Stability of Stochastic Approximations with `Controlled Markov' Noise and Temporal Difference Learning
This work addresses stability issues in reinforcement learning algorithms for researchers and practitioners, offering a more general theoretical framework than existing analyses, though it is incremental in extending prior stochastic approximation theory.
The paper tackles the problem of ensuring stability and convergence for stochastic approximation algorithms driven by controlled Markov processes, which are crucial in reinforcement learning, by presenting verifiable sufficient conditions that generalize to continuous state spaces and non-ergodic processes, and applies this theory to analyze generalized TD(0) and a TD formulation for supervised learning.
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.