Research map / Learning theory and algorithms / Combinatorial and search optimization
A Novel Integer Linear Programming Approach for Global L0 Minimization
Diego Delle Donne, Matthieu Kowalski, Leo Liberti
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 forbidden-support cuts and rounding heuristic (BC0/p) to the big-M MIP (ablation) | MIP0/2 · Deconvolution P0/2 | average time on commonly solved instances: 360 seconds against 104 | Improved |
| add forbidden-support cuts and rounding heuristic (BC0/p) to the big-M MIP (ablation) | MIP0/2 · Deconvolution P0/2 (404 instances) | instances solved to optimality: 250 -> 319 | Improved |
| replace big-M MIP with M-independent covering ILP (IPcov) | MIP0/2 · Deconvolution P0/2 (404 instances) | instances solved to optimality: 250 -> 379 | Improved |
| replace big-M MIP with M-independent covering ILP (IPcov) | MIP0/2 · Deconvolution P0/2 | average time on commonly solved instances: 360 -> 12 seconds | Improved |
| add forbidden-support cuts (BC0/p) to the big-M MIP (ablation) | MIP0/2 · Underdetermined deconvolution P0/2 | average time on commonly solved instances: 150 -> 233 sec. | Hurt |
| replace big-M MIP with covering ILP (IPcov) | MIP0/2 · Underdetermined deconvolution P0/2 | average time on commonly solved instances: 150 -> 49 sec. | Improved |
| add forbidden-support cuts (BC0/p) to the big-M MIP (ablation) | MIP0/inf · Deconvolution P0/inf Dantzig selector | average time on commonly solved instances: 366 -> 290 sec. | Improved |
| replace big-M MIP with covering ILP (IPcov) | MIP0/inf · Deconvolution P0/inf Dantzig selector | average time on commonly solved instances: 290 -> 954 sec. | Hurt |
| replace big-M MIP with covering ILP (IPcov) | MIP0/inf · Deconvolution P0/inf Dantzig selector | average support size on unsolved instances: 9.04 -> 13.57 | Hurt |
| replace big-M MIP with covering ILP (IPcov) | MIP0/2 · Pathological adversarial instances 20x40, K=8 | average support size: 20.5 -> 8.7 | Improved |
| replace big-M MIP with covering ILP (IPcov) | MIP0/2 · Pathological adversarial instances P0/2 | instances with best solution: 34 -> 82 out of 90 | Improved |
| add forbidden-support cuts (BC0/p) to the big-M MIP (ablation) | MIP0/2 · Pathological adversarial instances P0/2 | instances with strictly better solution: 1 -> 0 | No clear effect |
Closest papers by meaning
- Optimization over Sparse Support-Preserving Sets: Two-Step Projection with Global Optimality Guarantees (ICML 2025 · similarity 0.86)
- Parameterized Algorithms for MILPs with Small Treedepth (AAAI 2021 · similarity 0.85)
- Scalable First-order Method for Certifying Optimal k-Sparse GLMs (ICML 2025 · similarity 0.85)
- Simple and Fast Algorithm for Binary Integer and Online Linear Programming (NeurIPS 2020 · similarity 0.85)
- Predicting Lagrangian Multipliers for Mixed Integer Linear Programs (ICML 2024 · similarity 0.84)
- Semi-infinite Nonconvex Constrained Min-Max Optimization (NeurIPS 2025 · similarity 0.83)
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.
