MILP Solving on Multiple Knapsack (Statistical Performance)
0.52Geometric Mean Execution TimeSCIP default
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SCIP defaultConfiguration=Default2023.06 | 0.52 | — | |
| TreeDQNLoss Function=Modified learning objective2023.06 | 0.9 | — | |
| tmdp+DFS2023.06 | 0.9 | 5.28 | |
| FMSTS2023.06 | 0.93 | 9.77 | |
| TreeDQN (per)Experience Replay=Prioritized2023.06 | 1.08 | 8.95 | |
| TreeDQN (MSE)Loss Function=Standard MSE2023.06 | 1.11 | 4.31 | |
| ILApproach=Imitation Learning2023.06 | 1.83 | 4.38 | |
| Strong BranchingStrategy=Strong Branching2023.06 | 1.94 | — |