【5月14日】【管理科学与工程学院学术论坛】Towards an Adaptive Memory Programming Approach for the Robust Capacitated Veh发布日期：2019-09-12 21:12:58
主题：Towards an Adaptive Memory Programming Approach for the Robust Capacitated Vehicle Routing Problem
报告人：Panagiotis P. Repoussis, PhD，Assistant Professor in Operations (tenure track)， STEVENS INSTITUTE of TECHNOLOGY
Panagiotis P. Repoussis，博士，史蒂文斯理工学院技术管理学院助理教授，毕业于雅典经济与商业大学，获管理科学与技术方向博士学位，主要研究兴趣为：Quantitative analysis, modeling and optimization methods for a variety of problems in transportation and distribution logistics; network design; manufacturing and service operation management; terminal management; production scheduling; vehicle routing and scheduling。曾经在Transportation Science、European Journal of Operations Research、 Optimization Letters等本领域知名学术期刊上发表十多篇学术论文。
报告摘要：We present an Adaptive Memory Programming (AMP) meta heuristic to address the Robust Capacitated Vehicle Routing Problem under demand uncertainty. Contrary to its deterministic counterpart, the robust formulation allows for uncertain customer demands, and the objective is to determine a minimum cost delivery plan that is feasible for all demand realizations within a prespecied uncertainty set. A crucial step in our heuristic is to verify the robust feasibility of a candidate route. For generic uncertainty sets, this step requires the solution of a convex optimization problem, which becomes computationally prohibitive for large instances. We present two classes of uncertainty sets for which route feasibility can be established much more eciently. While we discuss our implementation in the context of the AMP framework, our techniques readily extend to other meta heuristics. Computational studies on standard literature benchmarks with up to 483 customers and 38 vehicles demonstrate that the proposed approach is able to quickly provide high quality solutions. In the process, we obtain new best solutions for a total of 123 benchmark instances.
时 间：2014年5月14日 14:00-16:00