Path Integral Policy Improvement with Covariance Matrix Adaptation

International audienceThere has been a recent focus in reinforcement learning on addressing continuous state and action problems by optimizing parameterized policies. PI2 is a recent example of this approach. It combines a derivation from first principles of stochastic optimal control with tools from statistical estimation theory. In this paper, we consider PI2- as a member of the wider family of methods which share the concept of probability-weighted averaging to iteratively update parameters to optimize a cost function. At the conceptual level, we compare PI2 to other members of the same family, being Cross-Entropy Methods and CMAES. The comparison suggests the derivation of a novel algorithm which we call PI2-CMA for ''Path Integral Policy Improvement with Covariance Matrix Adaptation''. PI2-CMA's main advantage is that it determines the magnitude of the exploration noise automaticall