26.11.2012 Views

Technical Sessions – Monday July 11

Technical Sessions – Monday July 11

Technical Sessions – Monday July 11

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

TB-07 IFORS 20<strong>11</strong> - Melbourne<br />

cedex 3, France, Pierre.Dejax@emn.fr, Hugues Dubedout,<br />

Thomas Yeung<br />

Real-life IRPs are often subject to uncertainties such as customer demand and<br />

supply disruption. In order to deal with these uncertainties, robust solutions<br />

are often proposed. To that aim, we propose a methodology consisting of the<br />

1. scenario generation to model the uncertainties, 2. obtaining robust IRP<br />

solutions from custom heuristics, and 3. evaluation of the impact of the different<br />

uncertainties on the robust solutions. This methodology was tested on<br />

both simulated test-cases and real-life problems as encountered in cryogenic<br />

gas distribution at Air Liquide.<br />

� TB-07<br />

Tuesday, <strong>11</strong>:00-12:30<br />

Meeting Room 106<br />

Discrete Optimisation<br />

Stream: Discrete Optimisation<br />

Invited session<br />

Chair: François Vanderbeck, Institut de Mathématiques de Bordeaux,<br />

Université Bordeaux1 & INRIA Bordeaux, 351 cours de la<br />

Libération, F-33405, Talence- CEDEX, France,<br />

fv@math.u-bordeaux1.fr<br />

1 - Bi-dynamic Constraint Aggregation and Subproblem<br />

Reduction for Crew Pairing Problem<br />

Francois Soumis, GERAD, 3000 Cote Ste-Catherine, H3T 2A7,<br />

Montreal, Québec, Canada, francois.soumis@gerad.ca,<br />

Mohammed Saddoune, Issmail Elhallaoui, Guy Desaulniers<br />

Bi-dynamic constraint aggregation is a new approach introduced to speed up<br />

the column generation method when solving degenerate set partitionig problems.<br />

The agregation profits from degeneracy to reduce the size of the master<br />

problem. The subproblems reduction eliminates arcs with less promising<br />

marginal costs. This reduces significantly the number of column generation iterations<br />

when we start from a good heuristic solution. These reductions lead to<br />

big savings in solution time. Tests on some hard real life crew pairing problems<br />

will be presented with the trivial initial solution: "follow the aircraft" and with<br />

an initial solution obtained with a linear programing approximation.<br />

2 - Large Sized Extended Formulations for Routing and<br />

Scheduling Problems<br />

Eduardo Uchoa, Engenharia de Produção, Universidade Federal<br />

Fluminense, Rua Passo da Pátria 156, São Domingos,<br />

22430-210, Niterói, Rio de Janeiro, Brazil,<br />

uchoa@producao.uff.br, Artur Pessoa, Marcus Poggi de Aragão,<br />

Rosiane de Freitas Rodrigues, Lorenza Moreno<br />

We present an extended formulation that allow viewing and parallel machine<br />

scheduling and capacitated vehicle routing problems in a very similar way:<br />

the machines correspond to the vehicles, jobs to the clients, and the processing<br />

times to client demands. This means that known VRP cuts can be effectively<br />

used in those scheduling problems and vice versa. However, the pseudopolynomially<br />

large size of the formulation requires a sophisticated branch-cutand-price<br />

implementation. Computational results show good results in comparison<br />

with other known methods<br />

3 - Using the Cost Splitting Dual for Multiple Resource<br />

Constrained Shortest Path Problems<br />

Olivia Smith, Mathematics and Statistics, University of<br />

Melbourne, 3052, Parkville, Vic, Australia,<br />

livsmith21@gmail.com, Natashia Boland<br />

Shortest path problems with a single resource constraint are much easier to<br />

solve than problems where there are many resource constraints. This makes decomposition<br />

approaches attractive as they can exploit the easier problem structure.<br />

We explore a reformulation in which we define a set of variables for each<br />

resource with extra constraints enforcing the equality of these variables. Relaxing<br />

these constraints in a Lagrangean fashion leads to a dual known as the Cost<br />

Splitting Dual. We investigate the use of this dual to assist in preprocessing and<br />

solving the multiple resource problem.<br />

4 - Numerical Experiments with Column Generation for Extended<br />

Formulations<br />

42<br />

François Vanderbeck, Institut de Mathématiques de Bordeaux,<br />

Université Bordeaux1 & INRIA Bordeaux, 351 cours de la<br />

Libération, F-33405, Talence- CEDEX, France,<br />

fv@math.u-bordeaux1.fr<br />

Extended formulations for integer programs are often tight but too large for<br />

a direct treatment. When they stem from a decomposition principle, a hybrid<br />

column generation procedure applies that mimics that for the Dantzig-Wolfe<br />

reformulation. Pricing subproblem solutions can be expressed in the extended<br />

space and added to the current restricted formulation along constraints that become<br />

active. Such “column-and-row generation” has a comparative advantage<br />

over standard column generation: in the extended space, subproblem solutions<br />

can be recombined resulting in faster convergence.<br />

� TB-08<br />

Tuesday, <strong>11</strong>:00-12:30<br />

Meeting Room 107<br />

Health Care Operations II<br />

Stream: Service & Health Care Operations<br />

Invited session<br />

Chair: Karen Smilowitz, Industrial Engineering and Management<br />

Sciences, Northwestern University, 2145 Sheridan Road,<br />

Technological Institute M233, 60208-3<strong>11</strong>9, Evanston, Illinois, United<br />

States, ksmilowitz@northwestern.edu<br />

1 - Improving Access to Community-based Chronic Care<br />

through Improved Capacity Allocation<br />

Karen Smilowitz, Industrial Engineering and Management<br />

Sciences, Northwestern University, 2145 Sheridan Road,<br />

Technological Institute M233, 60208-3<strong>11</strong>9, Evanston, Illinois,<br />

United States, ksmilowitz@northwestern.edu, Sarang Deo,<br />

Seyed Iravani<br />

Most health care operation models focus either on efficiency improvements in<br />

the delivery system or improvements in clinical decisions. We consider a novel<br />

setting of community-based delivery of chronic asthma care, where it is necessary<br />

to integrate these two approaches. We develop and analyze a joint disease<br />

progression and capacity allocation model to investigate how operational decisions<br />

can improve population level health outcomes. We test our findings using<br />

data provided by Mobile C.A.R.E, a community-based provider of asthma care<br />

to public school students in Chicago.<br />

2 - Role of Flexibility in Managing Operating Room Resources<br />

in Hospitals<br />

Kannan Sethuraman, Melbourne Business School, 3441<br />

Woodcliff Road, Melbourne, Australia,<br />

K.Sethuraman@mbs.edu, Devanath Tirupati<br />

We examine the role of flexibility in managing operating room resources in a<br />

tertiary hospital providing elective surgeries in a variety of specialties. We consider<br />

two kinds of flexibility and develop a queueing theoretic stochastic model<br />

based on approximations that is useful for a preliminary analysis to assess the<br />

impact of alternative strategies. The analytic model is complemented by a simulation<br />

model to provide detailed performance evaluation. Managerial insights<br />

are derived from computational results based on representative data from a real<br />

life case study.<br />

3 - Drivers of Excellence in Veterans’ Administration Clinics<br />

Subramaniam Ramanarayanan, Anderson School of<br />

Management, UCLA, 90403, Los Angeles, CA, United States,<br />

subbu@anderson.ucla.edu, Kumar Rajaram<br />

Although there is clear consensus about a demonstrable need for improvement<br />

of outcomes such as access, wait times and quality in healthcare, there is little<br />

clarity about how this might be achieved. While there is a significant amount of<br />

research focused on the impact of external structural forces on outcomes, our<br />

understanding of the importance of various intra-organizational characteristics<br />

in determining outcomes is still fairly limited. We address this research gap by<br />

focusing on the structural determinants of outcomes in a sample of Veteran’s<br />

Administration Clinics in the US.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!