基于关联矩阵的两两组合测试算法

A Pairwise Test Suite Algorithm Based on Correlation Matrix

组合测试是软件黑盒测试中的一种常用方法,能有效检测软件系统中由各个因素相互作用所引发的软件故障。基于参数顺序的约束满足算法IPO-SAT是一种常用的组合测试用例生成算法,该算法约束控制能力突出,能生成精简的两两组合测试用例集,但其计算过程频繁调用约束求解器,导致较多的时间开销。针对该问题,论文提出一种基于关联矩阵的两两组合测试改进算法MIPO,采取将约束信息存储于关联矩阵中的方式来避免调用约束求解器,以减少测试用例生成过程中的整体时间。针对5组不同的参数组合进行实验,结果表明,改进后的算法与IPO-SAT算法相比,生成两两组合测试用例集的时间可节省90%以上。

Combination test is a method usually used in software black box testing, which can effectively detect the software fault caused by the interaction of factors in the software system. The IPO-SAT algorithm used to generate the combined test case has strong control ability and can generate right pairwise test suites, but the calculation process frequently calling the constraint solver is complex and time consuming. To solve this problem, this paper proposes an improved pairwise test algorithm based on the association matrix, storing the constraint information in the correlation matrix without calling the constraint solver to reduce the computation time. Experiment shows that compared to the IPO-SAT algorithm, the algorithm generates pairwise test suite with 90% time reduction.