Speaker: Kübra Tanınmış
Abstract:
In this work, we study bilevel optimization problems with integer linear upper-level and integer quadratic lower-level. Relying on bilevel-free sets derived from improving directions, we develop a disjunctive cut approach to exclude bilevel-infeasible points. We show that, for a given improving direction, the associated disjunctive cuts can be obtained by solving a linear program. We derive conditions that allow the number of disjunctions to be reduced, which may in turn enlarge the bilevel-free set, and we propose several strategies to identify improving directions and generate cuts efficiently. We evaluate the proposed framework on a test set comprising purely binary instances, general integer instances, and instances with a non-convex follower objective function.
Bio:
Kübra Tanınmış is an Assistant Professor of Industrial Engineering at Koç University since 2023. She has received her Ph.D.degree in Industrial Engineering from Boğaziçi University in 2020. In her doctoral research she addressed bilevel optimization problems related to spreading information over social networks in a competitive environment. She worked as a post doctoral researcher at the Institute of Production and Logistics Management at Johannes Kepler University Linz from 2020 to 2022. Her research interests include bilevel optimization, interdiction games, influence spread on networks, and designing algorithms for discrete optimization problems.
