Multi-Agent Reinforcement Learning on SMAC 3m to 8m
100Final Win RateGCT-MARL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GCT-MARL2026.06 | 100 | 1.5 | |
| MAIL2026.06 | 100 | 4.91 | |
| ASALT2026.06 | 82 | — | |
| CORALtransfer_scheme=agent-wise transfer2026.06 | 81 | — | |
| LA-QTransformer2026.06 | 81 | — | |
| CORAL2026.06 | 81 | 2.8 | |
| LA-QT2026.06 | 81 | 1.72 | |
| Baselinetraining=from scratch using MAPPO2026.06 | 80 | — | |
| Baseline2026.06 | 80 | 5.1 | |
| DANNtransfer_scheme=agent-wise transfer2026.06 | 79 | — | |
| DANN2026.06 | 79 | 3.1 | |
| CycleGANtransfer_scheme=agent-wise transfer2026.06 | 75 | — | |
| CycleGAN2026.06 | 75 | 4.1 |