科研之友
微信
新浪微博
Facebook
分享链接
摘要
The hybrid hierarchical strategy, which is inspired by group collaboration, has been successfully used to maintain the diversity of the population and avoid premature convergence. A modified hybrid hierarchical JAYA algorithm (HHJAYA) is proposed here to solve global optimization problems and ameliorate the optimization performance of JAYA. In this method, all individuals are randomly divided into several sub-swarms at the bottom layer, each individual is updated by using a combination of DE and an improved JAYA algorithm, and the best individuals are selected to the top layer. In order to exchange information between the top and bottom layers, the random individual of the sub-swarm at the bottom layer will be replaced by the best individual at the top layer in each iteration. After a certain number of iterations, individuals are reorganized to increase the diversity of different sub-swarms. Moreover, experimental results on the CEC2014 and CEC2017 test suits have validated the feasibility and optimization behavior of HHJAYA. Finally, HHJAYA is extended to fuzzy clustering based on entropy-like divergence-based kernel for image segmentation, and the segmentation results on different benchmark images indicate that HHJAYA has excellent performance in most cases.
