摘要
针对多元非线性Rosenbrok函数优化问题,结合生物免疫系统相关机理,设计对应的人工免疫算法对其进行求解。讨论了相似度、亲和力、群体多样性以及免疫记忆等功能的设计和实现方法,算法设计中,结合信息熵原理设计抗体相似度,并引入克隆抑制、亲和突变及募集新成员算子来保持群体多样性。最后,通过实验比较分析,该算法稳定,且具有较强的全局搜索性能,能够获得较满意的解。
This paper, in accordance with the optimization problems of the multivariate nonlinear and combined with the relative mechanisms of biological immune system Rosenbrok functions, aims to design a corresponding artificial immune algorithm to solve this fucntion. It discusses the design and realization of such functions like similarity, affinity, group diversity, and immunological memory. In the algorithm design, combined with the entropy principle, group diversity can be maintained by the introduction of cloning inhibition, affinity mutation and the collection of the new operator. Finally, by an experimental comparative analysis, this algorithm proves to be stable and has strong global search performances. Also, it can get a relatively satisfactory answer.
出处
《铜仁学院学报》
2012年第3期141-144,共4页
Journal of Tongren University
基金
铜仁学院科研启动基金项目(TS1105)
