Goal-driven reasoning for fuzzy knowledge-based systems using a Petri net formalism

Abstract

In this paper we describe a goal-driven reasoning algorithm over fuzzy knowledge bases. The approach to the problem permits characterizing the chaining between rules without requiring a complete identification between the propositions. The certainty degree of the rules is implemented by means of linguistic truth values and we include different considerations on the goals. A simple and adequate description of the process is achieved through the use of a Petri net formalism for the representation and manipulation of fuzzy knowledge bases.<>