基于神经网络的直觉模糊推理方法

Method of intuitionistic fuzzy inference based on neural network

模糊推理是不确定性推理领域的一种很重要的方法。直觉模糊推理基于Atanassov直觉模糊逻辑,具有灵活性、合理性和可信性的优点,但并没有克服推理结果依赖于蕴涵关系的选取及有时不能满足推理一致性要求的缺点。通过理论推导,提出基于神经网络的直觉模糊推理方法,证明了该方法在正规直觉模糊集上满足推理的一致性。通过算例分析,验证了该方法的合理性。

Fuzzy inference is an important approach in the uncertainty reasoning community. Intuitionistic fuzzy inference based on Atanassov intuitionistic fuzzy logic has a flexible, rational and credible framework, but does not overcome the limitation that reference results rely on the choice of implication relations, so that it can not satisfy the request of reference correspondence sometimes. A method for intuitionistic fuzzy inference based on neural network is proposed, the reference correspondence of the method is verified on the basis of formal intuitionistic fuzzy sets, and the rationality of the method are verified with a particular instance.