A New Clonal Selection Immune Algorithm with Perturbation Guiding Search and Nonuniform Hypermutation

Abstract

A new clonal selection immune algorithm with perturbation guiding search and non-uniform hypermutation (nCSIA) is proposed based on the idea of perturbed particle swarm algorithm and non-uniform mutation. The proposed algorithm proportional clones antibody based on the affinity, adaptively adjusts the searching steps of antibodies with hypermutation according to the adaptive variation rule of non-uniform mutation and chooses the promising antibody based on the affinity by clonal selection principle. In order to keep the balance of exploration/exploitation better, perturbation guiding search strategy is presented, which is actually an elitist learning mechanism and is borrowed from the perturbed particle swarm algorithm. In order to validate the effectiveness of nCSIA, comprehensive experiments and analysis are done based on fifteen unimodal or multimodal benchmark functions. Compared with standard and the recent algorithms, it indicates that the proposed algorithm is feasible, effective and has better performance in terms of convergence, accuracy and stability. More evident predominance emerges from further experimental comparisons with expanding search space and increasing dimensions.