Accelerate Reinforcement Learning with PID Controllers in the Pendulum Simulations

Abstract

We propose a Proportional Integral Derivative (PID) controller-based coaching scheme to expedite reinforcement learning (RL).