Offline Reinforcement Learning on D4RL Gym v2 (test)
105.9Score (HalfCheetah, Medium-Expert)MBOP
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MBOP2022.06 | 105.9 | 55.1 | 70.2 | 44.6 | 48.8 | 41 | 42.3 | 12.4 | 9.7 | 47.8 | — | |
| D-QL2025.10 | 96.8 | 111.1 | 110.1 | 51.1 | 90.5 | 87 | 47.8 | 101.3 | 95.5 | 87.9 | — | |
| IDQL-A2025.10 | 95.9 | 108.6 | 112.7 | 51 | 65.4 | 82.5 | 45.9 | 92.1 | 85.1 | 82.1 | — | |
| MISA2022.10 | 94.7 | 109.8 | 109.4 | 47.4 | 67.1 | 84.1 | 45.6 | 98.6 | 86.2 | — | 742.9 | |
| χ-QL2025.10 | 94.2 | 111.2 | 112.7 | 48.3 | 74.2 | 84.2 | 45.2 | 100.7 | 82.2 | 83.7 | — | |
| BooT2022.06 | 94 | 102.3 | 110.4 | 50.6 | 70.2 | 82.9 | 46.5 | 92.9 | 87.6 | 81.9 | — | |
| GTP2025.10 | 93.8 | 112.2 | 114.2 | 53.9 | 90.3 | 89.5 | 50.8 | 101.7 | 94.2 | 89 | — | |
| QGPO2025.10 | 93.5 | 108 | 110.7 | 54.1 | 98 | 86 | 47.6 | 96.9 | 84.4 | 86.6 | — | |
| OneStep RL2022.10 | 93.4 | 103.3 | 113 | 48.4 | 59.6 | 81.8 | 38.1 | 97.5 | 49.5 | — | 684.6 | |
| BDM2025.10 | 93.2 | 104.9 | 111.1 | 57 | 98.4 | 87.4 | 51.6 | 92.7 | 89.2 | 87.3 | — | |
| 10%BC2022.10 | 92.9 | 110.9 | 109 | 42.5 | 56.9 | 75 | 40.6 | 75.9 | 62.5 | — | 666.2 | |
| DTvariant=GPT22022.06 | 91.8 | 110.9 | 91 | 44 | 67.4 | 81.3 | 40.3 | 94.4 | 72.7 | 80.1 | — | |
| CQL2022.06 | 91.6 | 105.4 | 108.8 | 44 | 58.5 | 72.5 | 45.5 | 95 | 77.2 | 77.6 | — | |
| CQL2022.10 | 91.6 | 105.4 | 108.8 | 44 | 58.5 | 72.5 | 45.5 | 95 | 77.2 | — | 698.5 | |
| CQL2025.10 | 91.6 | 105.4 | 108.8 | 44 | 58.5 | 72.5 | 45.5 | 95 | 77.2 | 77.6 | — | |
| ARQ2025.10 | 91 | 110 | 109 | 45 | 61 | 81 | 42 | 81 | 66 | 76.2 | — | |
| TD3+BC2022.10 | 90.7 | 98 | 110.1 | 48.3 | 59.3 | 83.7 | 44.6 | 60.9 | 81.8 | — | 677.4 | |
| TTvariant=Re-train2022.06 | 90.5 | 67 | 107.6 | 44.1 | 64.9 | 81.6 | 36.5 | 90.8 | 50.3 | 69.4 | — | |
| TTvariant=+S4RL (All)2022.06 | 87.5 | 92.6 | 100.5 | 46.7 | 53.9 | 80.2 | 43.1 | 76.1 | 52.1 | 71.1 | — | |
| DTvariant=Original2022.06 | 86.8 | 107.6 | 108.1 | 42.6 | 67.6 | 74 | 36.6 | 82.7 | 66.6 | 74.7 | — | |
| DT2022.10 | 86.8 | 107.6 | 108.1 | 42.6 | 67.6 | 74 | 36.6 | 82.7 | 66.6 | — | 672.6 | |
| IQL2022.10 | 86.7 | 91.5 | 109.6 | 47.4 | 66.3 | 78.3 | 44.2 | 94.7 | 73.9 | — | 692.6 | |
| IQL2025.10 | 86.7 | 91.5 | 109.6 | 47.4 | 66.3 | 78.3 | 44.2 | 94.7 | 73.9 | 77 | — | |
| TTvariant=+S4RL (Last)2022.06 | 84.3 | 96.9 | — | 46 | 56 | 80 | 40.2 | 72.8 | 61.1 | 70.9 | — | |
| C-AC2025.10 | 84.3 | 100.4 | 110.4 | 69.1 | 80.7 | 83.1 | 58.7 | 99.7 | 79.5 | 85.1 | — | |
| TTvariant=Reproduce2022.06 | 81.4 | 66.2 | 99.1 | 46.1 | 67.1 | 71.1 | 43.1 | 86.4 | 66.9 | 70.1 | — | |
| BC2022.06 | 59.9 | 79.6 | 36.6 | 43.1 | 63.9 | 77.3 | 4.3 | 27.6 | 36.9 | 47.7 | — | |
| BC2022.10 | 55.2 | 52.5 | 107.5 | 42.6 | 52.9 | 75.3 | 36.6 | 18.1 | 26 | — | 466.7 | |
| AWAC2022.10 | 42.8 | 55.8 | 74.5 | 43.5 | 57 | 72.4 | 40.5 | 37.2 | 27 | — | 450.7 | |
| TTvariant=Original2022.06 | 40.8 | 106 | 102.2 | 42.8 | 79.1 | 78.3 | 44.1 | 99.4 | 79.4 | 72.6 | — |