拉伸技术克隆选择算法多模态函数优化研究  

Stretching Clonal Selection Algorithm for Multi-modal Function Optimization

作　　者：洪露[1] 龚成龙[1] 王经卓[1] 纪志成[2] 

机构地区：[1]淮海工学院电子工程学院,江苏连云港222005 [2]江南大学物联网工程学院,江苏无锡214122

出　　处：《小型微型计算机系统》2014年第5期1151-1154,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61174013)资助;江苏高校优势学科建设工程项目资助

摘　　要：为了提高克隆选择算法对复杂多模态函数优化问题的全局最优解搜索能力,基于"Stretching"拉伸技术提出了一种拉伸克隆选择算法(SCSA),该技术能在算法的搜索过程中,不断缩小目标函数局部极值点的搜索范围,从而提高算法的优化效率.为增加算法种群的多样性及提高算法的全局收敛性,算法中引入了混沌变异机制和基于抗体的浓度及亲和度差的选择机制.多模态函数优化实验结果表明,基于该技术的SCSA算法相比传统的人工免疫算法能有效地抑制早熟收敛,具有更好的收敛速度和精度,是一种有效的多模态函数优化算法.In order to enhance the global optimal search ability of clone selection algorithm in solving complicated multi-modal func- tion optimization problem, a new clone selection algorithm （ SCSA ） based on the mechanism of ＂Stretching＂ technique is proposed in the paper. The algorithm equipped with this technique can narrow the range of extreme value of the objective function, and improve the optimization efficiency of the algorithm. To improve diversity of the population and global convergence performance, the chaos mutation mechanism and selection mechanism based on the density and fitness vector of the antibody are introduced in the algorithm. Simulation results indicate that the SCSA equipped with the ＂Stretching＂ technique can inhibit prematurity and has better convergence speed and precision compared with traditional artificial immune algorithms, and is an effective multi-modal function optimization algo- rithm.

关 键 词：多模态函数优化 克隆选择算法 拉伸技术 函数优化问题 人工免疫算法 全局最优解 局部极值点 全局收敛性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]