Toward a New Approach to Reasoning with Z-Number-Valued If-Then Rules

Existing works on reasoning with “if-then” rules with Z-number-valued components are scarce. In this work, an approach to reasoning with such kind of rules is proposed. The approach relies on a new concept of similarity of Z-numbers. This allows to reason with sparse rule base under relatively low computational complexity. Fuzzy and probabilistic information is used in the main stages of reasoning. An application of the proposed approach is considered.