Polynomial-Objective Integer Programming on RandQCP 10k
0.14Gap (%)HNN
Evaluation results
| Method | Links | |
|---|---|---|
| HNNTrain=Mini, Base solver=Gurobi2026.03 | 0.14 | |
| HNNTrain=Mini, Base solver=SCIP2026.03 | 0.17 | |
| HNNTrain=1000, Base solver=Gurobi2026.03 | 0.17 | |
| HNNTrain=2000, Base solver=Gurobi2026.03 | 0.18 | |
| NeuralQPTrain=2000, Base solver=Gurobi2026.03 | 0.19 | |
| HNNTrain=1000, Base solver=SCIP2026.03 | 0.21 | |
| NeuralQPTrain=Mini, Base solver=Gurobi2026.03 | 0.22 | |
| NeuralQPTrain=1000, Base solver=Gurobi2026.03 | 0.22 | |
| TriGNNTrain=2000, Base solver=Gurobi2026.03 | 0.23 | |
| TriGNNTrain=1000, Base solver=Gurobi2026.03 | 0.25 | |
| HNNTrain=2000, Base solver=SCIP2026.03 | 0.26 | |
| TriGNNTrain=Mini, Base solver=Gurobi2026.03 | 0.27 | |
| NeuralQPTrain=2000, Base solver=SCIP2026.03 | 0.29 | |
| TriGNNTrain=Mini, Base solver=SCIP2026.03 | 0.29 | |
| NeuralQPTrain=Mini, Base solver=SCIP2026.03 | 0.31 | |
| NeuralQPTrain=1000, Base solver=SCIP2026.03 | 0.32 | |
| TriGNNTrain=1000, Base solver=SCIP2026.03 | 0.32 | |
| TriGNNTrain=2000, Base solver=SCIP2026.03 | 0.7 | |
| GNNQPTrain=Mini, Base solver=Gurobi2026.03 | 3.36 | |
| GNNQPTrain=1000, Base solver=Gurobi2026.03 | 3.36 | |
| GNNQPTrain=2000, Base solver=Gurobi2026.03 | 3.38 | |
| GNNQPTrain=1000, Base solver=SCIP2026.03 | 3.4 | |
| GNNQPTrain=2000, Base solver=SCIP2026.03 | 3.41 | |
| GNNQPTrain=Mini, Base solver=SCIP2026.03 | 3.42 | |
| Exact solverTrain=-, Base solver=Gurobi2026.03 | 9.88 | |
| Exact solverTrain=-, Base solver=SCIP2026.03 | 53.07 |