Optimal Control Inspired Q-Learning for Switched Linear Systems

Chen，Hua1; Zheng，Linfang1; Zhang，Wei2

10.23919/ACC45564.2020.9147818

2020-07-01

0743-1619

978-1-5386-8267-8

Proceedings of the American Control Conference

2020-July

4003-4010

1-3 July 2020

Denver, CO, USA

This paper studies Q-learning for quadratic regulation problem of switched linear systems. Inspired by the analytical results from classical model-based optimal control, a structured Q-learning algorithm is developed. The proposed algorithm consists of a carefully designed parametric approximator that respects the analytical structure of the exact Q-function and an associated parameter training algorithm. Based on a geometric insight gained from the analysis of the exact Q-function structure, training of approximation parameters is formulated as a matrix identification problem. Probabilistic guarantee on successful identification of all matrices using the proposed algorithm is rigorously proved under moderate conditions. Several numerical studies are conducted to demonstrate the effectiveness of the overall proposed Q-learning algorithm.

Approximation algorithms Optimal control Switches Heuristic algorithms Training System dynamics Learning (artificial intelligence)

第一

英语

EI

20203409087210

Parameter estimation ; Learning algorithms ; Linear systems

Machine Learning:723.4.2 ; Algebra:921.1 ; Systems Science:961

2-s2.0-85089592193

Scopus

1.Southern University of Science and Technology (SUSTech),Department of Mechanical and Energy Engineering,Shenzhen,China
2.Ohio State University,Department of Electrical and Computer Engineering,Columbus,United States

Chen，Hua,Zheng，Linfang,Zhang，Wei. Optimal Control Inspired Q-Learning for Switched Linear Systems[C],2020:4003-4010.