基于变异点的选择性变异测试

Position-Selective Mutation Testing

变异测试是一种基于错误的软件测试方法,相对于路径分析等其他测试技术,能够更为有效地发现软件错误,但由于变异体的数量巨大,极大地提高了测试成本,严重制约变异测试的广泛应用。因此,提出一种基于变异点的选择性变异测试方法,通过分析变异体之间的冗余关系,发现非冗余变异体通常出现在条件和循环等关键语句,仅对控制语句进行变异,能够降低45.02%的测试成本,仅牺牲7.93%的测试有效性。另外提出一种更为有效的变异测试评估方法。

Mutation Testing is a fault-based technique which can be more effective in finding software errors, comparing to other native techniques such as path analysis. As the number of the mutants is enormous, the cost of mutation testing is greatly expensive, which seriously restricts the wide application of mutation testing. Therefore, proposes a selective mutation testing method based on mutation point. By analyzing the relationship between the redundant mutants, finds that the non-redundant mutants usually appear in the control statements. Therefore, the test cost is reduced by 45.02% if the control statements are mutated only, and just 7.93% of the test effectiveness is lost. In addition, pro- poses a more effective evaluation method of mutation testing validity.