报告题目：Robust Stationary Base-Stock Policies for Inventory Management with Lost Sales and Constant Lead Times

报告人：The University of Sydney Business School, Prof. Zhaolin (Erick) Li

报告时间：2025年6月3日 上午9:30 –11:00

报告地点：综合楼710B会议室

邀请人：人工智能系 黄敏教授

报告内容摘要：

We study multiperiod inventory models with lost sales and constant lead times. Nonnegative demands are independently and identically distributed (i.i.d.), and the available information on demand distribution contains only the mean and variance. When applying the canonical prime-dual method to solve the model, demand independence makes the problem intractable. To resolve the technical hurdle, we propose a zero-sum game model in which the firm chooses an ordering policy to maximize the long-run average profit while Nature (which is the firm’s opponent) chooses a distribution to minimize the firm’s profit in steady state. For the lost-sales model, we characterize the steady-state equilibrium and derive each player’s strategy in closed forms. We show that Nature’s equilibrium strategy is an i.i.d. two-point distribution while the firm can implement several equivalent strategies to attain the same optimal outcome. For instance, base-stock, capped base-stock, projected inventory level, and fixed non-stock probability policies yield the same outcome when playing against Nature’s equilibrium strategy. As our robust base-stock policy involves a distribution-free parameter and facilitates distribution-free implementation, it serves as a valuable reference for scholars and practitioners. We also perform numerical experiments to evaluate the performance of various policies. When the underline distribution is unknown and the firm uses a fitted distribution based on moments, our robust base-stock policy delivers superior performance than all the other candidates, effectively mitigating the consequence resulting from mis-specifying the demand distribution.

报告人简介：

Dr. Zhaolin (Erick) Li received his Ph.D. in Business Administration from The Pennsylvania State University, a Master of Commerce in Accounting from The University of New South Wales, and a Bachelor of Engineering in Materials Science and Industrial Engineering from Shanghai Jiao Tong University. He has been a faculty member at The University of Sydney Business School since January 2009. Prior to his academic career in Sydney, Dr. Li worked at Ernst & Young LLP and City University of Hong Kong.Dr. Li has published seven articles in Production and Operations Management, establishing him as the most prolific Australian scholar in this flagship journal. He has also had two articles published or recently accepted by Management Science. Both journals are highly regarded, appearing in the Financial Times, Bloomberg MBA, and UT Dallas Business School Research Rankings. According to studies by Babbar et al. (2017, 2018) in the International Journal of Production Economics, Dr. Li is ranked among Asia's top 25 scholars in Operations Management and Supply Chain Management, respectively.His recent research applies mean-variance analysis to develop probabilistic bounds and create robust solutions for various business challenges. He has also pioneered a joint optimization method to efficiently solve a class of moment-based robust optimization models, achieving novel closed-form solutions that are otherwise unattainable through traditional two-stage methods. His newest article, accepted by Management Science, applies mean-variance analysis to moral hazard, demonstrating that ambiguity can fundamentally change the characteristics of the optimal contract between farmers and landowners.