Stabilizing reinforcement learning control: A modular framework for optimizing over all stable behavior

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lawrence, Nathan P., Loewen, Philip D., Wang, Shuyuan, Forbes, Michael G., Gopaluni, R. Bhushan
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909145826852864
author Lawrence, Nathan P.
Loewen, Philip D.
Wang, Shuyuan
Forbes, Michael G.
Gopaluni, R. Bhushan
author_facet Lawrence, Nathan P.
Loewen, Philip D.
Wang, Shuyuan
Forbes, Michael G.
Gopaluni, R. Bhushan
contents We propose a framework for the design of feedback controllers that combines the optimization-driven and model-free advantages of deep reinforcement learning with the stability guarantees provided by using the Youla-Kucera parameterization to define the search domain. Recent advances in behavioral systems allow us to construct a data-driven internal model; this enables an alternative realization of the Youla-Kucera parameterization based entirely on input-output exploration data. Perhaps of independent interest, we formulate and analyze the stability of such data-driven models in the presence of noise. The Youla-Kucera approach requires a stable "parameter" for controller design. For the training of reinforcement learning agents, the set of all stable linear operators is given explicitly through a matrix factorization approach. Moreover, a nonlinear extension is given using a neural network to express a parameterized set of stable operators, which enables seamless integration with standard deep learning libraries. Finally, we show how these ideas can also be applied to tune fixed-structure controllers.
format Preprint
id arxiv_https___arxiv_org_abs_2310_14098
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Stabilizing reinforcement learning control: A modular framework for optimizing over all stable behavior
Lawrence, Nathan P.
Loewen, Philip D.
Wang, Shuyuan
Forbes, Michael G.
Gopaluni, R. Bhushan
Machine Learning
Artificial Intelligence
Systems and Control
Optimization and Control
We propose a framework for the design of feedback controllers that combines the optimization-driven and model-free advantages of deep reinforcement learning with the stability guarantees provided by using the Youla-Kucera parameterization to define the search domain. Recent advances in behavioral systems allow us to construct a data-driven internal model; this enables an alternative realization of the Youla-Kucera parameterization based entirely on input-output exploration data. Perhaps of independent interest, we formulate and analyze the stability of such data-driven models in the presence of noise. The Youla-Kucera approach requires a stable "parameter" for controller design. For the training of reinforcement learning agents, the set of all stable linear operators is given explicitly through a matrix factorization approach. Moreover, a nonlinear extension is given using a neural network to express a parameterized set of stable operators, which enables seamless integration with standard deep learning libraries. Finally, we show how these ideas can also be applied to tune fixed-structure controllers.
title Stabilizing reinforcement learning control: A modular framework for optimizing over all stable behavior
topic Machine Learning
Artificial Intelligence
Systems and Control
Optimization and Control
url https://arxiv.org/abs/2310.14098