A Hybrid Immune Evolutionary Algorithm for Global Optimization Search

Abstract

Optimization is an important issue in many kinds of application areas, whereas expediting optimizing process and jumping out of the local optimums are keys in optimization researches. This article presents an immune evolutionary algorithm for optimizing search in continuous space. The proposed algorithm adopts immune network model & evolutionary strategy, adjusts self-adaptively the metrics of evolutionary space on immune affinity, such as the evolutionary steps and directions. The algorithm realizes search diversity by restraining most individuals within one immune shape-space measured in restrain radius. The experimental results on multimodal functions show that the proposed algorithm got the whole optimal solutions and a lot of suboptimal ones in lesser amount of evolutionary generations and minor populations compared with the contrasted algorithms, such as CSA, GA and aiNet, and the effect of global optimizing capability are verified with excellent population diversity.