Apollo-MILP: An Alternating Prediction-Correction Neural Solving Framework for Mixed-Integer Linear Programming
About
Leveraging machine learning (ML) to predict an initial solution for mixed-integer linear programming (MILP) has gained considerable popularity in recent years. These methods predict a solution and fix a subset of variables to reduce the problem dimension. Then, they solve the reduced problem to obtain the final solutions. However, directly fixing variable values can lead to low-quality solutions or even infeasible reduced problems if the predicted solution is not accurate enough. To address this challenge, we propose an Alternating prediction-correction neural solving framework (Apollo-MILP) that can identify and select accurate and reliable predicted values to fix. In each iteration, Apollo-MILP conducts a prediction step for the unfixed variables, followed by a correction step to obtain an improved solution (called reference solution) through a trust-region search. By incorporating the predicted and reference solutions, we introduce a novel Uncertainty-based Error upper BOund (UEBO) to evaluate the uncertainty of the predicted values and fix those with high confidence. A notable feature of Apollo-MILP is the superior ability for problem reduction while preserving optimality, leading to high-quality final solutions. Experiments on commonly used benchmarks demonstrate that our proposed Apollo-MILP significantly outperforms other ML-based approaches in terms of solution quality, achieving over a 50% reduction in the solution gap.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Maximum Independent Set | Maximum Independent Set Medium (m) | Execution Time89 | 34 | |
| Combinatorial Auction | Combinatorial Auction (CA) medium-scale | Objective Value7.13e+3 | 20 | |
| Capacitated Facility Location | Capacitated Facility Location (CFL) medium-scale | Objective Value9.07e+3 | 12 | |
| Mixed Integer Linear Programming | Large-scale Set Cover (SC) (test) | Objective Value24.45 | 12 | |
| Combinatorial Auction | CA Large-scale | Objective Value1.35e+4 | 12 | |
| Maximum Independent Set | MIS Large-scale | Objective Value319.1 | 12 | |
| Mixed Integer Linear Programming | Large-scale Maximum Independent Set (MIS) (test) | Objective Value319.1 | 12 | |
| Mixed Integer Linear Programming | Large-scale Combinatorial Auction (CA) (test) | Objective Value1.35e+4 | 12 | |
| Mixed Integer Linear Programming | Large-scale Capacitated Facility Location (CFL) (test) | Objective Value1.75e+4 | 12 | |
| Set Cover | Set Cover (SC) medium-scale | Objective Value29.09 | 12 |