二项分布的可靠性增长模型

RELIABILITY GROWTH MODELS FOR BINOMIAL DISTRIBUTION

本文考虑了二项分布的可靠性增长模型,给出了各研制阶段参数的约束极大似然估计。我们提出了二项分布折合因子的概念,利用它获得了最末阶段参数的经典近似限。最后讨论了如何选择联合验前分布中的参数的问题,并给出了趋势检验方法。

In this paper, reliability growth model for binomial distribution is considered, and the constrained MLE of parameter for each development stage are given. We propose a concept of conversion factor of binomial distribution. By using this factor, the approximate classical limit of parameter for the last stage is obtained. At last, the problem of how to select the parameters of joint prior distribution is discussed and the trend testing approaches are given.