摘要
为了提高免疫算法的求解性能,在设计优良基因片的提取与保护方案、进化的抗体经解码后作为约束子函数的实参以剔除不满足约束条件的抗体以及在抗体选择概率中利用抗体对抗原适应度概率、相似度概率以及退火概率3个方面对传统免疫算法进行了改进;在对改进的免疫算法参数分析的基础上给出了其参数的自适应调整方法以及改进后的算法步骤;并通过设计实例对采用改进的参数自适应调整的免疫算法和改进的固定参数的免疫算法在最优解精度、平均值、标准偏差以及平均迭代时间等方面作对比;比较结果表明:改进的参数自适应调整的免疫算法求得的最优解比改进的固定参数的免疫算法获得的解以及传统文献获得的最优解在精度上分别提高了7%及10%,同时在优化平均时间上分别减少了14s及18s。
In order to improve the performance of immune algorithm （IA）, the traditional immune algorithm is modified in three aspects： a good extraction and gene chip proteetion scheme is designed, the evolved antibody after decoding is used as the arguments of constrained subroutines to remove the antibody which does not meet the constraints, and the information of the probability of antibody--antigen fitness, the probability of similarity and probability of annealing is utilized to compute the selection probability of antibody. Based on the analysis of parameters of the immune algorithm, the adaptive algorithm of parameter adjusting and the steps of improved algorithm are given, The im- proved immune algorithm based on self--adaptive parameter adjusting （ISAPA--IA） is compared with the improved immune algorithm based on fixed parameter optimization （IFP--IA） in terms of optimal solution aeeuraey, mean, standard deviation and the average iteration time in an industrial design example. Comparison results show that the optimal solution of the proposed ISAPA--IA method has accuracy 7 % higher than that of IFP--IA and 10% higher than that of traditional method in the literature. At the same time, the average iteration time of ISAPA --IA is 14 seconds less than that of IFP--IA and 18 seconds less than that of traditional method in the literature.
出处
《计算机测量与控制》
北大核心
2013年第5期1297-1300,共4页
Computer Measurement &Control
基金
河南省自然科学基金项目(122300410310)
河南省高等教育教学改革研究省级项目(2012SJGLX205)
关键词
改进的免疫算法
自适应参数调整
基因片
优化方法
improved immune algorithm
self-- adaptive parameter adjusting
gene segment
optimization design
