基于区间相容技术与GA的测试数据自动生成方法  

An Arithmetic for Automated Test Data Generation Based on Consistency Techniques & GA

作　　者：张毅坤[1] 赵明 张保卫[1] 崔杜武[1] 

机构地区：[1]西安理工大学计算机科学与工程学院,陕西西安710048 [2]国际软件自动化上海公司,上海200040

出　　处：《西安理工大学学报》2006年第4期350-354,共5页Journal of Xi'an University of Technology

基　　金：陕西省自然科学基金资助项目(2005F07);陕西省教育厅科研基金资助项目(06JK230)

摘　　要：针对测试数据自动生成完全依赖约束集求解问题(Constraint Solving Problem,CSP)进行求解会导致耗时较大甚至求解不出最终测试用例,以及采用动态GA算法又无法确定变量的最初论域空间,首次将基于CSP求解与GA的动态算法进行了有机结合,摒弃了二者固有的缺陷,吸取了静态算法对变量论域空间削减速度快的优点。采用eBox区间相容削减标准,过滤变量的论域空间,并在经过削减的空间上采用GA搜索算法自动产生测试数据,应用常变量的逆向推导技术、表达式直接遗传等技术,大幅度地提高了测试数据的生成速度。With an aim at the constraint solving problem（CSP） of the automated test data generation（ATDA）,seeking for solution could result in the fact that the final test case can not be derived with much time consumption, and also,using dynamic GA can not determine the initial domain of variables. Accordingly, the organic combination is first made of the solution based on CSP with GA in such a way that the inhereted defects of the both are discarded and that the ad- vantages of the static algorithm to eliminating the domain of variable at a fast speed can be absorbed, eBox-consistency domain eliminating standared is used to filter the domain of variables, and then GA searching algorithm via the eliminated space is adopted to generate automated test data. Also,the constant variable backward derivation tech and the expression direct genetic tech are used to improve the speed of ATDG by a wide margin.

关 键 词：测试数据自动生成 约束集求解问题 eBox相容 GA 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]