多重模糊推理情形下的CRIP法

CRIP Method of Multidimensional Fuzzy Inference

该文对模糊专家系统、模糊决策支持系统和模糊控制系统中常用的Zadeh的模糊推理合成法则CRI进行了推广。基于 ∨λ 算子 ,该文提出了多重模糊推理情形下的带参数的模糊推理合成法则 (简称CRIP法 )。它克服了传统的CRI法则的若干不足之处 ,能在极弱条件下满足假言推理、拒取式推理 ,具有推理的逼近性。通过参数的调整 ,CRIP法具有推理的灵活性和准确性。

Zadeh's Compositional Rule of Fuzzy Inference (CRI for short) is generalized in the paper, which is often applied to Fuzzy Expert System, Fuzzy Decision Support System and Fuzzy Control System. Based on macro union operator, a Compositional Rule of Fuzzy Inference with Parameters (CRIP for short) in multidimensional fuzzy inference case is proposed in the paper. The new method overcomes a few shortcomings of the conventional CRI. It has approximation capability, and can satisfy modus ponens and other inference principles if the parameters are adjusted in a suitable way. By adjusting the parameters, CRIP holds on flexibility and accuracy. The method of obtaining suitable parameters is presented in the paper.