A Classification System for Reliability Models

There has been little attempt to introduce a general classification system for reliability models analogous to those in other areas of applied probability, notably queueing systems. Thus there has been an unavoidable confusion in the literature and considerable unnecessary duplication of investigative effort. This paper argues the case for a classification system analogous to that in use for queues; identifies necessary features of such a system; and proposes a partial classification system. To be successful such a system must not only be a relatively convenient summary of diverse models, but must have the general support of the reliability community.