Portfolio Optimization on 15 financial assets (7-fold cross-val)
0.8237Sharpe RatioClassical DQN
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Classical DQNNumber of Parameters=160k, Hardware=Nvidia RTX 8000 GPU2026.01 | 0.8237 | 0.5367 | 27 | — | |
| Classical DDPGNumber of Parameters=160k, Hardware=Nvidia RTX 8000 GPU2026.01 | 0.7926 | 0.6276 | 14 | — | |
| Quantum DDPGNumber of Parameters=60, Hardware=IBM Eagle r3 QPU2026.01 | 0.7281 | 0.7395 | — | 23 | |
| Quantum DQNNumber of Parameters=60, Hardware=IBM Eagle r3 QPU2026.01 | 0.6537 | 0.6155 | — | 43 | |
| Mean-Variance OptimizationHardware=Nvidia RTX 8000 GPU2026.01 | 0.5919 | 0.917 | — | — | |
| Classical DQNNumber of Parameters=27k, Hardware=Nvidia RTX 8000 GPU2026.01 | 0.4939 | 0.6623 | 18 | — | |
| Quantum DQNNumber of Parameters=30, Hardware=IBM Eagle r3 QPU2026.01 | 0.4776 | 0.665 | — | 43 | |
| Classical DQNNumber of Parameters=13.5k, Hardware=Nvidia RTX 8000 GPU2026.01 | 0.4446 | 0.6568 | 11 | — | |
| Classical DDPGNumber of Parameters=27k, Hardware=Nvidia RTX 8000 GPU2026.01 | 0.4384 | 0.6905 | 10 | — | |
| Equal Weights PortfolioHardware=Nvidia RTX 8000 GPU2026.01 | 0.4375 | 0.6558 | — | — | |
| Classical DDPGNumber of Parameters=13.5k, Hardware=Nvidia RTX 8000 GPU2026.01 | 0.4287 | 0.6151 | 7 | — | |
| Quantum DDPGNumber of Parameters=30, Hardware=IBM Eagle r3 QPU2026.01 | 0.4179 | 0.6113 | — | 23 |