----------------------------------------------------------------------------------------------------

    Department of Systems Engineering and Engineering Management

                       The Chinese University of Hong Kong

----------------------------------------------------------------------------------------------------

Date: Tuesday, May 23, 4:30 pm – 6:00 pm

Venue: ERB 513, The Chinese University of Hong Kong

Title: Data-driven Piecewise Affine Decision Rule Methods for Stochastic Optimization with Covariate Information

Speaker: Prof. Junyi Liu, Tsinghua University

Abstract:

In this talk, we focus on a class of stochastic programming problems that minimize the conditionally expected cost given a new covariate observation. Since in many real-world OR applications, both the conditional probability distribution and scenarios under the given covariate observation are not available to the decision-maker, the classical SAA or SA approaches are not applicable for solving this class of problems. To deal with this challenge, we propose a data-driven piecewise affine decision rule (PADR) method based on historical data pairs for solving the contextual decision-making problem. We provide the non-asymptotic consistency of the data-driven PADR-based method by quantifying the approximation accuracy of piecewise affine functions. To solve the PADR-based empirical risk minimization problem with a coupled nonconvex and nondifferentiable structure, we develop an enhanced stochastic majorization minimization algorithm and provide the nonasymptotic convergence rate analysis in terms of directional stationarity.  Numerical results for both convex and nonconvex problems with various nonlinear generating models indicates the superiority of the proposed data-driven method compared with the state-of-the-art data-driven methods.

Biography:

Junyi Liu is currently an (untenured) associate professor at Department of Industrial Engineering, Tsinghua University. She obtained Ph.D. degree in 2019 at Department of Industrial and Systems Engineering at University of Southern California. She worked as a post-doc associate with Prof. Jong-Shi Pang before joining Tsinghua University in April 2021. Dr. Liu’s research interests lie in the broad area of stochastic programming, and in particular the intersections of stochastic programming with statistics and data science. 

Everyone is welcome to attend the talk!

SEEM-5202 Website: http://seminar.se.cuhk.edu.hk

Email: seem5202@se.cuhk.edu.hk