Inference Based on Z-Probability Trees

Decision-making is based on information. Often, decision-relevant information is imperfect and partially reliable. Zadeh proposed a concept of Z-number to account for this issue. In this paper, we consider a problem of inference with Z-valued probability trees. A statement of problem and solution procedures including the use of Bayes rule for Z-valued probabilities, computation of Z-valued expected values of alternatives and comparison of these values are described. An application of computation of posterior probabilities for oil extraction problem is considered.