Asynchronous Parallel Policy Gradient Methods for the Linear Quadratic Regulator

TL;DR

This paper investigates the convergence and efficiency of asynchronous parallel policy gradient methods applied to the linear quadratic regulator, demonstrating linear speedup through theoretical analysis and simulations.

Contribution

It provides the first rigorous convergence analysis of asynchronous parallel policy gradient methods for LQR, establishing their linear speedup property.

Findings

AZOPG achieves linear convergence rate.

Parallel workers significantly speed up policy learning.

Theoretical and simulation results confirm linear speedup.

Abstract

Learning policies in an asynchronous parallel way is essential to the numerous successes of RL for solving large-scale problems. However, their convergence performance is still not rigorously evaluated. To this end, we adopt the asynchronous parallel zero-order policy gradient (AZOPG) method to solve the continuous-time linear quadratic regulation problem. Specifically, as in the celebrated A3C algorithm, there are multiple parallel workers to asynchronously estimate PGs which are then sent to a central master for policy updates. Via quantifying its convergence rate of policy iterations, we show the linear speedup property of the AZOPG, both in theory and simulation, which clearly reveals the advantages of using parallel workers for learning policies.