Constrained Optimization on Concentric Circles Quadratic objective
100Feasibility (%)Projected Gradient
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Projected GradientMethod Category=Classical2026.04 | 100 | 31.24 | 0.96 | |
| FABMethod Category=ML-Based2026.04 | 100 | 0.45 | 1.75 | |
| FSNetMethod Category=ML-Based2026.04 | 99.9 | 1.75 | 1.29 | |
| FAB: 2 Decoder NetsMethod Category=ML-Based2026.04 | 99.8 | 0.5 | 1.49 | |
| FAB: 3 Decoder NetsMethod Category=ML-Based2026.04 | 99.3 | 0.86 | 1.5 | |
| FAB: Phase 1 OnlyMethod Category=ML-Based2026.04 | 99.1 | 0.77 | 0.89 | |
| FAB: No Hinge LossMethod Category=ML-Based2026.04 | 99.1 | 0.74 | 1.3 | |
| FAB: No DξMethod Category=ML-Based2026.04 | 98.3 | 0.74 | 1.02 | |
| Homeomorphic ProjectionMethod Category=ML-Based2026.04 | 93.7 | 166 | 4.83 | |
| Interior PointMethod Category=Classical2026.04 | 93.4 | 54.53 | 2.67 | |
| Penalty NNMethod Category=ML-Based2026.04 | 92.3 | 0.24 | 0.92 | |
| Augmented LagrangianMethod Category=Classical2026.04 | 37.2 | 57.59 | 1.16 | |
| Penalty MethodMethod Category=Classical2026.04 | 35.5 | 56.68 | 1.12 |