Continuous Control on MuJoCo Walker2d (Robustness Sweep)
4,056.6Return (pact=0.1)DR-SPCRL
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DR-SPCRLAlgorithm=SAC2025.11 | 4,056.6 | 2,438.1 | 960.6 | 425.7 | 183.5 | 4,187.8 | 1,444.6 | 319.7 | 155.4 | 108 | 5,211.5 | 4,233.1 | 3,317.9 | 2,501.1 | 1,661.9 | |
| SACAlgorithm=SAC2025.11 | 3,241 | 1,603.8 | 639 | 199.8 | 77.7 | 2,600.2 | 274.2 | 207.7 | 183.2 | 151.3 | 4,129.8 | 3,375.2 | 2,219.1 | 1,382.2 | 848.9 | |
| Dom. Rand.Algorithm=SAC2025.11 | 2,743.9 | 1,160 | 472.1 | 229.5 | 140.7 | 3,007.2 | 518 | 260.1 | 172.2 | 116.6 | 4,765.8 | 3,856.9 | 2,916.2 | 1,833.6 | 1,059.8 | |
| FixedAlgorithm=SAC2025.11 | 2,303.6 | 1,130.8 | 496.4 | 227 | 122.7 | 1,983.3 | 343.6 | 214.8 | 198.4 | 172.3 | 4,001.2 | 2,826.1 | 1,996.6 | 1,252.5 | 804.5 | |
| SPACEAlgorithm=SAC2025.11 | 2,240.5 | 1,071.5 | 368.3 | 194.4 | 117.4 | 2,758.2 | 562.8 | 242.5 | 195.5 | 151.7 | 3,225.8 | 2,592.9 | 1,965 | 1,244.8 | 721.4 | |
| LinearAlgorithm=SAC2025.11 | 1,997.8 | 990.8 | 418.3 | 220.5 | 142.7 | 1,924.7 | 266.1 | 169.8 | 110.2 | 97.1 | 3,500.3 | 2,583 | 1,741 | 1,285.4 | 816.6 | |
| ACCELAlgorithm=SAC2025.11 | 1,854.5 | 921.6 | 222.5 | 115.7 | 52 | 2,084.4 | 421.2 | 246.7 | 192.7 | 153.4 | 2,596.3 | 1,981 | 1,307.4 | 882 | 616.6 |