Multi-Agent Reinforcement Learning on Multi-Agent MuJoCo HalfCheetah back shin
7,176Average ReturnMAT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MAT#examples=10%, pre-training status=Pre-trained2022.05 | 7,176 | 74 | |
| MAT#examples=5%, pre-training status=Pre-trained2022.05 | 6,077 | 209 | |
| MAT#examples=1%, pre-training status=Pre-trained2022.05 | 4,143 | 230 | |
| MAT#examples=0%, pre-training status=Pre-trained2022.05 | 4,005 | 316 | |
| MAPPO#examples=10%2022.05 | 2,955 | 127 | |
| MAPPO#examples=5%2022.05 | 2,796 | 137 | |
| MAPPO#examples=1%2022.05 | 2,542 | 108 | |
| MAPPO#examples=0%2022.05 | 2,406 | 32 | |
| MAT-from scratch#examples=10%, pre-training status=From scratch2022.05 | 1,252 | 1,123 | |
| MAT-from scratch#examples=5%, pre-training status=From scratch2022.05 | 743 | 537 | |
| MAT-from scratch#examples=0%, pre-training status=From scratch2022.05 | -0.31 | 0.1 | |
| MAT-from scratch#examples=1%, pre-training status=From scratch2022.05 | -3.95 | 17 |