An improved strategy based on center of solutions for differential evolution algorithm

This paper proposed center of solutions to improve the solutions and convergence speed of Differential Evolution (DE) algorithm. Center of mass has been used in many studies and it is assumed that each particle has the same mass to be regarded as mean vector of target vector. In our method, we regard the solutions as mass of each particle and calculates the center of solutions. We handle the center of solutions as one element in mutation strategy to lead the direction of evolution. The performance of this strategy is reported and proves it can improve performance of conventional DE.