Sequential multiple hypothesis testing and efficient fault detection-isolation in stochastic systems

This paper develops information-theoretic bounds for sequential multihypothesis testing and fault detection in stochastic systems. Making use of these bounds and likelihood methods, it provides a new unified approach to efficient detection of abrupt changes in stochastic systems and isolation of the source of the change upon its detection. The approach not only generalizes previous work in the literature on asymptotically optimal detection-isolation far beyond the relatively simple models treated but also suggests alternative performance criteria which are more tractable and more appropriate for general stochastic systems.