Comparison of Reinforcement Learning algorithms applied to the Cart Pole problem

TL;DR

This paper compares various reinforcement learning algorithms for controlling the cartpole system without prior knowledge, introduces a novel RL-swing-up integration, and evaluates their performance against traditional LQR control.

Contribution

It provides a comparative analysis of RL algorithms for cartpole control and proposes a new method combining RL with swing-up controllers.

Findings

RL algorithms perform comparably to LQR in control tasks

The proposed RL-swing-up integration improves control efficiency

Different RL methods show varying strengths in nonlinear control

Abstract

Designing optimal controllers continues to be challenging as systems are becoming complex and are inherently nonlinear. The principal advantage of reinforcement learning (RL) is its ability to learn from the interaction with the environment and provide optimal control strategy. In this paper, RL is explored in the context of control of the benchmark cartpole dynamical system with no prior knowledge of the dynamics. RL algorithms such as temporal-difference, policy gradient actor-critic, and value function approximation are compared in this context with the standard LQR solution. Further, we propose a novel approach to integrate RL and swing-up controllers.