软件可靠性模型的Bayes推断及Gibbs算法

Bayes inference and Gibbs algorithm for JM model

作为重要的软件可靠性模型,JM模型的研究具有重要意义.论文研究了JM模型Bayes估计及其Gibbs算法.在先验分布确定的情况下,给出了Bayes估计的Gibbs算法,并证明了其收敛性.最后通过模拟分析发现:在l≥100,k>600时,所得到的参数Bayes估计与初始值几乎无关.从而说明Gibbs算法的可行性.

As an important software reliability model, JM model is significant to study. The Bayes evaluation and Gibbs algorithm for JM model are investigated in this paper. Given the prior-distribution,it gives the Gibbs algorithm of Bayes evaluation and proves the astringency of Gibbs algorithm. The algorithm＇s possibility is analyzed by simulation method at the end.