基于搜索的面向路径字符串测试数据自动生成方法  

Search-Based Automatic Path Test Generation Method for Character String Data

作　　者：赵瑞莲[1,2] 

机构地区：[1]北京化工大学计算机科学系,北京100029 [2]中国科学院软件研究所计算机科学国家重点实验室,北京100080

出　　处：《计算机辅助设计与图形学学报》2008年第5期671-677,共7页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(60473032);北京市自然科学基金(4072021);中国科学院软件研究所计算机科学国家重点实验室开放课题基金(SYSKF0605)

摘　　要：针对字符串测试数据自动生成,讨论了字符串间的距离,将不满足路径条件的字符串谓词表示成一个实值目标函数;利用快速下降搜索算法实施目标函数极小化,实现了基于搜索的面向路径字符串测试数据自动生成方法;探讨了其测试数据生成效率与初始输入、路径处理顺序之间的关系,并与遗传算法等几种算法进行了比较.实验结果表明:该方法是一种更经济有效的测试数据生成方法.Aiming at the automatic test generation for character string data, this paper defines the distance between two strings. A real-valued objective function is constructed with respect to the string predicate which does not meet a given path condition, and its minimum is computed by using a speedy descent search algorithm. As a result, path-oriented string test data can be automatically generated. In addition, this paper investigates the relations among test generation effectiveness and original inputs as well as path orders processed, and the test generation effectiveness is compared with other three algorithms such as the genetic algorithm. The experimental results illustrated that this approach is more economical and effective.

关 键 词：路径测试数据生成 字符串数据 字符串间距离 快速下降搜索算法 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]