Offline Multi-objective Optimization on Omnitest
0MSEGPR (RBF)+DR
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GPR (RBF)+DRNumber of training samples=1,000, Tau (τ)=0.902025.11 | 0 | 0.0058 | 1.16 | |
| GPR (Matern)+DRNumber of training samples=1,000, Tau (τ)=0.902025.11 | 0 | 0.006 | 1.16 | |
| Prob-MOEA/DNumber of training samples=1,0002025.11 | 0 | 0.464 | 1.02 | |
| BNN+DRNumber of training samples=1,000, Tau (τ)=0.902025.11 | 0.0028 | 0.0081 | 1.16 | |
| QR+DRNumber of training samples=1,000, Tau (τ)=0.902025.11 | 0.0036 | 0.0142 | 1.16 | |
| Prob-RVEANumber of training samples=1,0002025.11 | 0.0116 | 0.145 | 1.13 | |
| TGPR-MONumber of training samples=1,0002025.11 | 0.0116 | 0.0161 | 1.16 | |
| Prob-MOEA/DStrategy=baseline2025.11 | 0.0445 | 0.57 | 0.91 | |
| Prob-RVEAStrategy=baseline2025.11 | 0.196 | 0.18 | 1.07 | |
| TGPR-MOStrategy=baseline2025.11 | 0.297 | 0.222 | 1.03 | |
| QR+DRStrategy=dual-ranking (DR), Surrogate Model=QR, Quantile level (τ)=0.9, Optimizer=NSGA-II2025.11 | 0.37 | 0.5 | 0.91 | |
| GPR(RBF)+DRStrategy=dual-ranking (DR), Kernel=RBF, Quantile level (τ)=0.9, Optimizer=NSGA-II2025.11 | 0.437 | 0.215 | 1.06 | |
| GPR(Matérn)+DRStrategy=dual-ranking (DR), Kernel=Matérn, Quantile level (τ)=0.9, Optimizer=NSGA-II2025.11 | 0.556 | 0.195 | 1.08 | |
| DDMOEA/GANStrategy=baseline2025.11 | 1.31 | 0.248 | 1.08 | |
| DDMOEA/GANNumber of training samples=1,0002025.11 | 1.61 | 3.13 | 0.2 | |
| BNN+DRStrategy=dual-ranking (DR), Surrogate Model=BNN, Quantile level (τ)=0.9, Optimizer=NSGA-II2025.11 | 1.78 | 0.147 | 1.12 |