Reliability Analysis of Stress-Strength Model under Adaptive Progressive Hybrid Censoring Accelerated Life Test
The reliability analysis of the stress-strength model under the step-stress partial accelerated life test based on the adaptive progressive type-Ⅱ hybrid censored samples is studied,addressing the issue of long life test times and test data is difficult to collect.Firstly,the maximum likelihood estimation of the reliability of the stress-strength model is obtained according to the maximum likelihood theory,and the asymptotic confidence interval is provided.Then,two kinds of Bootstrap confidence intervals are constructed using the Bootstrap method.Secondly,Bayesian estimation of the reliability of the stress-strength model is obtained by the Lindley approximation method and the Markov chain Monte Carlo method,and the highest posterior density credible interval is derived.Finally,a comparative study is conducted using numerical simulations on the maximum likelihood method,the Lindley ap-proximation method,and the Markov Chain Monte Carlo method,while also comparing the excellence of the highest posterior density credible interval,Bootstrap confidence interval,and asymptotic confidence interval.The simula-tion results show that Bayesian estimation outperforms maximum likelihood estimation for the reliability of the stress-strength model,and Bayesian estimation based on the Markov chain Monte Carlo method is superior to that based on the Lindley approximation method.As the sample size increases,the mean square error and mean absolute deviation of both Bayesian estimation and maximum likelihood estimation decrease.The highest posterior density credible interval is found to perform better than both the asymptotic confidence interval and the Bootstrap confidence inter-val.From these conclusions,it can be inferred that the proposed method is practical for estimating the reliability of the stress-strength model under accelerated life testing.Furthermore,the Markov Chain Monte Carlo method out-performs both the maximum likelihood method and the Lindley approximation method.Additionally,the highest pos-terior density credible interval derived from Bayesian estimation is superior to both the Bootstrap and asymptotic confidence intervals constructed using maximum likelihood estimation.
万方数据
维普
