Simple and optimal methods for stochastic variational inequalities, II: Markovian noise and policy evaluation in reinforcement learning

TL;DR

This paper introduces simple, optimal stochastic algorithms for variational inequalities with Markovian noise, significantly improving convergence rates and enabling effective parallelization, especially for policy evaluation in reinforcement learning.

Contribution

It develops new TD learning algorithms with non-asymptotic analysis, improving convergence rates and parallel implementation over prior methods in stochastic variational inequalities.

Findings

Improved analysis of standard TD algorithm with parallel benefits

Introduction of conditional TD (CTD) with reduced bias and better complexity

Development of fast TD (FTD) with optimal convergence rate

Abstract

The focus of this paper is on stochastic variational inequalities (VI) under Markovian noise. A prominent application of our algorithmic developments is the stochastic policy evaluation problem in reinforcement learning. Prior investigations in the literature focused on temporal difference (TD) learning by employing nonsmooth finite time analysis motivated by stochastic subgradient descent leading to certain limitations. These encompass the requirement of analyzing a modified TD algorithm that involves projection to an a-priori defined Euclidean ball, achieving a non-optimal convergence rate and no clear way of deriving the beneficial effects of parallel implementation. Our approach remedies these shortcomings in the broader context of stochastic VIs and in particular when it comes to stochastic policy evaluation. We developed a variety of simple TD learning type algorithms motivated by its original version that maintain its simplicity, while offering distinct advantages from a non-asymptotic analysis point of view. We first provide an improved analysis of the standard TD algorithm that can benefit from parallel implementation. Then we present versions of a conditional TD algorithm (CTD), that involves periodic updates of the stochastic iterates, which reduce the bias and therefore exhibit improved iteration complexity. This brings us to the fast TD (FTD) algorithm which combines elements of CTD and the stochastic operator extrapolation method of the companion paper. For a novel index resetting policy FTD exhibits the best known convergence rate. We also devised a robust version of the algorithm that is particularly suitable for discounting factors close to 1.