基于优化模型的类级测试数据自动生成研究  

The Automatic Generation of Class Test Data Based on Optimization Template

作　　者：屈迟文[1] 

机构地区：[1]百色学院数学与计算机信息工程系,广西533000

出　　处：《科学技术与工程》2012年第30期7920-7923,7930,共5页Science Technology and Engineering

基　　金：百色学院一般项目(2010KB16)资助

摘　　要：研究了一种基于优化使用模型的类级测试用例的自动生成方法。首先使用基于约束的非线性规划问题的求解方法获得Markov链使用模型的状态转移概率,建立优化的使用模型;然后以使用模型的概率转移为约束采用轮盘赌算法自动生成测试用例;最后以Euclidean Distane方法度量测试的充分性。实验表明此方法能够提高测试自动化水平,提高软件的可靠性。A improved auto-producing mechanism for class_level test data is introduced based on a using tem-plate. Firstly a better Markov train using template is built up by improving the probability of state transformation through the solution of constrain nonlinear programming problem. And then based on the constrain of using template state transformation probability , the testing cases are created randomly and automatically. Finally the testing cover-age is measured by Euclidean Distane Method. Test results indicate that, this proposal can improve the capability of auto software testing and software reliability.

关 键 词：类级测试数据 MARKOV链使用模型 测试充分性 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]