Reinforcement Learning on HalfCheetah Pure
7,459Average Reward (0.3/1.7)True env+MPC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| True env+MPCData source (optimality)=Oracle (True Environment)2021.02 | 7,459 | 42,893 | 66,675 | |
| PerSimData source (optimality)=Pure2021.02 | 1,984 | 997 | 714.7 | |
| BCQData source (optimality)=Pure, Algorithm variant=Planning2021.02 | 549.8 | 2,006 | -65.18 | |
| PE-TS CaDMData source (optimality)=Pure2021.02 | 481.1 | 503.7 | 553 | |
| Vanilla CaDMData source (optimality)=Pure2021.02 | 50.31 | -134 | 11.39 | |
| CQLData source (optimality)=Pure, Algorithm variant=Acting2021.02 | -65 | -257.9 | -279.6 | |
| BCQData source (optimality)=Pure, Algorithm variant=Acting2021.02 | -262.7 | -139 | 165.6 | |
| CQLData source (optimality)=Pure, Algorithm variant=Planning2021.02 | -353.5 | -453.6 | -476.7 | |
| MORELData source (optimality)=Pure, Algorithm variant=Acting2021.02 | -726.2 | -666.1 | -841.7 | |
| MORELData source (optimality)=Pure, Algorithm variant=Planning2021.02 | -1,297 | -1,256 | -1,470 |