Collaborative Multiagent Reinforcement Learning by Payoff Propagation

In this article we describe a set of scalable techniques for learning the behavior of a group of agents in a collaborative multiagent setting. As a basis we use the framework of coordination graphs...