Sequential testing of simple hypotheses on parameters of Markov chains

Abstract

The problem of sequential testing of hypotheses on parameters of Markov chains is considered. The special case of simple hypotheses is analyzed. The expressions for conditional error probabilities and for conditional expected sequence sizes are obtained. The robustness analysis under "contamination" of Tukey—Huber type is performed. The case of Markov chains of the order K>J is studied separately. The results of computer modelling illustrate the theory.