基于演化算法的软件形式化开发部件选取方法  

Selection of Software Formal Development Components Based on Evolutionary Algorithms

作　　者：谭棉 冯夫健 TAN Mian;FENG Fu-jian(Teaching Administration Office,Guizhou Minzu University,Guiyang 550025,China)

机构地区：[1]贵州民族大学教务处,贵州贵阳550025

出　　处：《软件导刊》2018年第7期64-66,共3页Software Guide

基　　金：国家自然科学基金项目(61263034);贵州省教育厅青年科技人才成长项目(黔教合KY字[2016]165)

摘　　要：形式化方法在软件开发过程中一直扮演着重要角色。由于其建立在严格的数学基础之上,因此开发成本较高。为了解决开发成本问题,将软件开发整体分为形式化开发集合类和非形式化开发集合类。在此基础上,开发部件采用二进制串描述并构建目标函数模型,通过演化算法对模型进行优化求解,判定不同的部件开发集合类归属。选用(1+1)EA作为研究对象验证开发部件选取方法的可行性。采用遗传算法、粒子群算法、二进制差分演化算法进行验证分析。结果表明,该分类方法可作为演化算法在软件开发部件选取方面的一种理论工具。Formal development has played an important role in software development,because it is based on rigorous mathematics,development costs are often more expensive.In order to solve the development cost problem,this paper divides the software development into two sets of the class：formal development collection class and unformal development collection class.On this basis binary string description is used to describe the development components and construct the objective model.The model is optimized by the evolutionary algorithm to determine the ownership of different development components.In order to verify the feasibility of the development component selection,we select（1＋1）EA as the research object,and adopts the genetic algorithm,particle swarm algorithm and binary difference evolution algorithm to carry out experimental verification analysis.The experimental results show that the proposed classification method can be used as theoretical tool in selection of software development components.

关 键 词：形式化方法 演化算法 目标函数 集合类 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]