Technical Sessions – Monday July 11

TC-03 IFORS 2011 - Melbourne
Networks are one of the essential building blocks of society. Social movements
make use of networks to exchange information, but on the negative side
so do criminal and terrorist networks. This paper investigates how players in a
network can use additional links to optimally defend themselves against the impending
attack on the links of the network by a network disruptor. We find that
contrary to the intuition of the public-good nature of network protection underprotection
is no problem. Instead, players will either go for "over-protection’
or will remain in a minimally connected network.
2 - A Dynamic Game Analysis of Social Networking Services
Emiko Fukuda, Department of Computer Science, National
Defense Academy of Japan, 2398686, Yokosuka, Kanagawa,
Japan, emiko@nda.ac.jp
Social networking Services (SNS), such as MySpace or FaceBook, are one of
the most popular websites. We model a diffusion process of users of a SNS
as a dynamic game in which (1) each user can choose not to use the SNS in
her/his turn, and (2) network effect of the SNS depends on a history of players’
actions.Then we derive the necessary condition under which the state where
every player is an active user is the unique Markov perfect equilibrium outcome.
Moreover, we propose an incentive mechanism that enables the number
of active users grows steadily.
3 - Pollution, Pigouvian Taxation and Cartel Stability
Luca Lambertini, Economics, University of Bologna, Strada
Maggiore 45, 40125, Bologna, Italy, luca.lambertini@unibo.it,
Arsen Palestini, Andrea Mantovani
We assess the effects of Pigouvian taxation on firms’ ability to collude in a
Cournot supergame where production pollutes the environment. It turns out
that (i) if marginal production cost is constant, taxation enhances the intensity
of collusion, creating a tradeoff between environmental regulation and antitrust
policy; (ii) if instead marginal production cost is increasing, Pigouvian taxation
has a definite pro-competitive effect.
4 - A Dynamic Game of Pollution and Environmental Absorption
Capacity
Fouad El Ouardighi, Operations Management, ESSEC Business
School, Avenue Bernard hirsch, BP 105, 95021, Cergy Pontoise,
France, elouardighi@essec.fr
In this paper, we extend the model in El Ouardighi and Benchekroun (2010)
with a dynamic environmental absorption capacity as a two-player dynamic
game. We notably compare the open-loop Nash equilibrium with the cooperative
solution. Also, the stability of both equilibria is analyzed.
� TC-03
Tuesday, 15:00-16:30
Meeting Room 102
Metaheuristics
Stream: Meta-heuristics
Invited session
Chair: Weiqi Li, School of Management, University of
Michigan-Flint, 303 East Kearsley Street, 48502, Flint, Michigan,
United States, weli@umflint.edu
1 - Static Data Segment Location Problem in Information
Networks
Goutam Sen, Industrial Engineering and Operations Research
(IITB), Clayton School of Information Technology (Monash),
IITB Monash Research Academy, C - 519, Hostel 12, Indian
Institute of Technology, Powai, 400076, Mumbai, Maharashtra,
India, goutam.sen@iitb.ac.in, Mohan Krishnamoorthy, David
Abramson, Vishnu Narayanan, Narayan Rangaraj
We consider the problem of locating data mirrors in multiple locations in a
communications network so as to provide fast access to data. Full replication
of massively large databases, however, incurs large transmission and bandwidth/access
costs. We propose a new data segment location model, and formulate
it as a mixed integer linear program. We develop a simulated annealing
algorithm for our problem and also provide results for an exact solution strategy
based on a relaxation approach. We test our models on real data and also
point to future research.
50
2 - Comparing the Performance of Different Metaheuristics
for Solving a Location Problem
Javier Alcaraz, Dept. Estadística, Matemáticas e Informática,
Universidad Miguel Hernández de Elche, Av. Universidad s/n,
03202, Elche, Alicante, Spain, jalcaraz@umh.es, Mercedes
Landete, Juan Francisco Monge
In this work we present two new metaheuristics to solve a location problem.
Both algorithms make use of the same solution encoding and some procedures
based on the problem specific knowledge have been incorporated. To study the
performance of the algorithms and to compare their efficiency, we have carried
out an extensive computational experiment solving well known benchmark instances.
We have also compared them with a genetic algorithm proposed to
solve the P-median problem. The results show a similar behavior of the new
methods, which are much more efficient than the genetic algorithm.
3 - Attractor-Based Approach To Dynamic Traveling Salesman
Problem
Weiqi Li, School of Management, University of Michigan-Flint,
303 East Kearsley Street, 48502, Flint, Michigan, United States,
weli@umflint.edu
The paper introduces the concept of solution attractor in local search for combinatorial
optimization problems. A solution attractor drives local search trajectories
to converge into a small region in the solution space that contains the
most promising solutions. A method for constructing the solution attractor for
the traveling salesman problem (TSP) is presented. Based on the attractor concept,
this paper introduces an attractor-based procedure to tackle dynamic TSP.
This procedure not only generates high quality solution, but also provides the
flexibility and stability of the search system.
� TC-04
Tuesday, 15:00-16:30
Meeting Room 103
Revenue Management 2
Stream: Revenue Management and Dynamic Pricing
Invited session
Chair: Massoud Talebian, University of Newcastle, NSW 2308,
Callaghan, NSW, Australia, Masoud.Talebian@newcastle.edu.au
1 - Competitive Effects of Minimum Sales Volumes to Trigger
Higher Commissions
Massoud Talebian, University of Newcastle, NSW 2308,
Callaghan, NSW, Australia,
Masoud.Talebian@newcastle.edu.au, Guillermo Gallego
We consider a game between two capacity providers that compete for customers
through a broker. The broker works on commission margins and sells
to both loyal and no-loyal customers. We study the effects of requiring minimum
sales volumes, or thresholds, to trigger an increase in commissions. This
setting is similar to quantity discount contracts, where suppliers require minimum
purchase volumes to give discounts on the total purchase. We analyze
two settings, when commission margins are endogenous and also when they
are exogenous. We show that with endogenous margins the introduction of
thresholds results in less revenue for the smaller provider and more revenue for
the broker, with the revenue for the larger provider remaining unchanged. This
is in sharp contrast to the case where margins are exogenous. In this case at
least one provider has an incentive to impose minimum sales volumes to trigger
commissions, usually at the expense of the broker. Consequently, the gains
from the case of exogenous margins are a mirage in the full equilibrium, where
commission margins are endogenous.
2 - Issues in Partitioned Revenue Management
Max Gerlach, Information Systems, Freie Universitaet Berlin,
Garystr. 21, 14195, Berlin, Germany, max-gerlach@web.de,
Catherine Cleophas, Natalia Kliewer
While maximum revenue may be achievable by optimization across the overall
inventory, within the partitioned structure of conglomerates as well as in
corporate alliances, this is often impossible due to legal, organizational and
technical boundaries. We present an overview of constellations in which inventory
can be split into sets of substituting or complementing products as well
as alternative ways to implement revenue management in these constellations.
We illustrate the effects of different strategies using stochastic simulations and
provide applications in the airline industry.