A Hybrid Genetic Algorithm for Solving Job Shop Scheduling Problems
Job shop scheduling problems are immensely complicated problems in machine scheduling area, and they are classified as NP-hard problems. Finding optimal solutions for job shop scheduling problems with exact methods incur high cost, therefore, looking for approximate solutions with meta-heuristics are favored instead. In this paper, a hybrid framework which is based on a combination of genetic algorithm and simulated annealing is proposed in order to minimize maximum completion time i.e. makespan. In the proposed algorithm, precedence preserving order-based crossover is applied which is able to generate feasible offspring. Two types of mutation operators namely swapping and insertion mutation are used in order to maintain diversity of population and to perform intensive search. Furthermore, a new approach is applied for arranging operations on machines, which improved solution quality and decreased computational time. The proposed hybrid genetic algorithm is tested with a set of benchmarking problems, and simulation results revealed efficiency of the proposed hybrid genetic algorithm compared to conventional genetic based algorithm.