题　目：Dynamic Assortment in Service Systems

嘉　宾：郭元 助理教授 乔治华盛顿大学

主持人：许欢 教授 上海交通大学安泰经济与管理学院

时　间：2025年6月5日（周四）14:00-15:30

地　点：安泰楼A511室

内容简介：

This paper studies dynamic assortment optimization in service systems where customer purchasing decisions depend on real-time system congestion, as observed in food delivery and 3D printing services. We model the problem as a continuous-time control process in an M/G/1 queueing system with impatient customers whose choices follow a general discrete choice model. Each product is characterized by its margin and service time, and the firm’s objective is to dynamically select assortments to maximize long-run average revenue. We establish the existence and concavity of the value function and derive optimal state-dependent policies under general choice behavior. Under the Multinomial Logit (MNL) model, we show that the optimal assortment follows an adjusted-margin ordering, where product rankings shift at most once as congestion increases. To investigate the structure of optimal assortment across congestion levels, we analytically and numerically characterize conditions under which the optimal assortment expands, contracts, or changes non-monotonically with congestion. Furthermore, we extend the model to incorporate dynamic pricing and identify an optimal policy that maintains a full assortment while adjusting prices to satisfy a constant adjusted-margin condition. Finally, our numerical experiments demonstrate substantial revenue improvements of dynamic policies over static benchmarks, underscoring the value of state-dependent assortment control in congested service environments. (co-authored with Chen-An Lin from Purdue University)

演讲人简介：

Yuan Guo is an Assistant Professor of Decision Sciences at the George Washington University School of Business. She earned her Ph.D. in Business Administration, with a concentration in Operations Management, from the Fuqua School of Business at Duke University in 2023. Her research develops analytical models to study revenue management, customer choice behavior, and dynamic assortment decisions, with a particular interest in operational trade-offs arising in new retail and service environments. Her work seeks to generate structural insights with practical implications for platform design and retail strategy.

欢迎广大师生参加！