TY - GEN
T1 - Online Policy Gradient for Model Free Learning of Linear Quadratic Regulators with √T Regret
AU - Cassel, Asaf
AU - Koren, Tomer
N1 - Publisher Copyright:
Copyright © 2021 by the author(s)
PY - 2021
Y1 - 2021
N2 - We consider the task of learning to control a linear dynamical system under fixed quadratic costs, known as the Linear Quadratic Regulator (LQR) problem. While model-free approaches are often favorable in practice, thus far only model-based methods, which rely on costly system identification, have been shown to achieve regret that scales with the optimal dependence on the time horizon T. We present the first model-free algorithm that achieves similar regret guarantees. Our method relies on an efficient policy gradient scheme, and a novel and tighter analysis of the cost of exploration in policy space in this setting.
AB - We consider the task of learning to control a linear dynamical system under fixed quadratic costs, known as the Linear Quadratic Regulator (LQR) problem. While model-free approaches are often favorable in practice, thus far only model-based methods, which rely on costly system identification, have been shown to achieve regret that scales with the optimal dependence on the time horizon T. We present the first model-free algorithm that achieves similar regret guarantees. Our method relies on an efficient policy gradient scheme, and a novel and tighter analysis of the cost of exploration in policy space in this setting.
UR - http://www.scopus.com/inward/record.url?scp=85125033220&partnerID=8YFLogxK
M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???
AN - SCOPUS:85125033220
T3 - Proceedings of Machine Learning Research
SP - 1304
EP - 1313
BT - Proceedings of the 38th International Conference on Machine Learning, ICML 2021
PB - ML Research Press
T2 - 38th International Conference on Machine Learning, ICML 2021
Y2 - 18 July 2021 through 24 July 2021
ER -