基于分层抽样确定软件可靠性测试最小测试量项目管理研究

传统的根据运行剖面生成测试数据的方法可以理解为一种简单的样本抽样。分层随机抽样是一种比简单随机抽样更复杂的抽样方式,其原理是将总体划分成若干个子总体,再在每个子总体里分别进行随机抽样,其优点是抽样精度更高,估计量的方差更小。本文简要介绍了现有的最小测试量的确定方法,并利用分层抽样对传统的测试用例生成过程进行了改进,提出了一种基于分层抽样技术的软件可靠性测试最小测试量的确定方法,指出了下一步的研究工作。

Study on the Minimal Test Suite in SRT Based on Stratified Sampling

Abstract The traditional way to generate test case can be interpreted as Simple Sampling. Stratified Sampling is more complicated than Simple Sampling. It＇s principle is to divide the population into several sub-population and execute random sampling respectively. This method has a better precision and smaller variance in terms of mean. This paper introduces some existing methods of determining the minimal test suite in SRT. Then, the process of generating test cased using Operational Profile was modified and a new method of determining the minimal test suite based on Stratified Sampling was proposed. In the end, the next step of the research is indicated.