摘要
为了衡量不同软件的可靠性,至今人们已经建立了一百多种软件可靠性模型,但是在实际工程中,由于可操作性等方面的制约,难以有效应用。针对这种现象,给出了一种衡量软件可靠性的简单方法:通过将Nelson模型推广到软件可靠性增长中,只需要知道失效发生时所实施的测试用例数,就能通过一个递推公式直接计算出相应的软件可靠度,并通过40组实验数据进行了测试,其测试结果与后期运用JM(Jelinsk iMo-randa)模型估计出软件的可靠度具有一致性,从而证明了该方法的有效性和可行性。
Since along with the computer development, the software reliability all has received people's attention continuously, in order to measure the different software the reliability, until now the people have already established 100 many kinds of software reliable model, but in the actual process, because of operational reason, it's hard to apply them effectively. In view of this kind of phenomenon, this article has produced one measure software reliable straightforward procedure: Through promotes the Nelson model to the software reliability growth in, only needs to know the expiration occurs when implements measured the test example number, can through a recurrence formula direct calculation corresponding software margin of safety, and carries on the test through 40 groups of empirical data, its test result and later period estimated using the JM (jelinski Moranda) model the software the margin of safety has the uniformity, thus has proven this method validity and the feasibility.
出处
《吉林大学学报(信息科学版)》
CAS
2006年第1期99-102,共4页
Journal of Jilin University(Information Science Edition)
关键词
软件可靠性模型
软件可靠度
初始缺陷数
剩余缺陷数
失效数据
software reliability model
software margin of safety
initial software defects
remaining software defects
failure data