Model-Free Policy Gradients for Multi-Agent Shape Formation

Abstract

We develop a model-free technique based on gradient ascent for multi-agent systems to learn distributed policies for tight coordination, and we demonstrate its performance on the problem of shape formation.