報告題目：Asymptotically Optimal Replenishment policies for Perishable Inventory Systems
時間：2023年12月6日下午14：30
地點：中關村校區主樓418會議
報告人：卜金枝
報告人簡介：
Jinzhi Bu is an Assistant Professor in the Department of Logistics and Maritime Studies at the Hong Kong Polytechnic University. Her research interests include stochastic modeling and optimization, statistical and machine learning, data-driven decision making, and their applications to supply chain management and revenue management. Her research has been published in Operations Research and Management Science. Prior to joining PolyU, she was a postdoctoral associate at Massachusetts Institute of Technology from 2019 to 2021. She obtained her Ph.D. degree from the Chinese University of Hong Kong in 2019 and B.S. degree from Nanjing University in 2015.
報告內容簡介：
We consider a fundamental class of periodic-review inventory systems for perishable products with a fixed product lifetime. The problem setting is quite general: the inventory replenishment lead time can be either zero or positive, the demands are satisfied by on-hand inventories of different ages according to a general issuance policy (e.g., first-in-first-out (FIFO) or last-in-first-out (LIFO)), and unsatisfied demands can be either backlogged or lost. The objective is to minimize the long-run average holding, penalty, and outdating costs. The optimal replenishment policy for this class of systems is very complex and computing the optimal cost is intractable due to the curse of dimensionality. We propose two classes of simple policies and develop the asymptotic-optimality results in various parameter regimes. For the systems with zero lead time, we show that the class of base-stock policies is asymptotically optimal with large unit penalty costs under the FIFO issuance policy for general demand distributions and under a general issuance policy for unbounded demand distributions. For the systems with positive lead times, we show that the class of projected inventory level policies is asymptotically optimal with large unit penalty costs for a large class of demand distributions. The asymptotic-optimality results in other parameter regimes are also developed for both classes of policies. Our numerical results demonstrate the near-optimal performances of both policies in these systems. This talk is based on joint work with Xiting Gong, Xiuli Chao and Huanyu Yin.
（承辦：管理科學與物流系、科研與學術交流中心）
