模因内三角概率选择混合蛙跳算法  

Meme triangular probability distribution shuffled frog-leaping algorithm

作　　者：朱光宇[1] 

机构地区：[1]福州大学机械工程及自动化学院,福建福州350002

出　　处：《计算机集成制造系统》2009年第10期1979-1985,共7页Computer Integrated Manufacturing Systems

基　　金：福建省高等学校新世纪优秀人才支持计划资助项目(XSJRC2007-08);福建省科技厅重点资助项目(2009H0025);福建省自然科学基金(2009J01246)~~

摘　　要：为优化拱架型多头贴片机贴装顺序,建立了贴装顺序数学模型。按照三角概率分布选择多只蛙实现模因进化的策略,改进混合蛙跳算法,构建模因内三角概率选择混合蛙跳算法,解决了贴装顺序优化问题。实验表明,新算法具有比混合蛙跳算法更高的收敛率,比基因遗传算法具有更好的准确性及更高的收敛率,有效解决了贴装顺序优化问题。通过算法参数分析得出,随着蛙群个体数量的增加,算法能够得到更优解;模因组数对算法的影响小于蛙群的个体数量;对应不同蛙群的个体数量,有一个与之对应的模因组数,使算法相对稳定。To optimize the component placement sequence of gantry surface mounting machine, a mathematical model of component sequence was established, and the triangular probability distribution shuffled frog-leaping algorithm was proposed with the strategy of selecting several frogs for meme evolution. These frogs were selected according to triangular probability distribution. The presented algorithm was used to resolve the optimization problem of component placement sequence. Experiment results indicated that the improved algorithm outperformed shuffled frog-leaping algorithm in convergence ratio and Genetic Algorithm （GA） in accuracy and convergence ratio. Known from parameter analysis that better solution could be obtained when the frog population number increased and the effect of memeplex number was weaker comparing to frog population number. There was a memeplex number corresponding to different frog population number which ensured the algorithm to be comparatively stable.

关 键 词：组合优化 贴装顺序优化 三角概率分布 混合蛙跳算法 数学模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]