An Object-oriented Fuzzy Knowledge Representation And Its Reasoning Research

In the field of artificial intelligence, especially to solve the problems that are associated with the processing of knowledge in certain domains, a flexible and operational knowledge representation is necessary. A wonderful knowledge representation can most naturally describe all the objects and their rules in certain problem space. It also can be computable on theory and not too complex to work out in practice. This paper will propose a knowledge representation based on both object-oriented technology and fuzzy logic theory. The ability of description and granularity of this representation are evaluated in this paper.With the object-oriented technology, an object is a single entity in which data and methods are capsulized. It can be used to represent knowledge or entities in simulated domain. Because we can regard objects as entities' representation for computer, we can pay our attention to these entities and construct some computable entities that are as abstract as these entities.Fuzzy theory has been widely applied to AI. For example, fuzzy inference is based on fuzzy theory. In the real world, there are very many "fuzzy" concepts(or knowledge). How to represent these concepts(or knowledge) is the major objective of this paper.The object-oriented fuzzy knowledge representation combines the above technology and theory. It represents and processes fuzzy knowledge by fuzzy class(or fuzzy object). The corresponding inference engine is also interesting.After proposing an object-oriented fuzzy knowledge representation, this paper explores the fuzzy inference engine(FIE) that is based on this knowledge representation.