分级变异动态克隆选择算法.pdfVIP

第22 卷第6 期 控 制 与 决 策 2007 年6 月 V ol . 22 No . 6 Control and Decision Jun . 2007 : 100 10920 ( 2007) 06060 805 a b a a 胡江强, 郭 晨 , 尹建川, 李铁山 ( a. , b . , 116026) : 基于浮点数编码, 提出一种分级变异的动态免疫克隆选择优化 法. 根据抗体的亲和度将种群分解为3 个子 种群, 分配以不同的搜索任务, 实施不同的变异策略. 在进化过程中动态改变种群规模克隆规模和变异参数, 从而加 快了全局搜索速度, 提高了局部搜索精度. 对5 个复杂函数的优化仿真实验表明了该 法的有效性. : 浮点数编码; 克隆选择; 进化 法; 动态参数; 函数优化 : T P 18 : A Dynamic clonal selection algorithm with classified mutation a b a a H UJ iangqiang , G UO Chen , YI N J ianchuan , LI T ieshan ( a. Colleg e o Nav ig ation , b . Co llege o Automat ion and Electr ical Eng ineering , Dalian M ar itime U niver sity, D alian 116026, China. Co rr espondent : HU J iangqiang, Email : huj iangqiang@ 126 . com) Abstract: A dynamic immune clonal selectio n alg or ithm w it h classi ied mutation is pro posed based on lo ating point coding . T o speed up the global search and impro ve the local co nv erg ence pr ecision , the ollow ing t w o main strat eg ies ar e intro duced . Accor ding to the antibody a inity in r elation to the antig en , the antibody po pulat ion is deco mpo sed into sev eral subsets, and they ar e submitted to respective mutation pro cesses o r t heir di er ent g iven tasks . T hen, the population size, the clone size and the mutation parameters are dynamically chang ed w ith evo lutio n processing. T he pro po sed algo rithm is u sed to optimize 5 co mplex unct ions or t esting and the results show its e ect iveness . Key words: Floating point number code; Clonal selectio n; Evo lutio nar y algo rit hm ; Dynamic parameter s;