报告题目：Assortment Optimization under Heteroscedastic Data
报告人：Xiaobo Li（李晓波）新加坡国立大学
报告时间：2022年4月27日(周三) 16:00-17:30
报告地点：中国科学院大学中关村校区教学楼S406 ；
Zoom 会议号：810 0362 5928
入会密码：024230
内容摘要
We study assortment problems under the marginal exponential model (MEM), which is an extension of the multinomial logit (MNL) model that can capture heteroscedasticity in the data. We first show that when the variance of the outside option is the largest, one of the profit-nested assortments is optimal. This result generalizes the well-known result for the MNL assortment optimization problem. Next, we show that the product assortment problem under MEM is NP-hard, but the best profit-nested assortment provides a good approximation to the optimal assortment. Furthermore, we improve existing MEM parameter estimation methods. Our numerical studies show that using MEM to capture choice behavior in assortment optimization leads to competitive results compared to other choice models that are also designed to capture heteroscedasticity.
主讲人简介
Xiaobo Li is an assistant professor in the Department of Industrial Systems Engineering and Management at the National University of Singapore. He received my Ph.D. in Industrial Engineering from the University of Minnesota in 2018. His research mainly focuses on robust optimization and demand learning, with applications in revenue management, data-driven decision making, and sharing economy.
