Differential evolution; Diversity; Experiment; Real-world problems; Statistical comparison

In this paper, a new variant of the Differential Evolution (DE) algorithm is proposed to control the diversity of individuals in the population. The proposed approach is based on the failure of individuals in successive generations. The positions of the unsuccessful individuals are refreshed by employing the position parameters of the successful individuals from the population. Two control parameters of the proposed approach are studied to eliminate inappropriate settings. These nine variants of newly designed DE variants are compared with the classic DE algorithm when solving the set of real-world problems CEC 2011. The results show a very promising ability to solve real-world problems when the DE uses the proposed mechanism with the appropriate settings.