一种新的软件测试数据的进化生成方法  

A New Software Method to Design Test Data

作　　者：姬秀荔[1] 李爱玲[1] 孙玉强[2] 

机构地区：[1]安阳工学院计算机工程系,河南安阳455000 [2]河南师范大学算机工程系,河南新乡453007

出　　处：《微计算机信息》2007年第21期249-251,共3页Control & Automation

基　　金：河南省自然科学基金资助(0324220079)

摘　　要：测试数据自动生成对降低软件开发成本是非常重要的。本文首先提出了基于程序结构和程序规范相结合来设计适应度函数的新方法:分别设计基于程序结构和基于程序规范的适值函数,再将二者结合作为最终的适应度函数值;然后,在采用上述适应度函数定义的基础上,构造了一种软件测试数据的进化生成方法。Automated generation of test data is very important to reduce the cost of software development. Genetic Algorithm is a well-used method to generate the test data and its key kernel lies in the design of fitness function. This article proposed a new method to design fitness function based on program structure and program criterion. The concrete way is designing fitness function based on the program structure and program criterion respectively, and then combining the twos as final fitness function. The experimental results show that this method can produce test data set with high program coverage level and enhance the ability of detecting error.

关 键 词：程序结构 程序规范 测试数据 适应度函数 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]