Abstract:

An improved differential evolution algorithm based on center mutation and adaptive crossover probability, named center mutation-based differential evolution (CMDE), is proposed for the high-dimension complex optimization problem. The CMDE has two improvements: a mutation operator composed of the modified base vector and differential vectors, the former is set as the center of all target vectors, and the latter is determined by the function fitness value of three randomly selected vectors; an adaptive crossover probability process according to the distribution of the function fitness value. Numerical tests on several benchmark functions are conducted. The results show the CMDE can improve the convergence speed and has higher optimizing precision and searching ability.