Optimal Robust Policy for Big Data Newsvendor 2022-05-12
Subject：Optimal Robust Policy for Big Data Newsvendor
Guest：Gao Rui, Assistant Professor, University of Texas at Austin
Host：Cao Yufeng, Assistant Professor, ACEM-SJTU
Time：Wednesday, May 4, 2022, 10:00-11:30
Venue: Tencent Meeting
（Please send email to firstname.lastname@example.org by 18:00 May. 3th for meeting number and password.)
We study policy optimization for the feature-based newsvendor, which seeks an end-to-end policy that renders an explicit mapping from features to ordering decisions. Unlike existing works that restrict the policies to some parametric class which may suffer from sub-optimality (such as affine class) or lack of interpretability (such as neural networks), we aim to optimize over all measurable functions of features. In this case, the classical empirical risk minimization yields a policy that are not well-defined on unseen features. To avoid such degeneracy, we consider a distributionally robust framework. This leads to an adjustable robust optimization, whose optimal solutions are notoriously difficult to obtain except for a few notable cases. Perhaps surprisingly, we identify a new class of policies that are proven to be exactly optimal and can be computed efficiently. The optimal robust policy is obtained by extending an optimal robust in-sample policy to unobserved features in a particular way and can be interpreted as a Lipschitz regularized critical fractile of the empirical conditional demand distribution. We compare our method with several benchmarks using real data and demonstrate its superior empirical performance. This is a joint work with Luhao Zhang and Jincheng Yang.
Rui Gao is an Assistant Professor in the Department of Information, Risk, and Operations Management at the McCombs School of Business at the University of Texas at Austin. His main research studies data-driven decision-making under uncertainty and prescriptive data analytics. In particular, he develops robust and computationally efficient methodology with strong performance guarantees for learning and decision-making problems involving offline/online/contextual/observational data, arising in the area of machine learning, risk management and operations management. His research has been recognized with several INFORMS paper competition awards, including Winner in INFORMS Junior Faculty Interest Group Paper Competition (2020), Winner in INFORMS Data Mining Best Paper Award (2017), Runner-up in INFORMS Computing Society Student Paper Award (2017), and Finalist in INFORMS George Nicholson Student Paper Competition (2016). He received a Ph.D. in Operations Research from Georgia Institute of Technology in 2018, and a B.Sc. in Mathematics and Applied Mathematics from Xi'an Jiaotong University in 2013.