基于约束组合的测试用例生成

发布时间：2019-07-23 06:33

【摘要】：系统输入参数常有多种约束条件,且约束之间相互关联。违背约束条件及约束之间的依赖关系,是软件中常见的缺陷。当参数数量多、输入空间大时,组合测试可在保证覆盖率的同时有效降低测试代价。该文针对约束及约束组合的故障检测问题,将约束覆盖作为测试充分性准则,提出约束条件的组合测试方法。以典型的在线交易平台会员注册服务为例,对比OA(orthogonal array)、IPO(in-parameter order)和OFOT(one factor one time)3种组合算法应用于约束组合时的性能表现。实验选取不同的故障模式和实验配置,对3种组合算法从生成时间、故障检测能力、用例规模等方面进行了比较。实验结果表明:OA算法生成时间短、用例规模小且相对稳定、故障检测能力中等,适合迭代优化。
[Abstract]:There are many constraints in the input parameters of the system, and the constraints are related to each other. Violation of constraints and the dependence between constraints is a common defect in software. When the number of parameters is large and the input space is large, the combined test can effectively reduce the test cost while ensuring the coverage. In this paper, the constraint coverage is used as the test adequacy criterion to solve the fault detection problem of constraint and constraint combination, and a combination test method of constraint condition is proposed. Taking the typical online trading platform member registration service as an example, this paper compares the performance of OA (orthogonal array), IPO (in-parameter order) and OFOT (one factor one time) when applied to constrained combination. Different fault modes and experimental configurations are selected in the experiment, and the three combined algorithms are compared in terms of generation time, fault detection ability, use case scale and so on. The experimental results show that the OA algorithm has the advantages of short generation time, small and stable use case size, medium fault detection ability and suitable for iterative optimization.
【作者单位】：清华大学计算机科学与技术系;
【基金】：国家自然科学基金资助项目(61472197)
【分类号】：TP311.53

【参考文献】