Testing Stationarity and Change Point Detection in Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Mengbing, Shi, Chengchun, Wu, Zhenke, Fryzlewicz, Piotr
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915089637965824
author Li, Mengbing
Shi, Chengchun
Wu, Zhenke
Fryzlewicz, Piotr
author_facet Li, Mengbing
Shi, Chengchun
Wu, Zhenke
Fryzlewicz, Piotr
contents We consider offline reinforcement learning (RL) methods in possibly nonstationary environments. Many existing RL algorithms in the literature rely on the stationarity assumption that requires the system transition and the reward function to be constant over time. However, the stationarity assumption is restrictive in practice and is likely to be violated in a number of applications, including traffic signal control, robotics and mobile health. In this paper, we develop a consistent procedure to test the nonstationarity of the optimal Q-function based on pre-collected historical data, without additional online data collection. Based on the proposed test, we further develop a sequential change point detection method that can be naturally coupled with existing state-of-the-art RL methods for policy optimization in nonstationary environments. The usefulness of our method is illustrated by theoretical results, simulation studies, and a real data example from the 2018 Intern Health Study. A Python implementation of the proposed procedure is available at https://github.com/limengbinggz/CUSUM-RL.
format Preprint
id arxiv_https___arxiv_org_abs_2203_01707
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Testing Stationarity and Change Point Detection in Reinforcement Learning
Li, Mengbing
Shi, Chengchun
Wu, Zhenke
Fryzlewicz, Piotr
Machine Learning
We consider offline reinforcement learning (RL) methods in possibly nonstationary environments. Many existing RL algorithms in the literature rely on the stationarity assumption that requires the system transition and the reward function to be constant over time. However, the stationarity assumption is restrictive in practice and is likely to be violated in a number of applications, including traffic signal control, robotics and mobile health. In this paper, we develop a consistent procedure to test the nonstationarity of the optimal Q-function based on pre-collected historical data, without additional online data collection. Based on the proposed test, we further develop a sequential change point detection method that can be naturally coupled with existing state-of-the-art RL methods for policy optimization in nonstationary environments. The usefulness of our method is illustrated by theoretical results, simulation studies, and a real data example from the 2018 Intern Health Study. A Python implementation of the proposed procedure is available at https://github.com/limengbinggz/CUSUM-RL.
title Testing Stationarity and Change Point Detection in Reinforcement Learning
topic Machine Learning
url https://arxiv.org/abs/2203.01707