广义逆指数分布应力-强度模型的可靠性评估  

Reliability Evaluation of Stress-strength Model with Generalized Inverted Exponential Distribution

作　　者：罗君念 蔡静 张峰源 LUO Junnian;CAI Jing;ZHANG Fengyuan(School of Data Science and Information Engineering,Guizhou Minzu University,Guiyang 550025,China)

机构地区：[1]贵州民族大学数据科学与信息工程学院,贵阳550025

出　　处：《湖北民族大学学报（自然科学版）》2023年第1期102-108,共7页Journal of Hubei Minzu University:Natural Science Edition

基　　金：国家自然科学基金项目(11901134)。

摘　　要：基于广义逆指数分布,研究应力-强度模型的可靠性评估问题。首先,利用极大似然法给出参数及应力-强度模型可靠度的极大似然估计。其次,通过形状参数选取Gamma共轭先验,在平方损失下运用Gibbs抽样算法及Lindley近似算法给出不同超参数选择下应力-强度模型可靠度的贝叶斯估计。最后,通过Monte-Carlo模拟对2种估计方法的精度进行比较。数值模拟结果表明,贝叶斯估计优于极大似然估计,基于Lindley近似算法的贝叶斯估计优于基于Gibbs抽样算法的贝叶斯估计。Based on the generalized inverted exponential distribution, the reliability evaluation problem of stress-strength model is studied.Firstly, the maximum likelihood method is used to give a maximum likelihood estimate of the reliability of the parameters and stress-strength models.Secondly, by selecting the Gamma conjugate prior distribution for the shape parameters, Bayesian estimate of the reliability of the stress-strength model under different hyperparameter selections is given by using Gibbs sampling and Lindley′s approximation algorithm under square loss.Finally, the Monte-Carlo simulation compares the accuracy of the two estimation methods.The simulation results show that Bayesian estimation is better than the maximum likelihood estimation, and Bayesian estimation based on Lindley′s approximation algorithm is better than Bayesian estimation based on Gibbs sampling.

关 键 词：广义逆指数分布 应力-强度模型 可靠度 极大似然估计 贝叶斯估计 

分 类 号：O213.2[理学—概率论与数理统计]