ABSTRACT
One of the most challenging and important versions of the change detection–isolation problem is the multidecision and multistream detection problem which generalizes the multistream changepoint detection problem. This chapter addresses a simplified multistream detection–identification scenario where the change can occur only in a single stream. It focuses on a semi-Bayesian setting assuming that the change point is random and that there is no prior distribution on post-change hypotheses. The chapter shows that under certain very general conditions the proposed multihypothesis detection–identification rule asymptotically minimizes the trade-off between positive moments of the detection delay and the false alarm/misclassification rates expressed via the weighted probabilities of false alarm and false identification.
