Offline Reinforcement Learning on OGBench antsoccer-arena-singletask 5 tasks
69Normalized ScoreSORL
Evaluation Results
| Method | Links | |
|---|---|---|
| SORLPolicy Class=Multi-step / Iterative Policy2026.04 | 69 | |
| DeFlowPolicy Class=Multi-step / Iterative Policy2026.04 | 67 | |
| DROLPolicy Class=One-step Method, K-selection protocol=benchmark-specific2026.04 | 65 | |
| DROLPolicy Class=One-step Method, K=162026.04 | 62 | |
| FQLPolicy Class=One-step Method2026.04 | 60 | |
| IFQLPolicy Class=Multi-step / Iterative Policy2026.04 | 33 | |
| FBRACPolicy Class=Multi-step / Iterative Policy2026.04 | 16 | |
| FAWACPolicy Class=Multi-step / Iterative Policy2026.04 | 12 | |
| IQLPolicy Class=Gaussian Policy2026.04 | 8 | |
| BCPolicy Class=Gaussian Policy2026.04 | 1 | |
| ReBRACPolicy Class=Gaussian Policy2026.04 | 0 |