摘要
提出了一种基于高木-关野模糊系统的pi-sigma神经网络结构、学习算法,并分析了学习算法的收敛性.用这种混合型pi-sigma神经网络系统去实现模糊规则及其隶属函数的修正,从而得到模糊推理的自适应性.在设计过程中,引入Zadel模糊取乘算子,使之适合基于梯度的学习算法.最后的仿真结果表明此种网络的有效性、优越性并且在非线性建模、控制等方面有重要的应用价值.
A fuzzy neural network structure and study algorithms,based on T-S fuzzy systems namely fuzzy pi-sigma neural network has been proposed.The neural network is used to enable improvement of fuzzy rules and modification the membership functions.The fuzzy controller adopts one of the Zadeh s operators in order to make the BP algorithms been used.Simulation results are carried out to prove that this neural network is valid and feasible.
出处
《淮阴师范学院学报（自然科学版）》
CAS
2009年第4期284-287,296,共5页
Journal of Huaiyin Teachers College;Natural Science Edition
基金
国家自然科学基金资助项目(60772072)
