题　目：Global Optimality for Nonconvex Optimization in Operations, Learning, and Causality

嘉　宾：Yifan Hu, Assistant Professor, Rutgers University

主持人：何斯迈  教授 上海交通大学安泰经济与管理学院

时　间：2025年7月9日（周三）14:00-15:30

地　点：上海交通大学 徐汇校区安泰浩然楼308室

内容简介:

Nonconvex optimization arises in diverse domains such as operations, machine learning, and causality. While nonconvexity poses significant challenges in achieving global optimality, we demonstrate how these barriers can be overcome in several critical applications. In this talk, we first discuss quantity-based network revenue management problems. Leveraging hidden convexity, i.e., a convex reformulation through a (possibly implicit) variable transformation, we propose methods that achieve global optimality without requiring explicit knowledge of the transformation. Moreover, the complexities of these methods match the lower bounds for stochastic convex optimization, implying that they are optimal. Extensive numerical experiments on airline revenue management showcase the superiority of our approach, achieving higher revenue and lower computational costs compared to state-of-the-art bid-price control policies. Beyond hidden convexity, we further explore benign nonconvexity in finite-horizon reinforcement learning and causal discovery problems and devise efficient algorithms. Leveraging structural information from applications, we highlight principles that can drive scalable algorithmic design with practical impacts.

演讲人简介：

Yifan Hu is an incoming assistant professor in Rutgers University statistics department. He is a postdoc researcher jointly advised by Prof. Daniel Kuhn from EPFL and Prof. Andreas Krause from ETH Zurich in Switzerland. Prior to that, he obtained PhD in Operations Research from the University of Illinois at Urbana-Champaign, jointly advised by Prof. Xin Chen and Prof. Niao He. His research interests lie in data-driven decision-making, focusing on understanding the foundations of optimization and learning, as well as developing easy-to-implement algorithms for operations and causal inference applications.

欢迎广大师生参加!