Tolga Seyhan PhD defense 2-3-12

DISSERTATION DEFENSE ANNOUNCEMENT

TITLE:              Network Design under Competition

SPEAKER:      Tolga Han Seyhan, Ph.D. Candidate,
                       Department of Industrial and Systems Engineering

DATE:             Friday, February 3, 2012 at 10:00 a.m.

LOCATION:      Room 453 Mohler, 200 W. Packer Avenue

ABSTRACT: Bilevel problems generalize the idea of Stackelberg games and are useful in modeling problems with competition. They are often not tractable and suffer heavily from computational complexity, mainly because the leader has to anticipate the best response of the follower. We present heuristic methods to reformulate these problems so that they are tractable (unlike the originals) and still bear a high representation power (like the originals).
First, we study the Stackelberg-game variant of the maximal covering location problem. We use a greedy heuristic as a proxy for the follower’s decision, pose it as a feasibility problem, and embed it into the leader’s problem as constraints. Computational studies indicate short solution times; and even if the follower responds optimally, the method still provides an effective heuristic for the leader’s problem.
Next, we consider a competitive distribution system design problem. We first develop a model to find equilibrium strategies in the simultaneous-move game. Then, for the Stackelberg setting, we develop a heuristic reformulation of the leader problem using a pruning algorithm for the follower. We obtain near-optimal design strategies and both models offer short solution times.
Finally, we revisit the college admissions problem to study networking in a labor market game. We currently study the applicants’ side of this game and have developed an algorithm that characterizes their optimal choices. Then, we formulate the problem as a non-additive longest-path problem and introduce an algorithm that also solves the joint application-signaling problem.

BIOGRAPHY: Tolga Han Seyhan is a Ph.D. candidate in the Department of Industrial and Systems Engineering at Lehigh University. He received his B.S. and M.S. degrees in Industrial Engineering in 2004 and 2007, from Middle East Technical University in Turkey. He joined Lehigh University in 2007. His research focuses on mathematical programming and game theory with applications to logistics, supply chain management and energy systems. He is a member of INFORMS, MSOM and TSL societies, and SOLA. Tolga will join Amazon.com as a Sr. Operations Research Analyst.