Portfolio Optimization on EU30 cross-market (val)
11.89ReturnPPO+Penalty
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PPO+PenaltyOptimizer Interface=PPO Reference and Fixed-Penalty Baseline2026.05 | 11.89 | 1.17 | 1.4 | 18.97 | 6.68 | |
| MACF-TRPOOptimizer Interface=Trust-Region / Safe Geometry Interface2026.05 | 11.8 | 1.17 | 1.42 | 19.56 | 6.67 | |
| MACF-CPOOptimizer Interface=Trust-Region / Safe Geometry Interface2026.05 | 11.3 | 1.15 | 1.42 | 20.83 | 5.17 | |
| MACF-PPOOptimizer Interface=Scalarization / Lagrangian Interface2026.05 | 10.9 | 1.07 | 1.27 | 16.6 | 5 | |
| PPO-LagrangianOptimizer Interface=Scalarization / Lagrangian Interface2026.05 | 10.87 | 1.1 | 1.28 | 16.63 | 4.73 | |
| PPOOptimizer Interface=PPO Reference and Fixed-Penalty Baseline2026.05 | 10.86 | 1.04 | 1.26 | 19.67 | 8.69 | |
| CRPOOptimizer Interface=First-Order Feasibility Switching Interface2026.05 | 10.86 | 1.04 | 1.26 | 19.67 | 8.69 | |
| TRPOOptimizer Interface=Trust-Region / Safe Geometry Interface2026.05 | 10.76 | 1.08 | 1.3 | 17.7 | 6.2 | |
| CPOOptimizer Interface=Trust-Region / Safe Geometry Interface2026.05 | 10.76 | 1.08 | 1.3 | 17.7 | 6.2 | |
| MACF-CRPOOptimizer Interface=First-Order Feasibility Switching Interface2026.05 | 10.37 | 1.07 | 1.27 | 16.54 | 4.55 |