Flexible job shop scheduling using improved cuckoo search algorithm by Markov clustering and Levy flight

Considering the high complexity of scheduling problems, classic approaches fail to find the solution efficiently. Therefore, meta-heuristic algorithms are used for this purpose. In this paper, Cuckoo optimization algorithm (COA) is used as one of the novel and most effective evolutionary optimization algorithms for flexible job shop scheduling. In the proposed approach, for better solutions, the initial population is determined using NEH-D algorithm, which considers the completion time minimization of each job. Then active machines are grouped using Markov clustering, so that the assigned machine is chosen from the active ones, hoping that the possible solutions of COA are bounded and the execution speed of the algorithm is increased. Finally, COA is applied for job-machine assignment and Levy flight is used to improve the global search of the algorithm. The proposed approach is evaluated on standard datasets such as Kacem, Brandimarte and other related data. The experimental results show that the proposed algorithm is capable of finding the final solution with lower computational complexity and has higher convergence rate.