所属 広島修道大学 商学部 職種 教授
|Converging and Moving Detection Using an Outward Vector Rate for Adaptive Differential Evolution
|Tetsuyuki Takahama and Setsuko Sakai
|The performance of Differential Evolution (DE) is affected by algorithm parameters, mutation strategies and so on. One of the most successful studies on controlling the parameters in DE is JADE (adaptive DE with optional external archive). In this study, we propose a method to detect that the population is converging or moving. The vectors from parent to child are classified into inward vectors and outward vectors, and the rate of the outward vectors is used to determine the search state. In order to improve search efficiency of JADE, if the population is converging, an algorithm parameter is adjusted to enhance the convergence. If the population is moving, the parameter is adjusted to enhance the movement. Also, the moving direction of the population is determined by the averaged outward vectors and the search in the direction is strengthened. The advantage of JADE with the proposed method is shown by solving thirteen benchmark problems.