Self-adaptive Differential Evolution and its application to job-shop scheduling

To improve the global convergence property and the avoidance premature convergence ability of differential evolution (DE), a self-adaptive differential evolution (SDE) was proposed. First, in order to simplify the difficulty of choosing suitable parameter values and improve the ability of breaking away form the local optimum, chaos theory was used to optimize the parameters of individuals in population. Second, for the sake of balancing the global search ability and local search ability of DE, a self-adaptive parameter setting strategy according to the fitness of individual was presented. At last, the SDE was applied to solving the job-shop scheduling problem. Experiment results show that the proposed method SDE is effective to avoid premature convergence and improves the global search ability remarkably.