Multi-agent coordination on Dense-Obstacle Map
88.37Final Cumulative Win RateDPP
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| DPPTraining steps budget=2.0M2026.03 | 88.37 | 88.37 | 0.87 | 2.8 | |
| QMIXTraining steps budget=2.0M2026.03 | 88.12 | 88.12 | 2.01 | 7.6 | |
| SOG(Vision)Training steps budget=2.0M2026.03 | 83.51 | 83.51 | 1.99 | 2.6 | |
| SOG(RL-Vision)Training steps budget=2.0M2026.03 | 81.11 | 83.6 | 0.49 | 7.5 | |
| CommFormerTraining steps budget=2.0M2026.03 | 79.37 | 79.37 | 1.99 | 7.6 | |
| EuclidTraining steps budget=2.0M2026.03 | 78.81 | 83.51 | 0.61 | 2.7 | |
| MAPPOTraining steps budget=2.0M2026.03 | 77.79 | 77.79 | 2.01 | 13.2 |