strategic initiatives 

remains at the top, strategic initiatives are distributed across hierarchical levels, 

depending on where expertise is concentrated. The organizations also use multiple 

mechanisms to coordinate decentralized actors. Coordination and top-down decision 

making is weighed against the creativity that stems from delegated search. 

3 - Managing New Product Development Knowledge for Competing 

Firms: Case of Joint Development 

Gulru Ozkan, Assistant Professor, Clemson University, 

Department of Management, 101 Sirrine Hall, Clemson, SC, 29634, 

United States of America, gulruo@clemson.edu, Cheryl Gaimon 

We introduce a stochastic game on knowledge sharing (KS) and knowledge 

development (KD) strategies for two NPD firms. First, leader sets allocations of 

profit, then firms decide on KS for joint development of a new product. Next, firms 

jointly pursue KD and launch the product. Insights include impact of uncertainty. 

4 - A Decision Model to Manage Network Security Technologies for 

Information Assurance 

Soumyo Moitra, Senior Member of Technical Staff, Software 

Engineering Institute, Carnegie Mellon University, 

4500 Forbes Ave, Pittsburgh, PA, 15213, United States of America, 

smoitra@sei.cmu.edu 

This paper describes a model to manage technologies to protect informational assets. 

The focus is on the valuation of information and a methodology to arrive at the 

value at risk is presented. This assessment is used by the model to evaluate the 

benefits of different levels of security technologies. Sensitivity analysis with respect 

to the value of information is presented. 

■ WD55 


Inventions, Innovation and Technology Management 



Chair: Kun Liu, Assistant Professor, Wayne State University, 

5201 Cass Ave, Detroit, United States of America, ek9525@wayne.edu 

1 - Abandonment of Patented Inventions in 

Innovation-Intensive Industries 

Kun Liu, Assistant Professor, Wayne State University, 5201 Cass Ave, 

Detroit, United States of America, ek9525@wayne.edu 

Little research has examined how firms abandoned some inventions in innovationintensive 

industries. Distant innovation search is associated with a greater 

technological distance between newly acquired inventions and abandoned 

inventions, as well as a younger average age of abandoned inventions. Abandoning 

younger inventions and more focused abandonment of inventions are associated 

with greater market value as measured by Tobin’s Q. 

2 - Continuous Quality/Time/Cost Tradeoffs 

Bruce Pollack-Johnson, Associate Professor of Mathematical 

Sciences, Villanova University, 800 Lancaster Avenue, Villanova, PA, 

19085, United States of America, bruce.pollackjohnson@villanova.edu, 

Matthew Liberatore 

We use an example of a translation project to illustrate how quality can be modeled 

as a continuous function of time and cost. Given a project deadline and budget, 

overall quality can then be maximized using a nonlinear programming model. 

Alternatively, iso-quality curves can be drawn, to visualize the continuous tradeoffs 

between time, cost, and quality, and then used to pick a combination that seems 

best for a particular situation. 

3 - Launching Technologically Advanced Products in 

Segmented Markets 

John N Angelis, Rochester Institute of Technology, 

105 Lomb Memorial Drive, Rochester, United States of America, 

jangelis@saunders.rit.edu 

We focus on how competing profit-maximizing firms should set price and quality 

for a new technologically-advanced product sold to a segmented market. We 

analyze a closed-loop Stackelberg game with perfect information. If the late entrant 

possesses a large enough cost disadvantage, it should only target the least innovative 

segment. We also find that a firm with a large cost advantage may not necessarily 

earn higher profits by being the first mover. 


450 

4 - Realizing the Value of RFID in a Global Enterprise 

Ann Marucheck, Professor, Kenan-Flagler Business School, McColl 

Bldg CB 3490, UNC-Chapel Hill, Chapel Hill, NC, 27599, United 

States of America, ann_marucheck@unc.edu, Noel Greis, 

Monica Nogueira, Anders Duus, Hong Tham Nguyen 

For many organizations, understanding the potential benefits of RFID and justifying 

its investment continue to be challenges. In this research, we comprehensively 

study over 4000 cases of RFID projects as reported in the IDTechEX database. 

Specifically, we contrast the benefits realized by early adopters of RFID with more 

recent adopters. We further suggest measures that can determine how RFID 

provides economic value to an enterprise and identify the drivers of determining 

that value. 

■ WD56 


Simulation and Optimization II 


Chair: Xueping Li, Assistant Professor, University of Tennessee, 

408 East Stadium Hall, Knoxville, TN, 37996, United States of America, 

Xueping.Li@utk.edu 

1 - Allocating Manpower to Minimize Lmax in a Job Shop 

Benjamin Lobo, North Carolina State University, 400 Daniels Hall, 

College of Engineering, Raleigh, NC, 27695, United States of 

America, bjlobo@gmail.com, James Wilson, Thom Hodgson, 

Russell King, Kristin Thoney 

Most job shops are constrained not only by machines, but also by the number of 

workers available to operate these machines. Different worker allocations to 

machine groups can impact the Lmax value of a schedule. Using a relaxation of the 

problem to generate a lower bound on Lmax, we develop a procedure to allocate 

workers to machines that minimizes this lower bound. Computational experience is 

presented. 

2 - Improved Fully Sequential Procedure with Mean and 

Variance Update 

Huizhu Wang, Georgia Institute of Technology, 765 Ferst Dr, Atlanta, 

GA, 30332-0205, United States of America, huizhuwang@gmail.com, 

Seong-Hee Kim 

Ranking and selection (R&S) procedures compare a number of simulated systems 

and try to find a system with the best performance measure. Fully-sequential R&S 

procedures are shown to be efficient but its probability of correct selection tends to 

be higher than the nominal level especially for a large number of systems. We study 

sources for conservativeness and present a procedure with improved efficiency. 

3 - Simulation-based Optimization and Stochastic Programming 

Tahir Ekin, PhD Candidate, George Washington University, 

2201 G Street, NW Funger 415, Washington, DC, 20052, United 

States of America, ekin@gwu.edu, Refik Soyer, Nicholas Polson 

We provide a simulation-based approach to optimization and stochastic 

programming. First, we consider a Monte Carlo solution to linear programming and 

then show that it naturally extends to two-stage stochastic programming with 

uncertainty. One advantage of our approach is that it is straightforward to add 

parameter uncertainty and does not need derivative information. 

4 - Parametric Moment Closure of Non-linear State Dependent 

Stochastic Systems 

Ritesh Arora, Graduate Research Assistant, Missouri University of 

Science and Technology, Dpt. of Engg. Management & Systems 

Engg., Rolla, MO, 65409, United States of America, 

ra95d@mail.mst.edu, Ivan Guardiola 

This research highlights the use of parametric moment closure approximations for 

non-linear stochastic systems. A general birth-and-death non-linear stochastic 

process will be analyzed and closed under various underlying parametric 

distribution assumptions. It results that closure under neglect is considerably less 

robust than parametric closure. We highlight this theoretical endeavor through an 

analysis of aphid population stochastic model. 

5 - Simulation and Optimization of Supply Chain Models 

for Supercomputing 

Xueping Li, Assistant Professor, University of Tennessee, 408 East 


Xueping.Li@utk.edu, Zhe Zhang 

Cache design is a great challenge in data management in supercomputing and a key 

component to improve the performance by storing data for future access. It is in the 

similitude of behaviors of a supply chain in which suppliers attempt to meet the 

demands from customers. We develop, simulate and optimize supply chain models 

for the cache design of supercomputers and evaluate the performances.

■ WD57 


Portfolio Analysis 


Chair: Mark Zschocke, University of Waterloo, 200 University Avenue 

West, Waterloo, ON, N2L3G1, Canada, mszschoc@uwaterloo.ca 

1 - The Efficient Frontier for Weakly Correlated Assets 

Xili Zhang, Research Associate, School of Business 

Administration,South China University of Technology, 

381 Wushan Road, TianHe District, Guangzhou, 510641, China, 

zhangxili831@gmail.com, Michael J. Best 

Best and Hlouskava have shown that for a Markowitz portfolio selection problem 

having a diagonal covariance matrix and no short sales constraints, the efficient 

frontier is traced out in a monotonic fashion whereby assets are reduced to zero and 

subsequently remain at zero in order of their expected returns. We show that if the 

correlation matrix of the assets is “nearly” diagonal (in a sense to be made precise) 

then the efficient frontier will be traced out in a similar way. 

2 - Sharpe Ratios and Implied Risk Free Rates 

Michael J. Best, Professor, University of Waterloo, 

200 University Avenue West, Waterloo, ON, N2L 3G1, Canada, 

mjbest@uwaterloo.ca 

For the usual mean-variance Markowitz portfolio optimization model having just a 

budget constraint, the choice of a market portfolio implies a unique risk free rate. 

However, practical portfolio optimization problems have many inequality 

constraints. These constraints may cause kinks or points of non-differentiability on 

the resulting efficient frontier. We show that choosing a kink point as a market 

portfolio results in a continuum of implied risk free rates and give a formula for 

them. 

3 - Goal Programming Models for Mutual Funds Portfolio Selection 

From Global Markets 

Mehrdad Tamiz, Professor, Kuwait University, College of Business 

Administration, Kuwait City, Kuwait, mtamiz@yahoo.co.uk 

A vast number of criteria can be taken into consideration in portfolio selection other 

than expected return and variance of return. This talk examines several factors 

representing criteria in a Goal Programming setting for Portfolio Selection of mutual 

funds. The selection is from 20 mutual funds from 10 countries representing 7 

regions, with 7 factors that belong to mutual funds attributes, macroeconomics and 

regional preferences. 

4 - Servitization, Alliance, and Firm Value 

Yeonsung Yoo, Korea University Business School, Anam-dong, 

Seongbuk-Gu, Seoul, 136-701, Korea, Republic of, 

yooys@korea.ac.kr, Hosun Rhim 

We examine impacts of strategic alliance to the shareholder value by measuring the 

stock market reaction associated with companies’ alliance announcements. Firms 

are categorized to understand the impacts of servitization and productization. 

Sample includes Fortune 500 companies. 

5 - Competitive Project Portfolio Management 

Mark Zschocke, University of Waterloo, 200 University Avenue 

West, Waterloo, ON, N2L3G1, Canada, mszschoc@uwaterloo.ca, 

Benny Mantin, Beth Jewkes 

We develop a Competitive Project Portfolio Management (CPPM) model wherein 

two competing firms consider investing into two projects targeting, separately, a 

mature and an emerging market. The returns firms obtain from investments into 

these markets follow an s-shaped curve and depend on both firms’ actions. We find 

that different forms of interactions may arise and outline optimal CPPM strategies. 

We also discuss the market conditions that can lead to these outcomes. 


451 

■ WD58 


Financial Engineering II 


Chair: Stephen Stoyan, University of Southern California, 3715 

McClintock Avenue, GER 240, Los Angeles, CA, 90089, United States of 

America, stoyan@usc.edu 

1 - Corporate Cash Holding, Production and Investments Applied to 

the Agribusiness Sector 

Astrid Prajogo, Princeton University, ORFE Deparment, 

Sherrerd Hall, Princeton, NJ, 08544, United States of America, 

aprajogo@princeton.edu, John Mulvey, Davi Valladao 

We propose a multistage stochastic program of a firm facing uncertain investment 

opportunities and a convex cost of external financing. The model endogenously 

determines the best production, investment, financing and cash holding policies 

over the planning horizon. The implications of corporate cash holdings are 

illustrated for the agricultural sector. 

2 - Efficient Monte Carlo Barrier Option Pricing Under a 

Jump-Diffusion Process 

Samim Ghamami, University of Southern California, 

University Park Campus, Los Angeles, CA, United States of America, 

ghamami@usc.edu, Sheldon Ross 

We present efficient simulation procedures for pricing barrier options when the 

underlying security price follows a geometric Brownian motion with jumps. 

Metwally and Atiya [2002] developed a simulation approach in the same setting for 

pricing knock-out options, but no variance reduction was introduced. We improve 

upon M&A’s approach by innovative applications of well-known variance reduction 

techniques. We also show how to use simulation to price knock-in options. 

3 - Optimal Execution Strategy in the Presence of Price Impact 

Mauricio Junca, University of California, 4141 Etcheverry Hall, 

Berkeley, CA, United States of America, mjunca@berkeley.edu 

An investor needs to execute a long position in the asset by selling at discrete points 

in time, affecting the price of the asset and possibly incurring in a fixed transaction 

cost. The objective is to maximize the discounted revenue. This problem is 

formulated as an impulse control problem and we characterize the value function 

using the viscosity solutions framework. We also analyze the case where there is no 

transaction cost and how this formulation relates with a singular control problems. 

4 - An Algorithm for Portfolio Problems with Discrete 

Choice Constraints 

Stephen Stoyan, University of Southern California, 

3715 McClintock Avenue, GER 240, Los Angeles, CA, 90089, 

United States of America, stoyan@usc.edu 

We consider portfolio models that incorporate a comprehensive set of realistic 

financial constraints, one of which includes the number of securities to hold. 

Uncertainty is considered in the design through the use of stochastic programming. 

The resulting large-scale problem forms a mixed-integer program, where we present 

a model specific decomposition algorithm that generates solutions in reasonable 

time. Computational limitations involved with the approach will also be discussed. 

■ WD61 


Supply Chain, Risk Management 


Chair: Cigdem Gurgur, Professor, Purdue University, Doermer School of 

Business, 2101 East Coliseum Blvd., Fort Wayne, IN, 46805, 

United States of America, gurgurc@ipfw.edu 

1 - Incentives to Invest in Multi-layer IT security Defense in 

Supply Chain Firms 

Tridib Bandyopadhyay, Assistant professor, Kennesaw State 

University, 1000 Chastain Road, Kennesaw, GA, 30144, 

United States of America, tbandyop@kennesaw.edu, Dengpan Liu, 

Srinivasan Raghunathan 

Supply chain firms have integrated business processes. This makes their IT risk 

interdependent. Multilayer IT security defense works like stage-gates, bringing 

interdependency between the efficacies of the successive layers. Thus SC firms face a 

complex inter-firm, inter-layer IT security investment decision scenario. We provide 

a game theoretic model that captures above multi-faceted interdependencies of IT 

security risk in SC firms, and analyze their incentives to invest in IT security.

WD63 

2 - Supplier Selection in Make-to-order Environment with Risks 

Tadeusz Sawik, Professor, AGH University of Science & Technology, 

Department of OR & IT, Krakow, 30059, Poland, 

ghsawik@cyf-kr.edu.pl 

A mixed integer programming approach that uses conditional value-at-risk via 

scenario analysis is proposed for supplier selection and order allocation in supply 

chains with risks. Given a set of customer orders for products, the decision maker 

needs to decide from which supplier and when to purchase parts required for each 

order to meet its due date and to mitigate the impact of low probability supply 

disruptions and high probability supply delays. 

3 - A Multi-Product Risk-Averse Newsvendor with Exponential 

Utility Function 

Sungyong Choi, Rutgers University, 1 Washington Street, 

Room #1072, Newark, NJ, 07102, United States of America, 

sungyongchoi@gmail.com, Andrzej Ruszczynski 

We consider a risk-averse multi-product newsvendor using an exponential utility 

function. We study the asymptotic behavior of the solution with respect to the 

number of products and degree of risk aversion. Then, under reasonable conditions, 

we obtain closed-form approximation which are easy to compute and much more 

accurate than the risk-neutral solution. Then we obtain the analytical and 

numerical insights for the interplay between risk aversion and demand dependence 

structure. 

4 - Flexible Supply Contract Design using Options in Apparel 

Supply Chain 

Bong-Sung Chu, Keio University, 3-14-1, Hiyoshi, Kohokuku, 

Yokohama, Japan, superkensin@yahoo.co.jp, De-Bi Cao 

The major risks in a supply chain mostly come from mismatch of supply and 

demand with the dynamic demand changes. To avoid the risks, nowadays, flexible 

supply contract models using options are attracting significant attention. This paper 

addresses the flexible supply contract design with various types of financial options 

which give holders the right and opportunity to be guaranteed a certain trading 

quantity without stock out or opportunity loss in a single-period two-stage supply 

chain. 

5 - Competition, Diversification and Supplier Selection Under 

Supply Disruptions 

Cigdem Gurgur, Professor, Purdue University, Doermer School of 

Business, 2101 East Coliseum Blvd., Fort Wayne, IN, 46805, 

United States of America, gurgurc@ipfw.edu 

In this study we consider supplier selection and quantity allocation decisions for a 

single firm facing supply unreliability and demand uncertainty. We use a traditional 

newsvendor framework to determine the optimal number of suppliers to place an 

order with and the corresponding quantities of those orders. We explicitly address 

the strategic behavior of suppliers in pricing decisions and we show how the pricing 

decisions of the suppliers change in response to procurement decisions. 

■ WD63 


Decision Analysis VI 


Chair: Jing Ai, A Decision Analysis Approach to Enterprise Risk 

Management, The University of Hawaii at Manoa, Shidler College of 

Business, 2404 Maile Way C305, Honolulu, HI, 96822, 

United States of America, jing.ai@hawaii.edu 

1 - Novel Methods to Support Decision Making and Communicating 

about Risks in Distributed Environments 

Bonnie Ray, Manager, Risk Analytics, IBM T. J. Watson Research 

Center, P.O. Box 218, Yorktown Heights, NY, 10598, United States of 

America, bonnier@us.ibm.com, Lea Deleris 

I provide an overview of a current joint research initiative between IBM Research, 

IBM Dublin, and University College Cork, partially funded by IDA Ireland. The 

research focuses on development of new methods to 1)extract risk information from 

text to form probabilistic graphical models using NLP techniques, 2)reason about 

risk when probabilistic information is imprecise, 3)plan for risk in complex 

environments using stochastic CP techniques and 4)communicate about risks in 

distributed settings. 

2 - A Comparison of Simultaneous Kelly Betting Strategies 

Andrew Grant, University of Sydney, H69 Economics and Business 

Bldg, Cnr Codrington and Rose Sts, Darlington, NS, 2008, Australia, 

andrew.grant@sydney.edu.au, Peter Buchen 

We consider the problem of Kelly betting on simultaneous games, and the relative 

performance of betting strategies that use multi-bets compared to those that do not. 

We develop a simulation model to test the performance of three Kelly betting 

strategies using the empirical odds distribution from the 2007-08 English Premier 

League Season. The results suggest that using multi-bets of all levels outperforms 

the portfolio optimization approach of betting on single game outcomes only. 


452 

3 - Census Uncertainty and Congressional Reapportionment 

Dennis Leber, National Institute of Standards and Technology, 

100 Bureau Dr, Gaithersburg, MD, 20899-8980, United States of 

America, dennis.leber@nist.gov, Jeffrey Herrmann 

The U.S. Constitution requires a decennial census, the results of which are used for 

the reapportionment of congressional representatives. The census results are only 

estimates and contain uncertainty. The Census Bureau can quantify this uncertainty 

through statistical sampling, but this uncertainty is ignored during reapportionment. 

This talk will present decision methods that consider attribute value uncertainty and 

apply these methods to the problem of reapportioning congressional seats. 

4 - A Decision Analysis Approach to Enterprise Risk Management 

Jing Ai, A Decision Analysis Approach to Enterprise Risk 

Management, The University of Hawaii at Manoa, Shidler College of 

Business, 2404 Maile Way C305, Honolulu, HI, 96822, 

United States of America, jing.ai@hawaii.edu, Tianyang Wang 

Enterprise Risk Management (ERM) is an emerging corporate risk management 

concept that proposes to manage risks in an integrated strategic system. One of the 

most important challenges in ERM is the modeling of corporate decision making 

process in light of dependent risks associated with fragmented business activities. 

This paper presents a decision analysis approach to accomplish this task. It serves as 

a practical guideline for managers to optimally allocate resources given risk 

considerations. 

■ WD64 


Health Care, Strategy and Policy 


Chair: Vikram Tiwari, Assistant Professor, University of Houston, 

2200 Business Center Dr, Apt 5104, Pearland, TX, 77584, 

United States of America, vtiwari@Central.UH.EDU 

1 - Clostridium Difficile: System Dynamics Modelling of Hospital 

Infection Outbreaks 

David Lane, Reader in Management Science, London School of 

Economics, Houghton St., London, WC2A 2AE, United Kingdom, 

d.c.lane@lse.ac.uk, Diogo Quintas, Alec Morton 

LSE and UK National Audit Office staff constructed a simulation model to 

understand and control Clostridium difficile outbreaks. Different contamination 

stages, various transmission mechanisms and bed, toilet and staff hand cleaning 

were represented. The model synthesised information from a range of sources. It 

also allowed users to explore and understand the complex consequences of the 

interaction of a number of transmission vectors and policy interventions aimed at 

combating outbreaks. 

2 - Outcome Quality and Efficiency - A Healthcare Perspective 

Scott Lindsey, University of Utah, Operations and Information 

Systems, Salt Lake City, UT, United States of America, 

scott.lindsey@business.utah.edu, Sriram Thirumalai 

This study explores relationships between and operational drivers of outcome 

quality and efficiency within healthcare operations. In the hospital setting, these 

drivers include process standardization, operational focus, and operational 

effectiveness. The study involves the use of stochastic frontier analysis based on data 

from the U.S. Department of Health and Human Services and State of California. 

Study findings, limitations and directions for future research are identified. 

3 - Comparison of Markov and Discrete Event Simulation Models for 

Advanced Prostate Cancer 

Lanting Lu, Associate Research Fellow, Peninsula College of 

Medicine & Dentistry, Veysey Building, Salmon Pool Lane, Exeter, 

EX2 4SG, United Kingdom, lanting.lu@pms.ac.uk 

This work compares discrete event simulation (DES) cost-utility model and Markov 

cost-utility model built for an economic evaluation of a health technology degarelix, 

for advanced prostate cancer, and highlights the differences in the assumptions, the 

input data requirements and outcome estimations. The DES model gives more 

accurate and detailed outputs and conducting a probabilistic sensitivity analysis 

using the simulation model is found to be more straightforward than using the 

Markov model.

4 - Operating Room Utilization Study 

Jihan Wang, Ph.D Student, Wayne State University, 4815 4th Street, 

Rm. 2033, Detroit, MI, United States of America, 

aw0984@wayne.edu, Kai Yang, Susan Yu 

The successful management of operating rooms (OR) is very important 

measurement to any hospital. The OR utilization is a well-acknowledged metric to 

evaluate the performance of the facility. This research investigated 11 factors that 

people think will impact the utilization, such as first case delays and case 

cancellations. We determined the most important factors affecting utilization 

through statistical analysis of the data from Detroit VA medical center. 

5 - Incentive Games in Healthcare Delivery: Pay-for-Performance in 

Primary Care 

Vikram Tiwari, Assistant Professor, University of Houston, 

2200 Business Center Dr, Apt 5104, Pearland, TX, 77584, 

United States of America, vtiwari@Central.UH.EDU, Ana Groznik 

In existing compensation system, primary care physicians’ reimbursement from 

health insurance companies is not tied to patients’ long-term health outcome. Using 

stylized two-stage game theoretical models we compare some structural options of 

payments for physicians in the pay-for-performance market. We identify the 

tradeoffs faced by physicians and insurance companies and derive necessary 

conditions for different reimbursement structures to be optimal (tied to different 

patient characteristics). 

■ WD65 


Supply Chain, Shipping and Transportation 


Chair: Fatih Mutlu, Assistant Professor, Qatar University, Industrial 

Engineering Department, Doha, 2713, Qatar, fatihmutlu@qu.edu.qa 

1 - Cost Benefit Analysis of the Cargo Screening Processes using 

Alternative Evaluation Methods 

David Menachof, Peter Thompson Chair in Port Logistics, University 

of Hull Logistics Institute, Cottingham Road, Hull, HU6 7RX, United 

Kingdom, D.Menachof@hull.ac.uk, Uwe Aickelin, Galina Sherman, 

Peer-Olaf Siebers 

Cost benefit analysis using three different methods, scenario analysis, decision trees 

and simulation, are conducted. These methods are examined in a real world 

situation and compared to actual data using different probabilistic methods of 

estimating costs for port security risk assessment studies. Results show that in simple 

situations, all methods can be equally used. As complexity increases, we show how 

these tools can be used and focus on the limitations of each method. 

2 - Distributor’s Integrated Inventory and Shipment Decisions 

Sudarsan Rangan, Texas A & M University, 4217 TAMU, College 

Station, TX, United States of America, srangan@mays.tamu,edu, 

Ismail Capar, Malini Natarajarathinam 

This analysis is based on a problem faced by a spare parts distributor that supplies n 

retailers. The distributor has to decide on his inventory system and replenish retailer 

inventories using a VMI policy. We analyze this problem to provide optimal 

inventory acquisition and shipment decisions to minimize the overall cost at the 

distributor. 

3 - The Impact of Ship Unloading Time Variability on Port Selection 

Cesar Meneses, Research Assistant, Arizona State University, 

University Drive and Mill Avenue, Tempe, AZ, 85287, United States 

of America, cesar.meneses@asu.edu, Rene Villalobos 

We develop a port selection model based on additional inventory costs caused by 

the variability of service times observed between the arrival of a ship to a port and 

the time a container is released from the port. In particular, we seek to minimize 

total landed costs by selecting the proper port. 

4 - Contract Problems Between a Retailer and a For-Hire Carrier 

Fatih Mutlu, Assistant Professor, Qatar University, Industrial 

Engineering Department, Doha, 2713, Qatar, fatihmutlu@qu.edu.qa 

We study transportation contract problems between a retailer and a for-hire carrier 

delivering retailer’s replenishments. We model a decentralized carrier-retailer 

channel and design frameworks to set contract terms according to individual 

incentives. The terms are i) freight rate, which is based on a two-part tariff 

structure, and ii) shipment size. We also analyze the centralized channel to 

benchmark the channel profits. We propose modifications on the initial contracts to 

assure coordination. 


453 

■ WD67 


Transportation, Freight II 


Chair: Dung-Ying Lin, Assistant Professor, National Cheng Kung 

University, No. 1 University Road, Department of Transportation and 

Commun., Tainan City, Taiwan - ROC, dylin@mail.ncku.edu.tw 

1 - An Integrated Analysis on Global-US Freight Network 

Jiahui (Carol) Wang, Research Assistant, University of Oklahoma, 

202 W. Boyd St., Room 124, Norman, OK, 73019, United States of 

America, jiahuiwang@ou.edu, Guoqiang Shen, P. Simin Pulat 

This research studies on US international freight transportation by an integrated 

view on global freight network and domestic freight network, in which, the 

international freight flows are extended between US ports to final 

destinations/origins (state, census tract, traffic analysis zone…, and the final census 

block level). This research also provides insights into the impact of international 

freight flows on domestic freight flows starting from each port. 

2 - Discrete Time Formulation for the Assignment Problem Applied in 

Cross Docking Facilities 

Georgios Saharidis, Professor, University of Thessaly and Kathikas 

Institute of Research and Technology, Pedion Areos, Volos, 38334, 

Greece, saharidis@gmail.com, Mihalis Golias 

A cross-docking facility is a freight distribution facility representing a critical point in 

a supply chain. The scheduling of the inbound trucks at a cross docking facility is a 

classical assignment problem. In this paper two mathematical formulations are 

presented to schedule inbound trucks to doors at a cross-docking facility, and 

compared to the classical machine scheduling formulation. 

3 - Scheduling Commercial Vehicle Queues at a Canada-US 

Border Crossing 

Michael Haughton, School of Business & Economics, Wilfrid Laurier 

University, Waterloo, ON, Canada, mhaughton@wlu.ca, 

K.P. Sapna Isotupa 

Using the context of queue operations at a major Canada-U.S. commercial border 

crossing for truck-borne trade flows, we report on a computer simulation study to 

predict the likely impacts of smoothing those flows. We quantify the operational 

and resource efficiencies of flow smoothing, not only for trans-border trucking 

companies and other commercial organizations involved in trans-border supply 

chains but also for government authorities with regulatory jurisdiction at border 

crossings. Our study achieves two major goals. First, it extends the queueing 

literature’s range of settings in which queue management systems based on flow 

smoothing are studied. Second, it adds quantitative precision to the post-9/11 

discourse on reducing impediments to the performance of commercial trucking 

operations that support trans-border supply chains. 

4 - Quantifying the Value of Information Sharing in Supply Chain 

Inventory Management 

Dung-Ying Lin, Assistant Professor, National Cheng Kung University, 

No. 1 University Road, Department of Transportation and Commun., 

Tainan City, Taiwan - ROC, dylin@mail.ncku.edu.tw, Nai-Wen Hsu 

We first develop a model in which each stakeholder in the supply chain 

independently chooses the optimal inventory policy that has the lowest total 

expected cost (TEC) with stochastic demand. Then a model recognizing the 

inventory coordination is developed to estimate the total expected cost of the 

coordinated supply chain. The value of information sharing in supply chain 

inventory management is quantified by the difference of TECs between these two 

models. 

■ WD68 


Strategy/Strategic Planning II 


Chair: Michael Bean, Forio Business Simulations, 333 Bryant Street, 

San Francisco, CA, 94133, United States of America, mbean@forio.com 

1 - A Representation Model and Innovation Heuristics for Business 

Ecosystems of Product-Service Systems 

Minjeong Baek, Master Student, Seoul National University, 599 

Kwanakro, Kwanakgu, Seoul, Korea, Republic of, 

kelly706@snu.ac.kr, Changmuk Kang, Yoo S. Hong 

A business ecosystem is growing more complex in recent days, as more products 

and services are integrated as a product-service system, and more stakeholders are 

involved. This study presents a rigorous model for representing and analyzing a 

business ecosystem consisting of various stakeholders that interact with products 

and services. Based on this model, six heuristic rules for innovating an ecosystem 

and industry cases have been proposed.

WD69 

2 - The Demand-side Dynamics of Firms’ Intra-industry Exit in a 

Geographically Fragmented Industry 

Lalit Manral, Assistant Professor, University of Central Oklahoma, 

100 N Universtiy Drive, Edmond, OK, 73034, United States of 

America, lmanral@uco.edu, Kathryn Harrigan 

This paper explores the demand-side dynamics of firms’ selective exit from 

geographic sub-markets in a geographically fragmented industry. It provides 

theoretical arguments and empirical evidence to explain how the demand-side 

structural conditions differentially influence firms’ choice to manage their 

geographic scope over time. A unique panel dataset drawn from a natural 

experiment in the US long-distance telecommunications services industry during 

1984-1996 is used to test the hypotheses. 

3 - Strategic Capacity Investment and Pricing Decisions for 

Substitutable Products with Partial Flexibility 

Sharethram Hariharan, Graduate Research Associate, Oklahoma 

State University, 322 Engineering North, OSU Stillwater, Stillwater, 

ok, 74078, United States of America, sharethramh@gmail.com, 

Tieming Liu, Ho-Yin Mak, Z. Max Shen 

We study the capacity investment and responsive pricing decisions for a firm facing 

random demands for two substitutable products. Resources are partially flexible. A 

resource can be used to produce the other type of product, but with some efficiency 

loss. We evaluate the impacts of demand variability, correlation and risk-aversion on 

these decisions. 

4 - Creating Web-based Simulations and Interactive Data Visualizations 

Michael Bean, Forio Business Simulations, 333 Bryant Street, San 

Francisco, CA, 94133, United States of America, mbean@forio.com 

Simulations and interactive data visualizations that run in browsers have the 

advantages of global accessibility, simple distribution, and the ability to monitor and 

collect data on usage. However, simulations need to be modified in order to 

effectively use the online medium. Usability design is critical to create simulations 

that will be used by a diverse, global audience with limited knowledge of 

simulation, short attention spans, and unarticulated use objectives. 

■ WD69 


Reverse Logistics II 






1 - An Optimization Model to Determine Product Servicing Policies in 

Green Supply Chains 

Ruth Gledhill-Holmes, PhD Student, University of Central Florida, 

4000 Central Florida Blvd, Orlando, FL, 32816, United States of 

America, ruth_gledhill_holmes@yahoo.com, Ola Batarseh, 

Dima Nazzal 

This research considers the possible service policy options necessary to extend and 

maximize a product’s usable life (hence resulting in reduced production rates and 

the associated environmental benefits of this) in conjunction with the possible 

recovery options for maximizing the re-capture of end-of-use products (hence 

resulting in reduced waste, reduced need for first-use materials, and recovery of 

value for the manufacturer). 

2 - Integrating Environmental Considerations into Pricing and 

Production Planning Models 

Dima Nazzal, Assistant Professor, University of Central Florida, 4000 

Central Florida Blvd., Orlando, FL, 32816, United States of America, 

dnazzal@mail.ucf.edu, Ali Bozorgi 

We propose and test a multi period optimization model that integrates the 

environmental impact of a product into the production and pricing decisions. The 

model considers the trade-off between economic and environmental objectives. 

3 - A Model for Reverse Logistics Under Uncertainty 

Theresa Barker, PhD, University of Washington, Box 352650, Seattle, 

WA, 98115, United States of America, barkertj@u.washington.edu, 

Zelda Zabinsky 

Many mixed-integer linear programming models have been developed for reverse 

logistics. These models typically require producers to make certain network design 

decisions in advance, such as how to collect the return product and where to 

perform testing. We present an integrated MILP model that combines high-level and 

detailed network design decisions, along with chance-constrained programming to 

analyze the tradeoffs between various network configurations. 


454 

4 - Creating Value Through Product Stewardship and Take-Back 

Ronald Lembke, Associate Professor, Supply Chain, University of 

Nevada, MGRS /0028, Reno, NV, 89557, United States of America, 

rtl@unr.edu, Zac Rogers, Dale Rogers 

Secondary markets provide a place for unwanted items to be bought and sold, 

which diverts them from landfills, reducing the products’ ecological impact and 

creating economic value. Secondary markets divert a large number of products from 

landfills and create jobs, resulting in substantial economic value in the process. 

Although not reflected in current government metrics, a conservative estimate is 

that the secondary market represents 2.28 percent of the 2008 U.S. GDP. 

■ WD70 


Optimization Models in Air Traffic Management 



Chair: Andrew Churchill, Graduate Research Assistant, 

University of Maryland, 1173 Martin Hall, College Park, MD, 20742, 

United States of America, churchil@umd.edu 

1 - The Impact of Induced Technological Change on Air Traffic 

Management 

Megan Ryerson, Institute of Transportation Studies, 109 McLaughlin 

Hall, UC Berkeley, Berkeley, CA, 94720, United States of America, 

msmirti@berkeley.edu 

Short-haul aviation is a focus of transportation policy today due to growing aircraft 

diversity and uncertainty surrounding carbon emissions policies. We develop an 

analytic total logistics cost model of short-haul corridor serving multiple passenger 

groups and find that turboprops, alone and in a fleet mix with jets, minimize cost 

for levels of policy seen today. This aircraft mix in response to a carbon emissions 

policy has implications for air traffic management. 

2 - Converging Upon Basic Feasible Solutions Through 

Dantzig-Wolfe Decomposition 

Joseph Rios, NASA, Mail Stop 210-15, Moffet Field, CA, 94035, 

United States of America, Joseph.L.Rios@nasa.gov, Kevin Ross 

We derive an important property for solving integer programs by examining the 

master problem in Dantzig-Wolfe Decomposition. Namely, if a linear program can be 

decomposed appropriately, a mapping exists between basic feasible solutions in the 

master and original problems. This has implications for integer programs where the 

convex hull has mostly integer corner points. We highlight the significance through 

experiments on a large-scale traffic flow model for the National Airspace System. 

3 - Collaborative Strategies for Traffic Management in the Airspace 

Flow Program 

Amy Kim, amykim0603@gmail.com, Mark Hansen 

We investigate methods that aim to minimize the user-cost impact of a future AFP 

that employs rerouting and ground delay. Methods are assessed using a simple flight 

cost specification; the deterministic part represents flight costs known to the FAA, 

while the stochastic part represents costs known only to the airline. Models 

featuring different allocation schemes, user inputs, and timing of information gather 

are compared. We explore tradeoffs between information quality and timeliness. 

4 - Assessing the Impact of Stochastic Capacity Variation on 

Coordinated Air Traffic Flow Management 

Andrew Churchill, Graduate Research Assistant, University of 

Maryland, 1173 Martin Hall, College Park, MD, 20742, United States 

of America, churchil@umd.edu, David Lovell, Michael Ball 

In this research, an integer optimization model for coordination between air traffic 

flow management initiatives and a three parameter characterization of resource 

capacity are used to understand the effects of random capacity variations. 

Employing the optimization model within a Monte Carlo simulation, the results 

suggest that each of the three parameters characterizing capacity have marginally 

increasing impacts over a wide range of conditions.

■ WD72 


Facility Logistics Session: Warehouse Order Picking 



Chair: Russell Meller, Hefley Professor of Logistics and Entrepreneurship, 

University of Arkansas, 4207 Bell Engineering, Fayetteville, AR, 72701, 

United States of America, rmeller@uark.edu 

1 - Boosting Productivity in Multiple-Aisle Order-Picking by Cellular 

Bucket Brigades 

Yun Fong Lim, Assistant Professor, Singapore Management 

University, 50 Stamford Road #04-01, Singapore, Singapore, 

yflim@smu.edu.sg 

We introduce a new design of bucket brigades for order-picking in multiple, parallel 

aisles. Workers pick products from one side of an aisle when they proceed in one 

direction and they pick from the other side of the aisle, possibly for other customer 

orders, when they proceed in the reverse direction. We assume hand-off times are 

significant and propose simple rules for workers to share work. The new design can 

be substantially more productive than traditional bucket brigades. 

2 - Using Storage Profiles with Bucket Brigade Order Picking to 

Improve Productivity 

Don Eisenstein, Professor of Operations Management, University of 

Chicago, Booth School of Business, 5807 South Woodlawn Avenue, 

Chicago, IL, 60637, United States of America, 

Don.Eisenstein@chicagobooth.edu, Yeming Gong 

We compare bucket brigade and zone order picking protocols using various storage 

profiles. We find that the flexibility of a bucket brigade protocol can lead to some 

considerable advantages in productivity. 

3 - Modeling the Throughput Performance of a Picking Machine Order- 

Fulfillment Technology 

Jennifer Pazour, PhD Candidate, University of Arkansas, 

4207 Bell Engineering Center, Fayetteville, AR, 72701, 

United States of America, jpazour@uark.edu, Russell Meller 

Picking machines are an example of a stock-to-picker piece-level fulfillment 

technology that consists of two or more pick stations, a common storage system, 

and an integrated conveyor system. We develop a stability function model based on 

queuing theory that is capable of determining if a picking machine will be able to 

meet a throughput requirement. We then illustrate how a manager could use our 

analytical model to answer picking machine design questions. 

■ WD73 


Joint Session TSL/ SPPSN: Mitigating Disaster Impact 

Through Planning for Emergency Response 




Chair: Elise Miller-Hooks, University of Maryland, 1173 Glenn Martin 

Hall, College Park, MD, 20742, United States of America, 

elisemh@umd.edu 

1 - A Robust P-center Model for Emergency Facility Location Planning 

Chung-Cheng Lu, Assistant Professor, National Taipei University of 

Technology, 1 Section 3, Chung-Hsiao E. Road, Information and 

Logistics Management, Taipei, 106, Taiwan - ROC, 

jasoncclu@gmail.com 

This study deals with the p-center problem with uncertain travel times between 

emergency facilities and affected areas which may arise in large-scale disasters. The 

uncertain travel times are represented by intervals and a robust optimization 

approach that minimizes the worst-case deviation from optimal solutions was 

developed. Numerical experiments were conducted to demonstrate the effectiveness 

of the approach and to examine the tradeoff between robustness and optimality. 

2 - Locating Support Facilities for Large Scale Emergencies 

Anurag Verma, Graduate Student, Texas A&M University, Industrial 

and Systems Engineering, College Station, TX, 77843-3131, 

United States of America, anuragverma@tamu.edu, Gary Gaukler 

We provide large scale emergency facility location models that account for the 

unavailability of a disaster response facility depending on its location and the 

position of the disaster. We find that the locations suggested by the models in this 

paper significantly reduce the expected cost of transportation of supplies when we 

consider the damage a disaster causes to the support facilities and areas near it. 


455 

3 - Decision Support for Urban Search and Rescue 

Lisa (Lichun) Chen, lcchen@umd.edu, Elise Miller-Hooks 

The urban search and rescue (USAR) team deployment problem seeks an optimal 

deployment of USAR teams to disaster sites so as to maximize the expected number 

of saved lives over the search and rescue period. A multistage stochastic program 

that captures problem uncertainty and dynamics is presented. A column generationbased 

technique is proposed for its solution. The techniques is illustrated on a case 

study involving the 2010 Haiti earthquake. 

4 - Using TRANSIMS for On-line Transportation System Management 

During Emergencies 

Adel Sadek, Associate Professor, University at Buffalo, the State 

University of New York, 233 Ketter Hall, Buffalo, NY, 14260, 

United States of America, asadek@buffalo.edu, Daniel Fuglewicz, 

Alan Blatt, Yunjie Zhao 

The topic of transportation systems management during emergencies has recently 

received national attention. This study has two primary objectives: (1) to assess the 

level to which the TRANSIMS model can be used for the online management of 

transportation systems during emergencies; and (2) to develop any additional 

functionality needed for that. As a case study, the research uses the Buffalo-Niagara 

area known for its winter weather and numerous snow storms. 

■ WD74 


Joint Session TSL/ Service Science: Emergency 

Transportation, Logistics, and Evacuation Service 

Sponsor: Transportation Science and Logistics Society/ 

Service Science 


Chair: Tao Yao, Assistant Professor, Pennsylvania State University, 349 

Leonhard Building, University Park, PA, 16802, United States of America, 

tyy1@engr.psu.edu 

1 - Distributing Supplies in a Multi-city Infectious Disease Outbreak 

Fernando Ordonez, Associate Professor, University of Southern 

California, Department of Industrial and Systems Eng, 3715 

McClintock Ave, Los Angeles, CA, 90007, United States of America, 

fordon@usc.edu, Yingtao Ren 

We consider the distribution of medicine to mitigate a multi-city infectious disease 

outbreak under parameter uncertainty. We build a two-stage stochastic 

programming model which has integer variables in both stages. This problem is 

solved with a modified Bender’s method to recover dual information from the 

second stage MIP and to speed up convergence. We illustrate the use of the model 

and the algorithm in planning an emergency response to a hypothetical national 

smallpox outbreak. 

2 - Evacuation Planning Under Uncertainty: A Distributional Robust 

Chance-Constrained Approach 

Tao Yao, Assistant Professor, Pennsylvania State University, 349 

Leonhard Building, University Park, PA, 16802, United States of 

America, tyy1@engr.psu.edu, Byung Do Chung, Bo Zhang 

This talk provides a Chance-constrained Programming approach for evacuation 

transportation planning which are robust to uncertainty. We focus on demand 

uncertainty with partial distributional information and develop an approximation 

scheme to reformulate the problem as a deterministic convex program which is then 

proved to be safe and computationally tractable. Numerical experiments are 

provided to illustrate the performance of the proposed approach. 

3 - Demand Management for Evacuation Planning 

Douglas Bish, Assistant Professor, Grado Department of Industrial 

and Systems Engineering, Virginia Tech, Blacksburg, VA, 24061, 

United States of America, drb1@vt.edu 

In hurricane evacuations traffic often overwhelms the infrastructure, causing 

congestion that can increase the risk to the population. Supply management 

strategies are used in evacuation planning, but they are often insufficient. We study 

the use of demand management strategies that attempt to structure the evacuation 

demand in order to minimize congestion. Specifically, we study network flow 

problems that incorporate important aspects of evacuee behavior and evacuation 

management issues. 

4 - Optimal Evacuation Routing Under Dynamic Stochastic Risks 

Chi Xie, Research Fellow, The University of Texas at Austin, 1 

University Station, Austin, TX, 78712, United States of America, 

chi.xie@mail.utexas.edu, Tao Yao, Travis Waller 

We propose a risk-based modeling framework and solution procedure for dynamic 

evacuation planning under hazardous material releases, which integrates the latest 

atmospheric dispersion modeling and network flow modeling techniques. A robust 

dynamic evacuation routing model is presented for devising evacuation plans that 

minimize the most risky exposure to the evacuating population under timedependent, 

uncertain hazardous effluents.

Wednesday

Transform your PDFs into Flipbooks and boost your revenue!

Delete template?

Save as template ?