A PARAMETERIZED FUZZY PROCESSOR AND ITS APPLICATIONS

In this paper, a parameterized fuzzy processor (PFP) is presented based on a designed fuzzy instruction set of the parameterized fuzzy number (PFN). This PFN is presented by some parameters of the universe of the discourse instead of a series of the membership grades. The algorithms of these instructions are based on the parameter operations to implement the fuzzy arithmetics and the fuzzy reasoning methods. The instruction set supports the versatility of the fuzzy information, e.g., fuzzy arithmetic operations, inference operations, fuzzy logic operations, data translations and data exchangements. Based upon these, the characteristics of the PFN are discussed, and the architecture of the PFP outlined is constructed in this paper and its processing time is saved due to the adaptation of the parameterized fuzzy number (PFN). To evaluate the performance of PFN instructions and PFP structure, two experiments are presented. One is the speed performance evaluated for the PFN to compare with the discrete fuzzy number (DFN). The other is the confirmation of the feasibility and the capability of the instructions, which is demonstrated by three examples related to human inference, intuitive control and decision-making.