摘要
针对基因表达式程序设计(GEP)收敛速度慢、收敛后适应度不高和易陷入局部最优等缺陷,利用GEP染色体简单、线性和紧凑、易于进行遗传操作和免疫算法(Immune algorithm,IA)抗体多样性和免疫记忆等优点,提出了一种免疫基因表达式程序设计算法(Immune Gene Expression Programming,IGEP)。将免疫算法的按抗体浓度进行调节和免疫记忆的机制用于GEP的遗传算子中,收敛速度比GEP要快、收敛后适应度高且有效地克服了GEP不成熟收敛,理论证明该算法是全局收敛的。函数优化的仿真实验结果,进一步验证了该算法的性能。
In view of the disadvantages of gene expression programming, such as low convergence rate, low fitting degree and easily falling into local best, an Immune Gene Expression Programming is proposed based on the advantages of GEP,such as simplicity, linearity, compact in Chromosome, easy generic operation, variety of antibody, memory of immunity in Immune Algorithm and so forth. A new idea of the proposed algorithm is that the mechanism of adjustment according to immunity＇ s density and immunity＇ s memory is used in the generic operation of GEP. It has high convergence speed, high fitting degree and overcomes the premature convergence the Furthermore, it is theoretically proved to be overall convergence. The performance of the IGEP is proven via computer simulation for the function optimization.
出处
《计算机仿真》
CSCD
2008年第3期189-191,317,共4页
Computer Simulation
关键词
基因表达式程序设计
免疫算法
浓度
全局收敛
Gene expression programming（GEP）
Immune algorithms
Density（IA）
Overall convergence
