测试用例自动生成中约束求解策略的选择

Constraint Programming Strategy Selection on Test Case Generation

在软件测试中,构造测试用例是非常重要的一环.传统的测试用例生成方法都由测试人员手动完成,这种生成方式效率低下成本高昂,同时生成的测试用例非常有限,覆盖率较低.首先探讨了通过约束求解的方法,通过分析程序路径中所有的约束表达式,逆推得到测试用例的过程;然后调查了不同类型的C语言工程的待测源码的统计结果,同时针对其特征,引入了不同的约束求解算法策略;最后通过各种策略的组合与配置,给出了一种具有普遍性的测试用例自动生成的约束求解方法.

Test case generation is an important process in software testing.The traditional work to generate a group of test cases is all done by manpower.However,the traditional method can only get few test cases and reach very low coverage,but costs a lot of time and resources.A method to generate test cases through constraint programming is put forward.All constraint expressions in one path of a program is analyzed,then test cases from constraint expressions are generated through several mathematic algorithms.Some softwares written in C are investigated,the statistic rules of constraint types are summarized.Finally,an optimized process to generate test cases using constraint programming is concluded.