When
Thursday, February 20, 2025 - 2:00 p.m.(MST)
Jianqiang Cheng
Associate Professor of Systems & Industrial Engineering
University of Arizona
"Dimensionality Reduction Techniques for Moment-based Distributionally Robust Optimization"
ENGR 301
Abstract: This talk explores recent advancements in dimensionality reduction techniques for moment-based Distributionally Robust Optimization (DRO). We first examine traditional PCA-based methods and then introduce an optimized dimensionality reduction (ODR) approach that integrates subsequent optimization steps, leading to improved computational efficiency and solution accuracy. Specifically, it enables two outer and one inner approximations of the original problem, yielding two lower bounds and one upper bound, respectively. As these approximations are nonconvex low-dimensional semidefinite programs, we apply Alternating Direction Method of Multipliers algorithms to solve them efficiently. Numerical results show significant advantages of our approach on the computational time and solution quality over existing benchmark approaches.
Bio: Jianqiang Cheng is an associate professor in the Department of Systems & Industrial Engineering at the University of Arizona, and a member of the Applied Mathematics and Statistics & Data Science GIDPs. He earned his PhD in informatique at the PARIS-SACLAY University, France. His research focuses on developing computationally efficient techniques to solve difficult optimization problems under uncertainty. He is the recipient of a Bisgrove Early Career Scholar Award (Arizona Science Foundation) and a NSF CAREER Award. His research has been supported by National Science Foundation, Arizona Science Foundation, and Office of Naval Research, among others.