约束优化问题的免疫混沌算法

Immune chaotic algorithm for constrained optimization problems

结合免疫算法极强的全局搜索能力以及混沌优化方法适合局部搜索的特点,提出了一种新的免疫混沌算法.从一组可行解出发,采用免疫算法通过克隆选择、克隆扩增、高频变异和审查形成记忆细胞,并将其作为全局近似最优解,然后采用混沌优化方法按照混沌运动规律在近似最优解的邻域内进行局部搜索并审查,从而获得全局精确最优解.审查过程包含了对约束条件的处理,即对新产生的候选解进行审查,保留满足约束条件的可行解.利用该算法对几个经典约束优化问题进行了仿真测试,与以往方法相比获得了更优的结果,表明该算法是一种解决约束优化问题的有效方法.

A novel immune chaotic algorithm （ICA） was proposed to combine the global exploration capability of immune algorithm （IA） and the local exploitation capability of chaotic optimization （CO）. A group of feasible solutions was set firstly. During the course of optimization, the memory cells, namely the global approximate optimums, were obtained by IA with clonal selection, clonal proliferation, hypermutation and censoring steps, and then the global accurate optimums were reached by using censoring and CO, which locally searched the neighborhood of the global approximate optimums according to the rules of chaotic motion. The censoring process consisted of handling constrains. It censored the newcomers, and only the feasible ones were reserved. The test results of several classical COPs using ICA proved the effectiveness of ICA for constrained optimization problems.