Research map / Graph learning / Graph algorithms and clustering
Accelerating Primal Solution Findings for Mixed Integer Programs Based on Solution Prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use GCN predictions to add a local branching cut in approximate solving | SCIP · Eight MIP problem types; average | primal gap (%): 1.151 -> 0.316 | Helped |
| Use GCN predictions to add a local branching cut in approximate solving | SCIP · MK | primal gap (%): 0.000 -> 0.003 | Hurt |
| Use GCN predictions to guide an exact root branch | SCIP · Seven of eight MIP types | optimality gap (%) | Helped |
| Use GCN predictions to guide an exact root branch | SCIP · TSP | optimality gap (%): 12.13 -> 12.84 | Hurt |
| Use a tripartite-graph GCN instead of XGBoost for prediction with basic features | Binary-variable solution prediction · Eight MIP problem types; average | average precision: 0.415 -> 0.537 | Helped |
| Use a tripartite-graph GCN instead of XGBoost for prediction with all features | Binary-variable solution prediction · Eight MIP problem types; average | average precision: 0.698 -> 0.707 | Helped |
| Train on small-scale instances and test on large-scale instances | GCN solution prediction · Eight MIP problem types; average | average precision: 0.621 -> 0.593 | Hurt |
| Train on small-scale instances and use the resulting predictions for approximate solving | SCIP with GCN-A · Eight MIP problem types; average | primal gap (%): GCN-A 0.338; GCNG-A 1.236 | Mixed |
Closest papers by meaning
- Contrastive Predict-and-Search for Mixed Integer Linear Programs (ICML 2024 · similarity 0.89)
- Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming (ICLR 2025 · similarity 0.88)
- Expressive Power of Graph Neural Networks for (Mixed-Integer) Quadratic Programs (ICML 2025 · similarity 0.88)
- Generalization Guarantees for Learning Score-Based Branch-and-Cut Policies in Integer Programming (NeurIPS 2025 · similarity 0.88)
- BIPNN: Learning to Solve Binary Integer Programming via Hypergraph Neural Networks (NeurIPS 2025 · similarity 0.87)
- Fast and Interpretable Mixed-Integer Linear Program Solving by Learning Model Reduction (AAAI 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.