February 9, 2023 Thursday
In this talk, I will describe how to design efficient learning and control algorithms for energy systems by combining physics and operation data. Specially, I will talk about two methods that we developed for this problem: one is reinforcement learning with stability guarantee, by bridging policy learning in Reinforcement Learning with Lyapunov stability theory in control; the other one is robust online control, that combines convex body chasing for model identification and a robust control oracle to obtain a finite mistake guarantee.
Dr. Yuanyuan SHI
Assistant Professor, the Department of Electrical and Computer Engineering, University of California San Diego
Yuanyuan Shi is an Assistant Professor in the Department of Electrical and Computer Engineering (ECE) at the University of California San Diego. She received her Ph.D. and M.Sc. in Electrical and Computer Engineering and M.Sc. in Statistics from University of Washington in 2020 and 2019 respectively. She was a postdoc fellow in the Computing and Mathematical Sciences Department at Caltech from 2020-2021, before she started as a faculty at UCSD since July 2021. Yuanyuan’s research lies in machine learning and control, with applications in sustainable energy and power systems. Her research has been interdisciplinary in nature, integrating various mathematical, computational, engineering and some economics tools, that can be deployed in the electricity grid, energy market, smart buildings and beyond. She is a recipient of multiple awards, including the Rising Star award in EECS by MIT in 2018, the Scientific Achievement Award from the University of Washington Clean Energy Institute in 2020, and ACM e-Energy best paper finalist in 2022.