用相似系数实现候选软件架构的量化决策  

The Quantitative Decision Implementation on Candidate Architectures with Similar Coefficient

作　　者：冯富霞[1] 李森贵 

机构地区：[1]安徽工程大学计算机与信息学院,安徽芜湖241000 [2]芜湖莫森泰克汽车科技有限公司,安徽芜湖241000

出　　处：《安徽工程大学学报》2017年第1期68-71,76,共5页Journal of Anhui Polytechnic University

基　　金：安徽工程大学计算机应用技术重点实验室开放基金资助项目(JSJKF201604)

摘　　要：基于智能算法优化软件架构的过程已实现自动优化,架构师输入原始设计架构即可得到一个优化的架构集,但是最后选取需要的架构仍然缺乏量化的决策方法,主要依赖于架构师的主观经验.架构的非功能指标具有不确定的自然属性,基于PCM的智能优化预测算法实现了影响非功能指标因素的自由度空间的覆盖,成功处理了不确定性.根据PCM优化结果的数据,基于Fuzzy数学相似矩阵的相似系数理论,提出了一种有效且新颖的架构选择决策方法.The software architecture optimization process based on intelligence algorithm is already automated,if the architect inputs the initial design architecture,an optimized architecture set will be given.But there is a lack of quantitative decision methods on the architecture selection,and the decision mainly depends on architect＇s subjective experience.There is an uncertainty of the architecture non-function indexes,based on PCM,the intelligence optimization algorithm may cover all the space of architecture nonfunction indexes,and it can handle the uncertainty successfully.Based on the PCM optimization results and similarity coefficient theory about similar matrixes in Fuzzy mathematic,the paper provides an efficient and novel decision method.

关 键 词：架构决策方法 不确定性 相似系数 PCM 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]