Reinforcement Learning on Hopper-Gravity Start OpenAI Gym Mujoco (initialization)
306.79Average RewardEWC
Evaluation results
| Method | Links | |
|---|---|---|
| EWCsamples per iteration=2x2022.10 | 306.79 | |
| LPG-FTWsamples per iteration=2x2022.10 | 305.63 | |
| VBLRLsamples per iteration=1x2022.10 | 265.11 | |
| T-HiP-MDPsamples per iteration=1x2022.10 | 230.21 | |
| Single-task MBRLmode=train from scratch2022.10 | 18.62 |