Multi-Agent Reinforcement Learning on SMAC 3s5z_vs_3s6z (test)
20.94Test Win RateDDN
Evaluation Results
| Method | Links | |
|---|---|---|
| DDNFramework=DFAC, Training Length=8M timesteps2021.02 | 20.94 | |
| QMIXTraining Length=8M timesteps2021.02 | 20.16 | |
| VDNTraining Length=8M timesteps2021.02 | 19.75 | |
| DMIXFramework=DFAC, Training Length=8M timesteps2021.02 | 19.7 | |
| DIQLTraining Length=8M timesteps2021.02 | 17.52 | |
| IQLTraining Length=8M timesteps2021.02 | 16.54 | |
| VDN2020.03 | 2 | |
| QMIX2020.03 | 2 |