Use of Markov methods in analysis of fault-tolerant systems with flexible repair policies

Begins by considering some of the problems that have arisen in the application of traditional reliability methods to fault tolerant systems, particularly with the widely used fault tree approach. The application of the Markov state-flow equation to reliability analysis is then considered and it is shown how many of the aforementioned problems disappear with this approach and how the basic equation can be manipulated to include repair policies, discrete events, and to calculate system reliability. Then the issues are taken on to the next step by considering how to set up a reliability model from system design information in such a way as to ensure the Markov states and transitions are correct and so as to ensure that the reliability analysis gives an upper bound for system failure. The concept of formulation of design information and automatic generation of a reliability model for any given system is explained and an example analysis given based on a typical jet engine control system