A Global Minimization Algorithm by Removing Points Converging on the Already Known Local Minima

摘要

We propose an algorithm for solving global minimization problems with multimodal functions. Since this algorithm has a step by removing the candidates that are expected to converge on a current local minimum in each iteration, it can avoid to converge on an already known minimum in local minimizer. We show that the algorithm stops after a finite number of iteration and finds one of global minima under certain conditions. This algorithm has been tested for standard test functions, and the results show that the algorithm finds one. of global minima effectively and with a high degree of accuracy and a short computational time.