Multi-agent Reinforcement Learning on SMAC 2s3z (Heterog.)
84.6Win RateMetaMind
Evaluation Results
| Method | Links | |
|---|---|---|
| MetaMindMethod Category=Multi-Agent World Model Methods, Execution Protocol=Decentralized, Training Steps=100K, Number of Seeds=1002026.02 | 84.6 | |
| MARIEMethod Category=Multi-Agent World Model Methods, Execution Protocol=CTDE, Training Steps=100K, Number of Seeds=1002026.02 | 82.3 | |
| DCWMCMethod Category=Multi-Agent World Model Methods, Execution Protocol=Decentralized, Training Steps=100K, Number of Seeds=1002026.02 | 54.8 | |
| MBVDMethod Category=Multi-Agent RL Methods, Execution Protocol=CTDE, Training Steps=100K, Number of Seeds=1002026.02 | 42.7 | |
| MAPPOMethod Category=Multi-Agent RL Methods, Execution Protocol=CTDE, Training Steps=100K, Number of Seeds=1002026.02 | 31.2 |