摘要
随着生产实际情况的不断变化,以及模糊神经网络不断的改进和发展,提出一种改进的构造神经网络的方法,并且提出混合学习算法,结合共扼梯度下降法与递归最小二乘估计来分别辨识网络中的前、后件参数,并对非线性系统进行仿真实验,达到控制要求。
Along with the changing of the actual situation of production and continuous improvement and development of the fuzzy neural network,this paper proposes an improved construction of neural networks and hybrid learning algorithm.Combined with conjugate gradient descent method and the recursive least squares estimation of the respective identification in the network before and after the consequent parameters,a simulation experiment of nonlinear systems is conducted and control requirement is achieved.
出处
《三明学院学报》
2010年第2期109-113,共5页
Journal of Sanming University
基金
福建省自然科学基金计划资助项目(2007J0207)
福建省教育厅科技项目(JB06189)
关键词
模糊神经网络(FNN)
结构算法
学习算法
优化
Fuzzy Neural Network(FNN)
structure algorithm
learning
algorithms
optimization