Accelerated Life Tests of a Series System With Masked Interval Data Under Exponential Lifetime Distributions

We will discuss the reliability analysis of a series system under accelerated life tests when interval data are observed, while the components are assumed to have statistically independent exponential lifetime distributions. In a series system, the system fails if any of the components fails. It is common to include masked data in which the component that causes failure of the system is not observed. First, we apply the maximum likelihood approach via the expectation-maximization algorithm, and use the parametric bootstrap method for the standard error estimation. When the proportion of the masking data is high, the maximum likelihood approach fails due to lack of information. A Bayesian approach is an appropriate alternative in such a case. Hence, we also study the Bayesian approach incorporated with a subjective prior distribution with the aid of the Markov chain Monte Carlo method. We derive statistical inference on the model parameters, as well as the mean lifetimes, and the reliability functions of the system and components. The proposed method is illustrated through a numerical example simulated from the underlying model under various masking levels.