【摘要】 为了提高软件统计测试的效率,使用较少的测试资源达到测试目的,对测试剖面进行优化成为必需.采用客户需求和测试需求作为约束,模型静态参数或其他统计质量标准作为优化目标,提出了一种基于约束优化的测试剖面最优化计算框架.通过扩展蚁群算法研究了一种求解优化测试剖面的方法.实验表明,相对于传统的函数最优化方法,使用启发式的蚁群算法具有更好的优化效果,并且更具通用性.在统计测试中使用经过优化的测试剖面可以生成统计意义上符合测试需要的用例集合,提高测试效率.更多还原

【Abstract】 In order to sufficiently explore the advance of statistical testing and quickly achieve the testing goal,it is necessary to optimize the testing profile.Custom requirements and testing requirements are used as constrains,and statistical static parameters of usage model or other statistical quality criteria are used as the optimizing aims.Then an optimizing framework of testing profile is setup.By extending the ant colony arithmetic,a method base on constrain optimizing to calculate the testing profile is developed.The experiment results show that using heuristic ant colony arithmetic can achieve better optimizing effect than conventional functional constrained optimization arithmetic,and it is more universal.Using optimized testing profile in software statistical testing can generate better test case set to satisfy test requirement on statistical meaning to improve testing efficiency.更多还原