组合优化多种预测基因区域的PSO算法  

PSO Algorithm of Combining and Optimizing Multi Predicted Gene Regions

作　　者：陈雄峰[1] 刘燕[1] 

机构地区：[1]闽江学院计算机科学系,福建福州350108

出　　处：《三明学院学报》2009年第4期407-410,共4页Journal of Sanming University

基　　金：福建省教育厅A类科研项目(JA08176)

摘　　要：针对不同基因预测工具可能分别预测不同的基因区域,为充分利用各种工具的优点而求最大概率基因区域的问题,提出了使用PSO算法组合优化多种预测基因区域的算法。分析基因区域的读框限定和分段组合等特点,着重介绍适合PSO算法的基因区域表示和区域组合微粒表示的数学模型;分预测基因区域概率和组合区域概率计算两个步骤,详细叙述PSO算法目标函数的计算方法;简要描述算法的更新等式和步骤,并分析说明该算法可以克服时间复杂度和空间复杂度随着基因工具的数量指数增长的缺点。Aiming at the problem of seeking the most possible gene regions to take the advantage of each gene predicting tool when various tools might predict various gene regions respectively, this paper brings forward a PSO algorithm of combining and optimizing multi predicted gene regions. After analyzing the characteristics ofgene region include ORF restriction and subsection combination, it emphatically introduces the mathematical model of the representations of gene region and region-combination particle to fit PSO algorithm and amply depicts the computational method of this PSO＇s target function with the probability computation of predicted gene region and combined gene region. Lastly, the author briefly describes the update equation and the steps of this algorithm, and then analyzes this algorithm to explain that its time complexity and space complexity will not increase exponentially with the number of gene predicting tools.

关 键 词：PSO算法 基因区域 区域概率 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]