Multi-Agent Reinforcement Learning on SMAC 3svs3z to 3svs5z
0.89Environment Steps to 80% RewardASALT
Evaluation results
| Method | Links | |
|---|---|---|
| ASALTScale=10^5 environment steps2026.06 | 0.89 | |
| MALTScale=10^5 environment steps2026.06 | 1.41 | |
| EPCScale=10^5 environment steps2026.06 | 2.18 | |
| Distilled PolicyScale=10^5 environment steps2026.06 | 3.42 | |
| Fine-tuneScale=10^5 environment steps2026.06 | 3.92 | |
| PSMARLScale=10^5 environment steps2026.06 | 4.05 | |
| BaselineScale=10^5 environment steps2026.06 | 4.17 |