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

Table 2

Average success rate for 4-agent/12-vertex box-pushing (evaluation episodes = 50,000).

	= 100,000	= 500,000	= 1000,000
EAQR	82.6%	98.6%	99.6%
WoLF-PHC	80.7%	87.1%	91.7%
EMA Q-learning	66.7%	76.6%	78.7%
Single-agent RL	60.2%	91.2%	95.9%