基于Choquet积分-OWA的自适应神经模糊推理系统及其应用

Choquet Integral-OWA Based Adaptive Neural Fuzzy Inference System and Its Application

针对已有的自适应神经模糊推理系统(ANFIS)在模糊规则后件表达上的缺陷和常见的模糊推理系统存在的主要问题,提出基于Choquet积分-OWA的模糊推理系统(AggFIS),在模糊规则的后件表达、模糊算子的普适性和输入及规则的权重等方面有很大优势,它试图建立能够充分体现模糊逻辑本质和人类思维模式的模糊推理系统.根据模糊神经网的基本原理将AggFIS与前馈神经网络相结合,得到基于Choquet积分-OWA的自适应神经模糊推理系统(Agg-ANFIS),并将该模型应用于交通服务水平评价问题.实验结果证明,基于Choquet积分-OWA的自适应神经模糊推理系统具有很好的非线性映射功能,它的本质是一类通用逼近器,为解决复杂系统的建模、分析及预测问题提供了有效的途径.

In order to solve the defects of consequent part expression in ANFIS（adaptive neural fuzzy inference system） model and several shortcomings in FTS（fuzzy inference system）,this paper presents a Choquet integral-OWA（outlook web access） based FTS,known as AggFIS.This model has advantages in consequent part of fuzzy rule,universal expression of fuzzy inference operator and importance factor of each input and each rule,aiming at establish fuzzy inference system that can fully reflect the essence of fuzzy logic and human thinking patterns.If AggFIS is combined with a feed forward-type neural network according to the basic principles of fuzzy neural network,the Choquet integral-OWA based adaptive neural fuzzy inference system（Agg-ANFTS） is obtained,which is applied to the evaluation of traffic level of service.Experimental results show that Agg-ANFTS is a universal approximator because of its nonlinear mapping capability by training and can be used in,analysis and prediction of complex systems modelling.