RL-SPH: Learning to Achieve Feasible Solutions for Integer Linear Programs
About
Primal heuristics play a crucial role in quickly finding feasible solutions for NP-hard integer linear programming (ILP). Although $\textit{end-to-end learning}$-based primal heuristics (E2EPH) have recently been proposed, they are typically unable to independently generate feasible solutions. To address this challenge, we propose RL-SPH, a novel reinforcement learning-based start primal heuristic capable of independently generating feasible solutions, even for ILP involving non-binary integers. Empirically, RL-SPH rapidly obtains high-quality feasible solutions with a 100% feasibility rate, achieving on average a 28.6$\times$ lower primal gap and a 2.6$\times$ lower primal integral compared to existing start primal heuristics.
Tae-Hoon Lee, Min-Soo Kim• 2024
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Integer Linear Programming | MIPLIB | Incumbent Value-9.35e+3 | 34 | |
| Integer Linear Programming | IS | FR100 | 14 | |
| Integer Linear Programming | SC | Feasibility Rate (FR)100 | 13 | |
| Integer Linear Programming Solving | CA | FR (%)100 | 8 | |
| Integer Linear Programming | NBI | FR100 | 6 | |
| Integer Linear Programming | MVC | Feasibility Rate (FR)100 | 6 | |
| Integer Linear Programming | CA | Feasibility Rate (FR)100 | 6 | |
| Integer Linear Programming | IS | FR100 | 3 |
Showing 8 of 8 rows