Polynomial-Objective Integer Programming on RandQCP 1k
0.32Gap (%)HNN
Evaluation results
| Method | Links | |
|---|---|---|
| HNNTrain=Mini, Base solver=Gurobi2026.03 | 0.32 | |
| HNNTrain=1000, Base solver=Gurobi2026.03 | 0.43 | |
| HNNTrain=Mini, Base solver=SCIP2026.03 | 0.46 | |
| NeuralQPTrain=Mini, Base solver=Gurobi2026.03 | 0.47 | |
| HNNTrain=1000, Base solver=SCIP2026.03 | 0.5 | |
| NeuralQPTrain=1000, Base solver=Gurobi2026.03 | 0.54 | |
| TriGNNTrain=Mini, Base solver=Gurobi2026.03 | 0.55 | |
| TriGNNTrain=1000, Base solver=Gurobi2026.03 | 0.57 | |
| NeuralQPTrain=Mini, Base solver=SCIP2026.03 | 0.6 | |
| NeuralQPTrain=1000, Base solver=SCIP2026.03 | 0.62 | |
| TriGNNTrain=Mini, Base solver=SCIP2026.03 | 0.67 | |
| TriGNNTrain=1000, Base solver=SCIP2026.03 | 0.72 | |
| Exact solverTrain=-, Base solver=Gurobi2026.03 | 3.21 | |
| GNNQPTrain=1000, Base solver=Gurobi2026.03 | 3.48 | |
| GNNQPTrain=Mini, Base solver=Gurobi2026.03 | 3.53 | |
| GNNQPTrain=Mini, Base solver=SCIP2026.03 | 3.6 | |
| GNNQPTrain=1000, Base solver=SCIP2026.03 | 3.62 | |
| Exact solverTrain=-, Base solver=SCIP2026.03 | 38.82 |