2021-02-02 | Yize Chen:Engineering Data-Driven Tools for Cyber-Physical Systems



With the growing amount of sensing data and increasing level of system complexity, there are emerging opportunities and challenges in designing reliable operation schemes for various tasks in cyber-physical systems such as robotics manipulation, building energy management and optimal power flow. This calls for a rethinking of control and optimization theory from the standpoint of data-driven approach.

 In the first part of the talk, I will focus on learning to solve optimization considering the data uncertainties, where the rich information of the underlying optimization problem along with the convex optimization theory provide design modules for an efficient learner to output the solutions satisfying all engineering constraints.

In the second part, I will touch on the decision-making problem where a model is not known a priori, and investigate how specifically designed neural networks can provide optimal control solutions. This result represents the design principles for learning an accurate and computationally tractable controller. Together, these examples highlight the important role of learning and closed-loop control in the design of sustainable and reliable information and physical systems.




Yize Chen is a Ph.D. candidate from the Department of Electrical and Computer Engineering at University of Washington, and he received his B.S. with honors from Chu Kochen College at Zhejiang University in 2016. His research lies in control, learning and optimization of cyber-physical systems with performance guarantees. He is the recipient of several awards, including the 2019 ACM e-Energy Best Paper RunnerUp. He held internship and research positions at Harvard Medical School, Los Alamos National Laboratory, and Microsoft Research. His research is supported by Keith and Nancy Rattie Fellowship and University of Washington Clean Energy Institute Fellowship. Homepage: https://blogs.uw.edu/yizechen/