Non-Homogeneous Markov Models and Their Application in Reliability

A Markov model is a stochastic model of an item or process that transitions between discrete, mutually exclusive states. Within the context of reliability and maintainability, there are at least two states: Operational and Not Operational. Additional states may be added in the form of specific failure modes or mechanisms. Most of the literature, as well as commercially available Markov model software, makes the assumption that the state transition probability from any given state to the next is constant. This assumption makes the execution of the Markov model simpler, but it often violates the reality of the system being modeled. The results from such a model may be informative for relative changes in system performance. However in some cases it may not offer enough precision to be useful for decision making.