Safe Reinforcement Learning on Constrained Pendulum (100 rollouts)
100ACS (%)Ours
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Oursrollouts=1002026.04 | 100 | 100 | |
| TD3rollouts=1002026.04 | 99 | 80 | |
| PPO Barrierrollouts=1002026.04 | 99 | 91 | |
| CPOrollouts=1002026.04 | 94 | 18 |
| Method | Links | ||
|---|---|---|---|
| Oursrollouts=1002026.04 | 100 | 100 | |
| TD3rollouts=1002026.04 | 99 | 80 | |
| PPO Barrierrollouts=1002026.04 | 99 | 91 | |
| CPOrollouts=1002026.04 | 94 | 18 |