摘要
针对2004年Pankaj Jalote和Brendan Murphy提出的P-B模型假设过于苛刻、适用范围较窄,文中提出了一种新的基于软件发布后的可靠性模型。该模型不仅考虑了原P-B模型中用户对可靠性产生的影响,而且还考虑了通过版本升级和打补丁修正软件本身错误对软件可靠性产生的影响,同时将这种影响反映到模型的衰减因子中,进而对原始模型的衰减因子进行修正,并通过最小二乘法对新模型参数进行估计。最后通过仿真实验证明了,改进后的新模型比原始模型的预测精度更高,更符合实际失效数目的变化。
Because the assumption of P-B model raised by Pankaj Jalote and Brendan Murphy was too rigorous and the applicable scope was narrow,a reliable model based on the publication of software was brought up.This model not only considered the effect caused by the users of the original P-B model,but also took the influences of modifying errors of software itself by edition upgrading and patching into account,by taking the effect into the decay factor of the mode,modifying the decay factor of the original model and estimating parameters of the new model by least square method.At last,the simulation experiment proved that the improved new model was more accurate on calculation and more suitable to the real changes of the amount of parameters of software failure.
出处
《计算机技术与发展》
2011年第9期85-87,91,共4页
Computer Technology and Development
基金
南京邮电大学校科研基金(01院32)
关键词
软件可靠性
衰减因子
失效率
版本升级
软件补丁
software reliability
decay factor
failure rate
edition upgrading
software patch