Abstract
An examination is made of some methods of determining Bayesian confidence limits for the reliability of series or parallel systems using binomial component test data. It is assumed that a prior distribution for the reliability of the component is a beta distribution. Then, some procedures of determining the posterior distribution of the system are considered. The results say that the posterior distribution of the system is determined based on the Mellin transformation explicitly for the cases in which some conditions are satisfied and an approximation method for deriving Bayesian confidence limits only for the series system reliability is presented when each component of the system is highly reliable.
