基于邻域抗体交叉的克隆选择算法

传统克隆选择算法由于没有交叉算子而无法在抗体间进行信息交流,进化过程中随机和盲目的变异会产生数量众多近似甚至相同的抗体,由此引起多样性丧失而导致早熟收敛。为解决该问题,提出了邻域抗体交叉的克隆选择算法,通过对匹配度近似抗体的交叉产生新抗体,在保留原有优质基因的同时,又能从其他抗体引入新信息。针对旅行商问题的仿真实验也证明了新算法的稳定性和有效性。

Adjacent Antibody Crossover Based Clonal Selection Algorithm

In traditional clonal selection algorithm（CSA）,there is no information communication during different antibodies because there is no crossover.The blind and random mutation will generate vast similar or same antibodies.Furthermore,diversity loosing usually leads to premature convergence.To solve this problem,an adjacent antibody crossover was proposed.New antibodies were generated by crossing adjacent antibodies with similar affinity.This crossover operator could not only maintain the good gene in original antibody,but it could also introduce useful information from other antibody.Simulation results on Traveling Salesman Problem（TSP） demonstrated the efficiency and robustness of the novel algorithm.