Statistical Analysis for Constant-Stress Accelerated Life Test With Weibull Distribution Under Adaptive Type-II Hybrid Censored Data

Wei Cui,Zaizai Yan,Xiuyun Peng

doi:10.1109/access.2019.2950699

Abstract

In this paper, we discuss the point and interval estimates of the parameters from Weibull distribution based on adaptive Type-II progressively hybrid censored data in constant-stress accelerated life test using maximum likelihood estimation (MLE) methods and Bayesian estimation (BE). The Bayes estimates cannot be obtained in explicit form. A hybrid Markov chain Monte Carlo method is applied to compute the approximate Bayes estimates and the corresponding credible intervals. Moreover, the Monte Carlo simulation is performed to compare the performance of the proposed methods. Finally, an actual data set is analyzed for illustrative purposes.

Highlights

Life test is necessary for investigating, analyzing and evaluating the reliability of the product
The parameter estimation was discussed for the constantstress accelerated life test under adaptive Type-II progressive censoring scheme when the lifetime of products follow the Weibull distribution
The point and interval estimates of unknown parameters, reliability and hazard functions were derived by using methods of the Maximum likelihood and Bayes estimates

Summary

INTRODUCTION

Life test is necessary for investigating, analyzing and evaluating the reliability of the product. Shi et al [11] considered a constantstress accelerated life test with competing risks for failure from exponential distribution under progressive type-II hybrid censoring, they derived the MLE and Bayes estimates. Under the constant-stress accelerated life test with progressive type-II hybrid censoring, the parameters of the exponential distribution are considered by Anwar [12]. Mahmoud et al [13] deliberated the inference for onstantstress partially accelerated life test model with progressive type-II hybrid censoring scheme using the MLE, Bayes and parametric bootstrap methods. The research on constant-stress accelerated life test with adaptive Type-II progressive hybrid censoring is very little, except Zheng and Shi [15], they estimated the unknown exponential distribution parameter and the reliability function through combining the EM algorithm and the square method in 2013.

LIFE TEST PROCEDURE AND BASIC ASSUMPTIONS

BAYESIAN ESTIMATION

SIMULATION STUDY

APPLICATION TO A REAL DATA EXAMPLE

CONCLUSION