BAYESIAN DYNAMIC FORECASTING FOR ATTRIBUTE RELIABILITY

In this paper, we consider a Bayesian dynamic forecasting model that u tilizes both the engineering knowledge about the product reliability a nd attributes (success or failure) data gathered from the inspection o f the early stage of development or storage. We assume that a prior di stribution of reliability follows a beta distribution. The expected re liability is represented as a cumulative logistic function of the leng th of time that a new product has been under development or a finished item has been stockpiled in storage. As periodic testing produces att ribute data, a prior distribution is updated. The expected reliability and forecasting errors are obtained from a posterior distribution tha t reflects the uncertainty involved in forecasting. The proposed metho d is applied to predict the expected reliability decay of a gyroscope in a missile stockpile. (C) 1997 Elsevier Science Ltd.