Fuzzy reasoning is an important part of fuzzy theory, and plays major role in fuzzy expert systems. In most fuzzy reasoning methods, the similarity-based weighted fuzzy reasoning is a kind of simple and flexible method. In this paper toward the limitation of two existing definitions of the similarity measure and the corresponding fuzzy reasoning methods, we propose two improved similarity measures and the enhanced weighted fuzzy reasoning methods. With respect to the limitation of the former similarity measures and reasoning methods, this paper proposes three new definitions of the similarity measure, constructs the corresponding weighted fuzzy reasoning methods and compares the proposed methods with several existing methods. We obtain some superiority by analyzing the properties of these methods, thus the new methods enrich the fuzzy reasoning theory and extend the field of its applications.
|