Learning Attentional Communication with a Common Network for Multiagent Reinforcement Learning

Table 2

Maximum median performance % of the algorithms tested.

Scenarios	MAACCN-QMIX	QMIX	VDN	IQL	Heuristic	QTRAN	Qatten	COMA
2s_vs_1sc	100	100	100	100	0	100	100	96
3s5z	97	91	87	9	42	20	95	0
5m_vs_6m	75	75	78	59	0	58	74	0
3s_vs_5z	100	97	73	46	0	15	97	0
6h_vs_8z	30	3	0	0	0	0	4	0