Warm-starting constraint generation for mixed-integer optimization: A Machine Learning approach
作者：
Highlights：
• A machine-learning-assisted tool to warm-start constraint generation is proposed.
• The proposed approach speeds up the running times in mixed-integer linear programs.
• The invariant constraint sets are used to build the initial set of constraints.
• Our approach theoretically guarantees to attain a feasible and optimal solution.
• The computational experiments show the benefits of the proposed methodology.
摘要
•A machine-learning-assisted tool to warm-start constraint generation is proposed.•The proposed approach speeds up the running times in mixed-integer linear programs.•The invariant constraint sets are used to build the initial set of constraints.•Our approach theoretically guarantees to attain a feasible and optimal solution.•The computational experiments show the benefits of the proposed methodology.
论文关键词：Mixed integer linear programming,Machine learning,Constraint generation,Warm-start,Feasibility and optimality guarantees
论文评审过程：Received 12 May 2022, Revised 14 July 2022, Accepted 27 July 2022, Available online 1 August 2022, Version of Record 12 August 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2022.109570
