Constrained Optimization (Quadratic Objective) on Two Moons 1,500 problems (test)
100Feasibility (%)FAB: 2 Decoder Nets
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FAB: 2 Decoder NetsType=ML-Based2026.04 | 100 | 0.59 | 1.31 | |
| FAB: 3 Decoder NetsType=ML-Based2026.04 | 100 | 0.8 | 1.02 | |
| FSNetType=ML-Based2026.04 | 98.7 | 4.42 | 1.13 | |
| FABType=ML-Based2026.04 | 96.8 | 0.55 | 0.85 | |
| Penalty NNType=ML-Based2026.04 | 91.1 | 0.2 | 0.61 | |
| FAB: Phase 1 OnlyType=ML-Based2026.04 | 82.5 | 0.64 | 0.86 | |
| Projected GradientType=Classical2026.04 | 80.5 | 79.07 | 1.18 | |
| Interior PointType=Classical2026.04 | 78.3 | 82.8 | 1.64 | |
| FAB: No DξType=ML-Based2026.04 | 75.1 | 0.64 | 0.86 | |
| Homeomorphic ProjectionType=ML-Based2026.04 | 73.9 | 247 | 1.74 | |
| FAB: No Hinge LossType=ML-Based2026.04 | 28.4 | 0.63 | 1.13 | |
| Augmented LagrangianType=Classical2026.04 | 20.9 | 85.71 | 1.33 | |
| Penalty MethodType=Classical2026.04 | 20.4 | 78.39 | 1.36 |