基于MCMC的MTBF值区间估计方法研究

摘要

针对Weibull分布参数的后验分布计算复杂的问题,提出采用MCMC方法直接构造贝叶斯后验分布的马尔科夫链,对分布参数的贝叶斯估计进行求解,解决了数值积分问题,保证可靠性评估的有效实施.利用提出的模型,分别对先验分布为伽马-逆伽马分布、二元正态分布情况下MTBF值区间估计进行了仿真,验证了该方法的有效性.%Calculating parameters of the posterior distribution for Weibull distribution is complicated, so we proposed MCMC approach to directly construct the Markov chain of Bayesian posterior distribution and Bayesian estimation of distribution parameters is solved, solving the problem of numerical integration, to ensure the effective implementation of the reliability assessment. Using the proposed model, the MTBF value of interval estimates is simulated corresponds to the priori distribution for the gamma-the inverse gamma distribution and the binary normal distribution respectively, to verify the validity of this method.