Matheuristics for the single-path design-balanced service network design problem

We introduce and study the single-path design-balanced service network design problem, where flow of a commodity must use only one path from origin to destination. We present an arc-based formulation. We propose two approaches respectively based on matheuristics: local branching and relaxation induc...

Full description

Saved in:
Bibliographic Details
Published in:Computers & operations research 2017-01, Vol.77, p.141-153
Main Authors: Li, Xiangyong, Wei, Kai, Aneja, Y.P., Tian, Peng, Cui, Youzhi
Format: Article
Language:English
Subjects:
Design-balanced constraints
Heuristic
Local branching
Mathematical analysis
Mathematical programming
Mixed integer
Network flow problem
Networks
Operations research
Origins
Relaxation induced neighborhood search
Searching
Service network design
Single-path constraints
Solvers
Strategy
Studies
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We introduce and study the single-path design-balanced service network design problem, where flow of a commodity must use only one path from origin to destination. We present an arc-based formulation. We propose two approaches respectively based on matheuristics: local branching and relaxation induced neighborhood search. These approaches hybridize metaheuristic strategies and mathematical programming, and use an idea of defining neighborhoods as small mixed integer programs (MIP), and exploring neighborhoods using MIP solvers. We also present an implementation combining these two approaches. In addition, we propose a Lagrangian heuristic to generate a starting solution. Our computational results demonstrate effectiveness of these approaches. •We introduce and study the single-path design-balanced service network design problem.•We propose three matheuristics to solve the introduced problem.•We design a Lagrangian relaxation method to get initial solution in the three methods.•Computational results show that these approaches reach good performance.
ISSN:0305-0548
1873-765X
0305-0548
DOI:10.1016/j.cor.2016.07.010