协同人工免疫计算模型的研究  

A Study on a Coordinative Immune-Computing Model

作　　者：王磊[1] 刘小勇[1] 

机构地区：[1]西安理工大学计算机科学与工程学院,陕西西安710048

出　　处：《电子学报》2009年第8期1739-1745,共7页Acta Electronica Sinica

基　　金：国家自然科学基金(No.60603026)

摘　　要：针对现有人工免疫模型在计算效能方面仍然存在的一些问题,本文借鉴生命科学中协同进化的概念与方式,探讨了一种协同人工免疫计算模型的实现方法,旨在改善当前该类计算模型中的不足.其主要思想在于分析种群进化过程中,在免疫疫苗亦随之进化的情况下,如何建立疫苗库与种群间的协同进化机制以提高算法全局搜索最优解或者满意解的概率.通过理论分析和针对多组TSP问题的仿真计算,结果表明该模型在搜索最优解或满意解均优于传统的遗传算法,同时在寻优效率上有较大提升.For the limited computing efficiency of existing artificial immune models,a novel method is proposed to improve their searching capabilities, which makes use of some coordinative mechanisms with referring to such kind of models or concepts in natural world. With regarded to the fact that a vaccine itself evolves along with population＇s evolution, this method aims at setting up a coordinative relation between the population and the vaccination-base during the whole evolutionary process, in order to raise the probability with which the algorithm finds the optimum or a satisfied solution. Both theoretical analysis and the simulation on multi examples of TSP problem, show that this model appears batter than traditional genetic algorithms, and the searching efficiency for the globally optimum is greatly improved as well.

关 键 词：智能计算 免疫算法 协同进化 TSP问题 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]