自适应采样免疫算法及其对随机优化问题的应用  

Adaptive Sampling Immune Algorithm and Its Application to Stochastic Optimization Problems

作　　者：涂歆[1] 

机构地区：[1]贵州大学理学院系统科学及信息技术研究所,贵州贵阳550025

出　　处：《贵州大学学报（自然科学版）》2009年第5期78-82,共5页Journal of Guizhou University:Natural Sciences

基　　金：教育部科学技术研究重点项目(208125)

摘　　要：基于体液免疫的简化机制,设计随机优化免疫算法处理高维随机优化问题.算法的关键在于设计:(1)动态分配群体中各抗体采样次数的自适应采样算法;(2)抗体进化算子;(3)进化群体和记忆集的采样数.该算法具有结构简单、灵活、易于应用的特点.数值实验及多种算法比较显示,此算法具有噪声抑制能力强和寻优速度快速等特点,对于不同噪声幅值的高维随机优化,均能在优化质量、噪声抑制、执行效率中获得满意效果.A stochastic optimization immune algorithm, suitable for stochastic optimization was proposed based on the simplified immune metaphors of humoral immune. The focus of design is concentrated on the three operators： adaptive sampling algorithm （ASA）, antibody evolution and sampling rule of population. The first allocates dynamically the sampling number of population to each antibody. The second designs immune operators to evolve usable antibodies so that they become better. The final one provides sampling numbers for the current evolution population and memory set. With the characteristics of simple structures, feasibility and convenient application, the algorithm can strongly suppress noise while searching the desired solution Experimental results and comparison with multiple relevant algorithms show that when being applied to stochastic optimization problems, it can achieve satisfactory effect with respect to optimization quality, noise suppression and performance efficiency.

关 键 词：随机优化 免疫算法 自适应采样 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]