基于搜索的Simulink测试数据生成  

Search-based test-data generation for Simulink

作　　者：邓绍鹏[1] 杨志义[1] 王宇英[1] 

机构地区：[1]西北工业大学计算机学院 陕西省嵌入式系统技术重点实验室,西安710129

出　　处：《计算机应用研究》2012年第7期2527-2530,共4页Application Research of Computers

基　　金：国家“863”计划资助项目(2011AA010101);西安市科技计划项目(CXY1008)

摘　　要：为满足基于模型的测试需求,研究了基于搜索的Simulink模型测试数据生成方法。对于严格约束,该方法利用代价函数规则构建目标函数,针对Simulink模型复杂性的特点,采用模拟退火算法对目标函数求优,最终搜索出满足目标约束的测试数据。由此设计了一种综合考虑宽松约束和严格约束的Simulink测试数据生成框架,最后实现了该方法在典型案例的应用。实践证明,该方法快速有效地解决了Simulink模型的测试数据生成问题。In order to apply model-based testing,this paper focused on search-based approach of Simulink test-data generation.This method firstly constructed the objective function using the principle of cost function which met the requirements of particular target constraint.According to the complexity of Simulink model,it calculated the optimal value of the target function through the algorithm of simulated annealing and searchet the test-data which met the target constraint.And put forward a test-data generation framework which took loose constraints and tense constraints into consideration and illustrated the usage of this framework through a case study.Finally,it shows that search-based test-data generation is useful for Simulink model testing.

关 键 词：测试数据生成 Simulink测试 基于搜索的测试 基于模型的测试 模拟退火 

分 类 号：TP311.56[自动化与计算机技术—计算机软件与理论]