基于混沌克隆的混杂多目标免疫优化算法

针对遗传算法求解高维多目标问题时易于早熟、陷入局部搜索等缺点.本文借鉴生物免疫系统的混沌现象及其自适应性和并行性,提出一种混杂多目标免疫优化算法(HMIOA).算法基于混沌映射产生初始抗体群并利用混沌规律克隆优秀抗体;根据抗体的被控度和抗体间的拥挤距离设计抗体的亲和力;算法在进化过程中对不同子群采取不同突变方式;借助Average linkage聚类法更新记忆细胞等.数值实验中,选取两种著名的多目标进化算法和一种克隆选择算法,应用于4种不同类型的高维多目标优化问题,比较结果表明:HMIOA所获Pareto有效面较大的控制其他算法所获Pareto有效面,且有效面的分布较均匀.

A Hybrid Multiobjective Immune Optimization Algorithm Based on Chaos Clone

Multiobjective immune optimization algorithm, based on chaos mechanism and adaptability and parallelism of the biological immune system, is proposed to solve multiobjective optimization problems in order to overcome the shortages of premature convergence and the poor local searching performance. In designs of the algorithm, a chaos antibody population are generated by chaos mapping;the excellent antibodies are cloned by chaos mechanism;the scheme of antibody affinity is designed by means of the strength and crowding distance of antibody;the subpopulation divided is mutated by the different methods, respectively; the average linkage method are used to update memory cells. The proposed algorithm is compared against two popular multiobiective evolutionary algorithms and a clonal selection algorithm for multiobjective optimization by means of five different kinds of multiobjective benchmark problems. Numerical experiments illustrate that the proposed algorithm is promising and competitive to the compared algorithms in solving multiobjective optimization problems with high-dimension.