Permutation flow-shop scheduling using a hybrid differential evolution algorithm

Abstract

In this paper, a hybrid differential evolution algorithm is developed for the permutation flow-shop scheduling problem to minimise the makespan. The iterative evolution procedure of the algorithm includes a DE-based global search for the population and an insert-based local search for the best individual. Experiments are performed, with the goal to discuss the impacts of different parameter values and the local search on the performance of the algorithm. The results reveal that choosing an appropriate CR is crucial on the global search performance for medium or larger-sized problems, and local search procedure brings the improvement in the optimisation ability, but leads to inevitable increase of computational time. Our future work is to exploit a more effective local search, which is capable of ensuring great optimisation power with reasonable computation time.