Constraint Programming for Modeling and Solving a Hybrid Flow Shop Scheduling Problem

Abstract

As an extension of a flow shop scheduling problem, hybrid flow shop scheduling problems (HFSP) have been widely concerned. Their characteristics are that every stage has parallel machines, and every job has more complicated production routes than a classical flow shop problem. Currently, most research about HFSP is based on meta-heuristic algorithms, especially evolutionary algorithms. In this article, we provide new models and solution methods based on constraint programming (CP). According to our experiments conducted on benchmark datasets, CP shows great performance in comparison with other competitive solution methods. It renews the best-found solutions of some benchmark instances. For the instances that cannot be solved exactly, it can provide a high-accuracy feasible solution as an upper bound and a relaxed infeasible solution as a lower bound.