摘要
为了评估安全关键软件的防危性 ,该文在分析软件可靠性与防危性异同的基础上 ,建立了相应的防危性评估指标 ,同时研究了传统统计测试方法确定所需防危性测试用例数的局限性 .紧接着 ,使用贝叶斯函数动态地表示软件失效率的密度函数 ,提出了一种增量记忆型的防危性测试方法 ,该方法可根据软件出现失效时间的早晚及次数动态地确定所需测试用例数的增量 ,达到了改进传统测试方法局限性的目的 ,使安全关键软件测试失效后可经受更严格的测试 ,同时减小了测试工作量 .为证明该方法的合理性、有效性 ,使用了大量例证 .最后 ,就该领域内的相关工作及发展方向进行了展望 .
In order to evaluate the safety of safety critical software, on the basis of analyzing the similarities and differences between safety and reliability of software, the safety criterion is firstly established in this paper, and the limitation of the classical testing method based on statistically independent approach used to decide the number of safety testing case is studied. Following this, the density function of software failure probability with Bayesian function is presented dynamically, and a kind of new safety testing method with function of increment and memory based on the time and number of software failure is presented, which is used to decide dynamically the increment number of safety testing cases, and to improve the limitation of the classical testing method. With this approach, while a failure occurs during the course of testing, this safety critical software is to endure more stringent than the original test after the removal of the offending fault, and the testing workload is to be reduced. In order to proof the rationality and validity on this approach, a large number of examples are used to exemplify it. Finally, related work and future trends of research in this field are listed.
出处
《计算机学报》
EI
CSCD
北大核心
2004年第4期442-450,共9页
Chinese Journal of Computers
基金
国家预研基金 ( 2 0 0 0J6.7.1.DZ0 2 0 6)
国家"十五"攻关项目基金( 4 13 15 0 40 10 6)资助
