Differential evolution based on a novel double-population strategy

Abstract

Differential evolution (DE) is a population-based stochastic search algorithm, which shows good performance when solving many optimization problems. In order to improve the performance of DE, this paper presents a new variant of DE based on a double-population strategy. The proposed approach is called DPDE, which consists of two populations. The first population focuses on original DE algorithm, and the second one concentrates on local search. To verify the performance of DPDE, ten famous benchmark functions were selected in the experiments. Simulation results show that DPDE outperforms DE and another variant of DE on most test functions.