Differential evolution utilizing a handful top superior individuals with bionic bi-population structure for the enhancement of optimization performance

摘要

In this paper, we propose a new DE variant incorporating a new mutation strategy and a bionic bi-population structure to enhance the overall performance of the canonical DE algorithm. As it is known to all that the guidance of a handful top superior elites is usually better than the guidance of only one elite in many cases, therefore, a handful top superior individuals are employed in the newly proposed mutation strategy for individual evolution. Moreover, hybridization of two populations of a certain species can improve the adaptation to the environment, and business and information exchanges between tribes can help their economical and technological development, so a bionic bi-population structure is also employed in the new DE variant for information exchanges. Both of the two aspects make the new algorithm achieve an overall better performance, and experiment results show that a handful top superior individuals with bionic bi-population structure has better performances on optimization accuracy, convergence speed and success rate under these tested benchmark functions especially those other algorithms may be trapped into local optima.