基于CBR的测试用例复用方法研究

A Study of Test Case Reuse Based on CBR

为了从已有的测试用例中提取经验知识、缩短测试用例设计时间,本文提出基于CBR(案例推理)的测试用例复用方法。首先对测试用例复用过程进行分析,指出测试用例检索是测试用例复用过程的关键。在测试用例检索中采用K近邻法,并对K近邻法进行改进,同时在改进的算法中使用带权重的距离度量算法。在此基础上提出遗传模拟退火算法,该算法可对测试用例属性的权重进行优化,是遗传算法和模拟退火算法的结合,可以有效避免遗传算法的早熟问题,增强算法的全局寻优能力,缩短搜索时间。通过实验可以证明,该算法比标准的遗传算法和模拟退火算法具有更高的求解质量和求解效率。

To extract experience knowledge from test cases and reduce the test cases design time, a method of test cases reuse based on Case-Based Reasoning (CBR) is presented. Firstly, the process of test cases reuse is analyzed. It is pointed out that test cases retrieval is the key to the process of test cases reuse. K-Nearest Neighbor is adopted and im-proved to retrieve test cases, and the weighted distance measurement algorithm is used in the improved algorithm. On this basis, optimization of the weights of attributes based on Genetic Simulated-annealing Algorithm (GSA) is proposed. The algorithm, which can optimize the weights of attributes, is the combination of Genetic Algorithm (GA) and Simu-lated-annealing Algorithms (SA). It can effectively avoid the premature convergence problem, improve the global-optimization capability and shorten the search time. Finally, results of the experiment indicated the proposed GSA had better efficiency and optimization performance than simple GA and SA, and would be one effective way to optimize the weights of attributes.