An Effective Meta-Heuristic for No-Wait Job Shops to Minimize Makespan

The no-wait job shop problem that exists with makespan minimization is well known to be a strongly NP-hard problem. In this paper, the properties of the problem are analyzed according to its characteristics. The problem is remodeled based on the introduced time difference. A traditional framework is adopted by decomposing the problem into two subproblems: the sequencing and the timetabling problems. An efficient Shift Penalty-Based Timetabling method is proposed, which constructs two initial timetables from time difference-based sets and improves them by an investigated timetable tightening method. A modified complete local search with memory is presented for the sequencing problem. The whole algorithm is tested on benchmark instances and compared with the two best existing algorithms. Computational results show that the proposed algorithm performs well on both effectiveness and efficiency.