GraphBU: MILP Instance Generation with Graph-Native Block Units
About
Mixed-integer linear programming (MILP) instances used for solver development are hard to obtain when models come from private or application-specific pipelines. A generator must keep the structure that solvers and learned policies rely on. Existing general generators usually choose their generation unit from a formulation template, summary statistics, local graph edits, or blocks found after recombination. These units do not explicitly record how a local part of the MILP is coupled to the rest of the instance. We propose GraphBU, a graph-native generator whose basic unit is a local subproblem plus its interface. The method promotes coupling nodes into master constraints or boundary variables and uses the resulting block units for compatibility-checked replacement. The analysis focuses on the properties needed by this construction: promotion separates interfaces, replacement can preserve feasibility under an interface-slack condition, and the graph construction is invariant to row-column permutations. On MILP instances generation, this unit keeps graph statistics close to the source family, preserves feasibility on most datasets, and improves downstream Predict-and-Search training. Genrated by GraphBU, The average graph-statistical similarity was approximately 0.934, the average feasibility was approximately 96.7%, and the average increase in the main index of downstream PS was approximately 8.0%.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| MILP Instance Generation | CA (Combinatorial Auction) | Similarity0.989 | 4 | |
| MILP Instance Generation | FA (Facilities Allocation) | Solution Similarity81.3 | 4 | |
| MILP Instance Generation | IP (Independet Set) | Similarity0.94 | 4 | |
| MILP Instance Generation | WA (Workload Assignment) | Similarity99.5 | 4 | |
| Predict-and-search | CA original (test) | Object Count9.75e+4 | 4 | |
| Predict-and-search | FA original (test) | Time3.61 | 4 | |
| Predict-and-search | IP original (test) | Obj. Score10.778 | 4 | |
| Predict-and-search | WA original (test) | Objective Score699.3 | 4 |