Wasserstein distributionally robust control and optimization for ambiguous stochastic systems

Insoon Yang

Associate Professor,
Seoul National University (SNU)

Seminar Information

Seminar Series
Dynamic Systems & Controls

Seminar Date - Time
February 11, 2022, 3:00 pm
-
4:00

Seminar Location
Seminar Recording Available: Please contact seminar coordinator, Lusia Veksler at (lveksler@eng.ucsd.edu)


Abstract

Standard stochastic control methods assume that the probability distribution of uncertain variables is available. Unfortunately, in practice, obtaining accurate distribution information is a challenging task. To resolve this issue, we investigate the problem of designing a controller that is robust against errors in the empirical distribution obtained from data. The proposed framework using the Wasserstein metric has several salient features, including an out-of-sample performance guarantee, an explicit solution in the LQ setting, and a theoretical connection to the classical H_infty method. We further discuss its MPC variant and application to motion planning and control in risky environments.  

Speaker Bio

Prof. Insoon Yang is an Associate Professor of ECE at Seoul National University (SNU). He received B.S. degrees in Mathematics and in Mechanical Engineering (summa cum laude) from SNU in 2009; and a Ph.D. in EECS from UC Berkeley in 2015. He was an Assistant Professor of ECE at University of Southern California from 2016 to 2018, and a Postdoctoral Associate with the Laboratory for Information and Decision Systems at Massachusetts Institute of Technology from 2015 to 2016. His research interests are in stochastic control and optimization, and reinforcement learning. He is a recipient of the 2015 Eli Jury Award and a finalist for the Best Student Paper Award at the 55th IEEE Conference on Decision and Control 2016.