MILP Solving on Facility Location
3.9Geometric Mean Execution TimeTreeDQN
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TreeDQNLoss Function=Modified learning objective2023.06 | 3.9 | — | |
| ILApproach=Imitation Learning2023.06 | 3.99 | 2.66 | |
| TreeDQN (per)Experience Replay=Prioritized2023.06 | 4.15 | 6.84 | |
| TreeDQN (MSE)Loss Function=Standard MSE2023.06 | 4.6 | 4.17 | |
| tmdp+DFS2023.06 | 5.29 | 1.4 | |
| FMSTS2023.06 | 5.39 | 1.44 | |
| SCIP defaultConfiguration=Default2023.06 | 7.27 | — | |
| Strong BranchingStrategy=Strong Branching2023.06 | 24.81 | — |