Stability of Stochastic Approximations with `Controlled Markov' Noise and Temporal Difference Learning

TL;DR

This paper establishes new stability and convergence conditions for stochastic approximation algorithms driven by controlled Markov processes, extending analysis to continuous state spaces and non-ergodic processes, with applications to reinforcement learning and supervised learning.

Contribution

It provides verifiable conditions for stability and convergence of SAs with controlled Markov noise, including continuous states and non-ergodic processes, and applies these results to TD learning and forecasting.

Findings

Stability conditions ensure almost sure boundedness of SAs.

Analysis applies to continuous state spaces in RL.

Generalized TD(0) and supervised learning formulations are analyzed.

Abstract

We are interested in understanding stability (almost sure boundedness) of stochastic approximation algorithms (SAs) driven by a `controlled Markov' process. Analyzing this class of algorithms is important, since many reinforcement learning (RL) algorithms can be cast as SAs driven by a `controlled Markov' process. In this paper, we present easily verifiable sufficient conditions for stability and convergence of SAs driven by a `controlled Markov' process. Many RL applications involve continuous state spaces. While our analysis readily ensures stability for such continuous state applications, traditional analyses do not. As compared to literature, our analysis presents a two-fold generalization (a) the Markov process may evolve in a continuous state space and (b) the process need not be ergodic under any given stationary policy. Temporal difference learning (TD) is an important policy evaluation method in reinforcement learning. The theory developed herein, is used to analyze generalized $TD(0)$, an important variant of TD. Our theory is also used to analyze a TD formulation of supervised learning for forecasting problems.