Research map / Learning theory and algorithms / Combinatorial and search optimization
Data-driven Mixed Integer Optimization through Probabilistic Multi-variable Branching
Yanguang Chen, Wenzhi Gao, Wanyu Zhang, Dongdong Ge, Huikang Liu, Yinyu Ye
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 PMVB as an external branching rule in Gurobi | Gurobi · SCP, CA, MKP | SGM solve time: 10% to 30% improvement on solution time | Improved |
| add PMVB as primal heuristic with logistic regression on MKP | Gurobi/COPT · MKP | speedup: speedups of at least 500% | Improved |
| add PMVB as primal heuristic with logistic regression on SCP | Gurobi/COPT · SCP | speedup: most cases exceeding 1000% | Improved |
| add PMVB as primal heuristic with logistic regression on SCUC | Gurobi/COPT · SCUC | average speedup: over 50% | Improved |
| use GNN predictions with varying problem size on SCP | PMVB · SCP | speedup | Mixed |
| use data-free PMVB with LP relaxation surrogate on MIPLIB | COPT · MIPLIB (57 instances) | SGM: improvement of up to 16% | Improved |
| set delta to 10^-8 instead of 0.01 in data-free PMVB (ablation) | data-free PMVB · MIPLIB (COPT) | SGM vs no PMVB: > 1.0 -> 0.84 | Improved |
| set delta to 10^-15 instead of 10^-8 in data-free PMVB (ablation) | data-free PMVB · MIPLIB (COPT) | SGM vs no PMVB: 0.84 -> 0.94 | Hurt |
| use data-free PMVB on SCP and CA with Gurobi and COPT | Gurobi/COPT · SCP, CA | speedup | Improved |
Closest papers by meaning
- Learning to Branch in Mixed Integer Programming (AAAI 2016 · similarity 0.92)
- Generalization Guarantees for Learning Score-Based Branch-and-Cut Policies in Integer Programming (NeurIPS 2025 · similarity 0.88)
- Sample Complexity of Algorithm Selection Using Neural Networks and Its Applications to Branch-and-Cut (NeurIPS 2024 · similarity 0.87)
- Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming (ICLR 2025 · similarity 0.87)
- Contrastive Predict-and-Search for Mixed Integer Linear Programs (ICML 2024 · similarity 0.86)
- Fast and Interpretable Mixed-Integer Linear Program Solving by Learning Model Reduction (AAAI 2025 · similarity 0.85)
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.
