Chapter 4 Multistate start-up demonstration tests

In this chapter, we study the cases in which each trial could result in more than two outcomes, leading to multistate start-up demonstration tests. The models included in this family specifically assume that the outcome from each trial is a discrete random variable with more than two possible values; theses values correspond to different types of successes and failures. The acceptance and rejection criteria are quite similar to those used in the family of binary models; they are based on the appearance of a consecutive/total number of successes/failures of a specific type. The Markov chain embedding method has been used extensively for the study of multistate demonstration tests, under both i.i.d. and Markov-dependent trials. The nature of multistate start-up demonstration tests is also directly related to problems from the general theory of waiting times of patterns on multistate trials or competing patterns or the multiple failure mode reliability theory.