A Local Search Method with Built-in Genetic Manipulation for Makespan Minimization in Job-shops that Considers Automatic Processing during Breaks

Abstract

Even though most machines installed in job-shops are equipped with automatic processing functions, there are cases in which manual processing by workers is also partially required. The working hours for these workers include break times. Manual processing cannot be performed during a break, but automatic processing can continue. A local search method incorporating genetic operations focusing on a block on the critical path is proposed to obtain the shortest makespan schedule in such an environment. In addition, this paper develops a mixed integer programming model (MIP) for exact optimization. In numerical experiments, the developed MIP obtained the optimal solution in one problem instance of ten machines and ten jobs. The proposed method found one optimal solution in ten trials. For problem instances of 15 machines and 15 jobs, the MIP model could not find optimal solutions in all instances within 24 hours of computation. A genetic algorithm (GA)-based optimization was also applied to these ten instances. The proposed method found better solutions than the best solutions by MIP in six instances by conducting ten trials. The proposed method was also superior to the GA-based approximate optimization method.