Abstract | ||
---|---|---|
•We propose an efficient and generalizable learning based cut selection policy for tackling combinatorial optimization problems.•We present a novel cut ranking formulation in the context of multiple instance learning.•Experiments demonstrate that cut ranking is superior to other manual heuristics and can generalize to problems of different properties•Experiments on real-world product planning problems of an industrial MIP solver demonstrate that cut ranking can significantly improve the efficiency. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1016/j.patcog.2021.108353 | Pattern Recognition |
Keywords | DocType | Volume |
Mixed-Integer programming,Cutting plane,Multiple instance learning,Generalization ability | Journal | 123 |
Issue | ISSN | Citations |
1 | 0031-3203 | 0 |
PageRank | References | Authors |
0.34 | 0 | 9 |
Name | Order | Citations | PageRank |
---|---|---|---|
Zeren Huang | 1 | 0 | 0.68 |
Kerong Wang | 2 | 0 | 0.34 |
Furui Liu | 3 | 232 | 15.41 |
Hui-Ling Zhen | 4 | 5 | 1.74 |
Weinan Zhang | 5 | 1228 | 97.24 |
Mingxuan Yuan | 6 | 310 | 28.34 |
Jianye Hao | 7 | 0 | 0.34 |
Yong Yu | 8 | 7637 | 380.66 |
Jun Wang | 9 | 2514 | 138.37 |