Continuous-state and discrete-action control on LunarLander v3
232.9Average RewardOracle QHD†
Evaluation Results
| Method | Links | |
|---|---|---|
| Oracle QHD†Encoder Setting=Hetero2026.05 | 232.9 | |
| FedQHDEncoder Setting=Homo2026.05 | 224.1 | |
| Oracle DQN†Encoder Setting=Hetero2026.05 | 202.6 | |
| FedAvg-DQNEncoder Setting=Homo2026.05 | 200.9 | |
| Indep. QHDEncoder Setting=Homo2026.05 | 187.3 | |
| Oracle DQN†Encoder Setting=Homo2026.05 | 183.5 | |
| Indep. QHDEncoder Setting=Hetero2026.05 | 174.6 | |
| Distill. FedDQNEncoder Setting=Hetero2026.05 | 131.4 | |
| Distill. FedDQNEncoder Setting=Homo2026.05 | 122.5 | |
| FedQHDEncoder Setting=Hetero2026.05 | 101.5 | |
| Oracle QHD†Encoder Setting=Homo2026.05 | 87.6 | |
| Trunc. FedAvg-QHDEncoder Setting=Homo2026.05 | 65 | |
| Trunc. FedAvg-QHDEncoder Setting=Hetero2026.05 | 24.8 |