Polynomial-Objective Integer Programming on RandQCP 5k
0.19Gap (%)HNN
Evaluation results
| Method | Links | |
|---|---|---|
| HNNTrain=Mini, Base solver=Gurobi2026.03 | 0.19 | |
| HNNTrain=2000, Base solver=Gurobi2026.03 | 0.2 | |
| NeuralQPTrain=1000, Base solver=Gurobi2026.03 | 0.22 | |
| HNNTrain=1000, Base solver=Gurobi2026.03 | 0.23 | |
| NeuralQPTrain=Mini, Base solver=Gurobi2026.03 | 0.25 | |
| NeuralQPTrain=2000, Base solver=Gurobi2026.03 | 0.25 | |
| TriGNNTrain=1000, Base solver=Gurobi2026.03 | 0.27 | |
| HNNTrain=Mini, Base solver=SCIP2026.03 | 0.27 | |
| TriGNNTrain=Mini, Base solver=Gurobi2026.03 | 0.28 | |
| TriGNNTrain=2000, Base solver=Gurobi2026.03 | 0.28 | |
| HNNTrain=1000, Base solver=SCIP2026.03 | 0.3 | |
| HNNTrain=2000, Base solver=SCIP2026.03 | 0.32 | |
| NeuralQPTrain=2000, Base solver=SCIP2026.03 | 0.34 | |
| NeuralQPTrain=1000, Base solver=SCIP2026.03 | 0.36 | |
| TriGNNTrain=2000, Base solver=SCIP2026.03 | 0.36 | |
| NeuralQPTrain=Mini, Base solver=SCIP2026.03 | 0.37 | |
| TriGNNTrain=Mini, Base solver=SCIP2026.03 | 0.37 | |
| TriGNNTrain=1000, Base solver=SCIP2026.03 | 0.38 | |
| GNNQPTrain=Mini, Base solver=Gurobi2026.03 | 3.23 | |
| GNNQPTrain=2000, Base solver=Gurobi2026.03 | 3.23 | |
| GNNQPTrain=1000, Base solver=Gurobi2026.03 | 3.25 | |
| GNNQPTrain=Mini, Base solver=SCIP2026.03 | 3.29 | |
| GNNQPTrain=2000, Base solver=SCIP2026.03 | 3.29 | |
| GNNQPTrain=1000, Base solver=SCIP2026.03 | 3.31 | |
| Exact solverTrain=-, Base solver=Gurobi2026.03 | 5.83 | |
| Exact solverTrain=-, Base solver=SCIP2026.03 | 53.12 |