A Consideration of the Bayesian Approach in Reliability Evaluation

The purpose of this paper is to demonstrate the application of Bayesian procedures to the field of reliability. First, the behavior of a family of Bayesian a posteriori distributions is examined pictorially and described asymptotically by a limit theorem. Then Bayesian point estimation is discussed and the properties of the mean of the a posteriori distribution as a point estimate are given. Following a discussion of the method of constructing Bayesian confidence intervals, Bayesian intervals are compared with Neyman intervals for a special case. The advantage that Bayesian intervals exhibit when system confidence intervals are desired from component tests is pointed out. The application of Bayesian theory in making typical reliability decisions via decision theory is also indicated.