Estimation of Generalized Rayleigh Components Reliability in a Parallel System Using Dependent Masked Data

Abstract

When the lifetimes data from the parallel system are masked and the masking probability is dependent with the failure component, we consider the reliability estimations of generalized Rayleigh components in the parallel system. Based on the given masking level and masking probability ratio, Maximum Likelihood Estimations (MLE) and Bayes estimations of parameters are obtained respectively. At last, numerical simulation demonstrates the correctness of theoretical results. We study the influence of masking level on the accuracy of the estimations, and then compare the effect of MLEs and Bayes estimations under non-informative priors and conjugate priors.