Abstract:A mixed-integer programming (MIP) problem contains not only constraints but also integer restrictions. Integer restrictions divide the feasible region defined by constraints into multiple discontinuous feasible parts with different sizes. Several popular methods (e.g., rounding and truncation) have been proposed to deal with integer restrictions. Although it is easy for these methods to generate an integer, they tend to converge to an integer which is located in a feasible part with a big size. If the optimal solution is not in this feasible part, they are very likely to converge to a local optimal solution due to the loss of diversity of the population. To overcome this shortcoming, a biobjective optimization-based two-phase method is proposed in this article. In the first phase, a measure function is designed to compute the degree that a solution violates integer restrictions. By employing this measure function as the second objective function and removing integer restrictions, a MIP problem is transformed into a constrained biobjective optimization problem (CBOP). It can be proven that the Pareto optimal solution of the transformed CBOP which satisfies integer restrictions is the optimal solution of the original MIP problem. To solve the transformed CBOP, a new comparison rule is designed. After the first phase, the population can approach the Pareto optimal solution which satisfies integer restrictions. Then, the second phase is implemented to enhance the convergence precision and obtain the optimal solution. In addition, we design 12 test problems to verify the effectiveness of the proposed method. The results demonstrate that the proposed method shows better performance against five state-of-the-art evolutionary algorithms for MIP.

Improving Primal Heuristics for Mixed Integer Programming Problems Based on Problem Reduction: A Learning-based Approach

Optimality-based domain reduction for inequality-constrained NLP and MINLP problems

An Enhanced Branch-and-bound Algorithm for Bilevel Integer Linear Programming.

A Parallel Method With Hybrid Algorithms For Mixed Integer Nonlinear Programming

Nested Heuristic and Gradient-Based Method for Generalized Disjunctive Programming

L2P-MIP: Learning to Presolve for Mixed Integer Programming

Learning to Branch in Mixed Integer Programming

Reinforcement Learning for Node Selection in Mixed Integer Programming

Improving the filtering of Branch-And-Bound MDD solver (extended)

An Efficient Strategy for the Activation of MIP Relaxations in a Multicore Global MINLP Solver

Pre-trained Mixed Integer Optimization through Multi-variable Cardinality Branching

A Biobjective Perspective for Mixed-Integer Programming

Towards Imitation Learning to Branch for MIP: A Hybrid Reinforcement Learning Based Sample Augmentation Approach

Learning a Large Neighborhood Search Algorithm for Mixed Integer Programs

Deep learning enhanced mixed integer optimization: Learning to reduce model dimensionality

An efficient branch and bound reduction algorithm for globally solving linear fractional programming problems

Learning to Solve Bilevel Programs with Binary Tender

Presolve Reductions in Mixed Integer Programming

Threshold-aware Learning to Generate Feasible Solutions for Mixed Integer Programs

Self-Paced Learning Based Graph Convolutional Neural Network for Mixed Integer Programming (student Abstract).