Matheuristic for the decentralized factories scheduling problem

Abstract

• This study considers a distributed production network with parallel factories. • Jobs can be transport from an overloaded factory to a factory with a lower workload. • The problem is formulated as mixed integer linear programming. • A matheuristic is proposed to minimize the makespan. • Guiding principles are used by ordering the neighborhood structures. In this study, a novel mixed integer linear programming (MILP) model is developed for the decentralized factories scheduling problem, where a set of transporters is used for shipping goods among parallel factories to minimize the production costs over all of the factories. Due to its typical features, such as multiple heterogeneous factories and transportation times, this problem is extremely difficult to solve, especially for large-scale problems. In order to address this difficulty, the main aim of this study is to develop a new solution algorithm based on the interoperation of metaheuristics and mathematical programming techniques to minimize the production costs for jobs. The algorithm comprises an electromagnetism-like algorithm and variable neighborhood search. In this hybridization based on MILP relaxation, the guiding principle involves the ordering of neighborhood structures. The results obtained by the proposed algorithm and MILP are analyzed and compared for various test problems.