Improved mind evolutionary computation for optimizations

Abstract

The paper introduced several strategies to mind evolutionary algorithm (MEC) and developed its global search ability on solving complex problems. Firstly, one region-partition initialization strategy was used to keep more potential niche into the initial population. Secondly, one self-adaptive mechanism was adopted to develop the similartaxis operator. Moreover, one global selection method based on the niche technique was applied to maximize the evolvability of groups. Finally, a series of experiments were performed on some well-known benchmark problems. The number results illustrate that improved MEC own a robust ability in global optimization and can alleviate the premature convergence validly.