Multi-agent Offline Reinforcement Learning on MPE World (Medium-replay) benchmark leaderboard · SOTA2 Research

Multi-agent Offline Reinforcement Learning on MPE World (Medium-replay)

73.4ScoreCFCQL

Evaluation Results

Method	Links
CFCQL2023.09		73.4	—
OMAR2023.09		42.9	—
IQL2023.09		41.5	—
ICQL2023.09		29.6	—
MATD3-BC2023.09		17.4	—
MACQL2023.09		13.2	—
MAICQ2023.09		12	—
AWAC2023.09		8.9	—
CFCQL2026.03		—	73.4
DoF2026.03		—	58.6
IQL2026.03		—	41.5
MA-TD3+BC2026.03		—	17.4
MACQL2026.03		—	13.2
OMAR2026.03		—	42.9
PLCQL2026.03		—	107.5
SPaCQL2026.03		—	105.2