EAQR: A Multiagent Q-Learning Algorithm for Coordination of Multiple Agents

Figure 11

Four-agent/12-vertex box-pushing.