Multi-Agent Reinforcement Learning on Simple Spread
86.7Success RateQMIX + EMA
Evaluation Results
| Method | Links | |
|---|---|---|
| QMIX + EMARegime=Augmentative, Algorithm=QMIX, Stabilization method=EMA, Number of seeds=52026.07 | 86.7 | |
| EMAalpha=0.22026.07 | 86.7 | |
| Freeze-Schedule2026.07 | 84.1 | |
| QMIX (Baseline)Regime=Augmentative, Algorithm=QMIX, Stabilization method=None, Number of seeds=52026.07 | 74.4 | |
| Baseline QMIXshaping=no2026.07 | 74.4 | |
| Dynamic LLMgranularity=per-episode2026.07 | 15.2 |