运作管理学术论坛（02）：龙卓瑜

时间：2025-07-11

运作管理学术论坛（02）：龙卓瑜

报告人：龙卓瑜教授

报告题目： Optimal Open-Loop Policy in Distributionally Robust Inventory Control with Uncertain Lead Times

报告时间：2025年7月12日（周六）上午9:30-10:20

报告地点：劝学楼425室

主办单位：运作管理与优化决策研究团队

报告人简介：Daniel Zhuoyu Long is a Professor in the Department of Systems Engineering and Engineering Management at The Chinese University of Hong Kong. Previously, he received his Bachelor's degree from Tsinghua University in 2005, Master's degree from the Chinese Academy of Sciences in 2008, and Ph.D. from the National University of Singapore Business School in 2013, joining CUHK in the same year. His research primarily focuses on distributed robust optimization theory and its applications to various operations management problems, such as logistics and supply chain management, project management, healthcare operations management, and revenue management. His work was elected as a finalist for the 2021 Best OM Paper in OR, and received the 2022 CSAMSE Best Paper Award (First Prize) and 2024 CSAMSE Best Paper Award (Second Prize). He currently serves as an Associate Editor for the MSOM Journal.

报告摘要：In this paper, we consider a multi-periods inventory control problem with uncertain lead times, which present significant challenges in both practice and academic. We consider the distributionally robust optimization (DRO) setting where only the marginal distribution is known and we aim at minimizing the worst-case expected cost. We focus on open-loop policies, which have attracted substantial attention recently due to their asymptotic optimality and appealing numerical performance. By showing the supermodularity property of the cost function, we develop an efficient algorithm that identifies the worst-case distribution and reformulate the DRO inventory problem into a Mixed Integer Linear Programming (MILP) problem. Further, we identify necessary and sufficient conditions under which there are no order-crossovers. When such conditions are satisfied, our problem can be further simplified into a shortest path problem on a directed acyclic network, and thus can be solved within polynomial time. Numerical examples are provided to demonstrate the practical effectiveness of our approach. In this paper, we consider a multi-periods inventory control problem with uncertain lead times, which present significant challenges in both practice and academic. We consider the distributionally robust optimization (DRO) setting where only the marginal distribution is known and we aim at minimizing the worst-case expected cost. We focus on open-loop policies, which have attracted substantial attention recently due to their asymptotic optimality and appealing numerical performance. By showing the supermodularity property of the cost function, we develop an efficient algorithm that identifies the worst-case distribution and reformulate the DRO inventory problem into a Mixed Integer Linear Programming (MILP) problem. Further, we identify necessary and sufficient conditions under which there are no order-crossovers. When such conditions are satisfied, our problem can be further simplified into a shortest path problem on a directed acyclic network, and thus can be solved within polynomial time. Numerical examples are provided to demonstrate the practical effectiveness of our approach.

撰稿： 赵永丽     审核：吴志樵  印明鹤       单位：管理科学与工程学院