Research map / Learning theory and algorithms / Combinatorial and search optimization
Scalable Mixed-Integer Optimization with Neural Constraints via Dual Decomposition
Shuli Zeng, Sijia Zhang, Feng Wu, Shaojie Tang, Xiangyang Li
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 |
|---|---|---|---|
| add dual update and augmented term (ablation) | DD · Water Potability n=20 | quality relative to SCIP: 0.454 -> 1.000 | Improved |
| add dual update and augmented term (ablation) | DD · Water Potability n=80 | iterations: 100 -> 48 | Improved |
| add dual update and augmented term (ablation) | DD · Water Potability n=80 | speedup vs SCIP: 20.9 -> 111.7 | Improved |
| add dual update and augmented term (ablation) | DD · Water Potability n=20 | speedup vs SCIP: 0.3 -> 0.3 | No clear effect |
| swap NN subsolver from log-barrier to projected gradient | DD · Tree Planting 24x128 | runtime (s): 111.0 -> 6.5 | Improved |
| use DD instead of Big-M linearisation (SCIP-ML) | DD-PGD · Tree Planting 24x128 | runtime (s): 860.3 -> 6.5 | Improved |
| use DD-Barrier instead of Big-M linearisation on smallest net | DD-Barrier · Tree Planting 8x32 | runtime (s): 8.5 -> 24.8 | Hurt |
| use DD instead of Big-M linearisation | DD · Water Potability | solution quality | No clear effect |
| swap MLP for LSTM backbone | DD · Water Potability | fraction converged in 50 iterations: 0.83 -> 0.50 | Hurt |
| increase network from 2x16 to 64x512 | DD · Water Potability n=100 | total time (s): 7.7 -> 113.6 | Hurt |
Closest papers by meaning
- Predicting Lagrangian Multipliers for Mixed Integer Linear Programs (ICML 2024 · similarity 0.88)
- Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming (ICLR 2025 · similarity 0.87)
- Expressive Power of Graph Neural Networks for (Mixed-Integer) Quadratic Programs (ICML 2025 · similarity 0.87)
- BPQP: A Differentiable Convex Optimization Framework for Efficient End-to-End Learning (NeurIPS 2024 · similarity 0.86)
- BIPNN: Learning to Solve Binary Integer Programming via Hypergraph Neural Networks (NeurIPS 2025 · similarity 0.86)
- Deep Distributed Optimization for Large-Scale Quadratic Programming (ICLR 2025 · similarity 0.86)
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.
