Transition probabilities for Markov chains having fuzzy states (Q4363897)

\textit{J. Norton} [Theory Decis. 25, No. 3, 287-313 (1988; Zbl 0675.60003)] has considered set-valued Markov chains where the state probabilities build a Dempster-Shafer mass function. Especially, he has presented a formula for transition probabilities in terms of this mass function. The authors discuss this Norton formula for Markov chains with fuzzy set-valued states. A direct application of this formula is possible only under rather restrictive assumptions, e.g. for hierarchical mass functions with fuzzy focal elements. For general mass functions with fuzzy focal elements, a generalized Norton formula is presented which involves a measure of the degree to which two fuzzy sets are equal. Finally, a result of \textit{Ph. Smets} [Uncertainty in Artificial Intelligence 7, 1-8 (1990)] is modified to find transition probabilities from fuzzy states to crisp states.