Efficient Tasks and Maintenance Scheduling in Parallel Machines With Single Server, Setup, and Unloading Times

ABSTRACT This study dealt with the scheduling problem of identical parallel machines with a single server, loading time, preventive maintenance periods, and unloading time. The loading and unloading activities are performed by the server. The main objective is minimizing makespan. Three metaheuristic algorithms known as simulated annealing (SA), tabu search (TS), and genetic algorithm (GA) have been adapted and proposed. Pilot runs with Taguchi design were used to conduct the best parameter settings of the proposed algorithms. To assess the performance of metaheuristics, a set of lower bounds is presented. Utilizing a set of test problems that were produced at random based on relevant research from works of literature, this study involves four types of instances in which the generated machines' availabilities and the jobs' processing times differ, and each case was replicated 10 times. The performance of the proposed algorithms was studied based on the effect of the number of jobs, the number of machines, varying processing times, and availability periods on the algorithm parameters. The performance was measured as the gap between the developed lower bounds and the obtained makespan; computational time was also considered. An extensive experimental study was conducted to assess the performance of the proposed procedures. The obtained results provided a strong evidence of the efficiency of the presented algorithms.