Constrained Reinforcement Learning on Safety Gymnasium CarCircle1 (val)
18.6RewardFOCOPS
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FOCOPSKnowledge Level=Oracle2025.04 | 18.6 | 40.7 | — | |
| CUPKnowledge Level=Oracle2025.04 | 17.5 | 27.5 | — | |
| TraCeSKnowledge Level=Zero Knowledge2025.04 | 17.4 | 11.3 | 4,005 | |
| TRPO-SauteKnowledge Level=Oracle2025.04 | 14.4 | 26.4 | — |