An Online Prediction Algorithm for Reinforcement Learning with Linear Function Approximation using Cross Entropy Method

TL;DR

This paper introduces two stable online algorithms for reinforcement learning prediction tasks using linear function approximation, leveraging the cross entropy method for optimization, with proven convergence and strong empirical performance.

Contribution

The paper presents novel stable online algorithms employing the cross entropy method for reinforcement learning prediction with linear function approximation, including convergence proofs.

Findings

Algorithms demonstrate good accuracy and stability.

Achieve computational efficiency on benchmark problems.

Proven convergence using the ODE method.

Abstract

In this paper, we provide two new stable online algorithms for the problem of prediction in reinforcement learning, \emph{i.e.}, estimating the value function of a model-free Markov reward process using the linear function approximation architecture and with memory and computation costs scaling quadratically in the size of the feature set. The algorithms employ the multi-timescale stochastic approximation variant of the very popular cross entropy (CE) optimization method which is a model based search method to find the global optimum of a real-valued function. A proof of convergence of the algorithms using the ODE method is provided. We supplement our theoretical results with experimental comparisons. The algorithms achieve good performance fairly consistently on many RL benchmark problems with regards to computational efficiency, accuracy and stability.