一种可求逆的模糊神经网络模型及其在热轧产品中应用

A Reversible Fuzzy Neural Network and Its Application to Hot Rolling Mill Production

提出了一种新的模糊神经网络及其学习算法.利用这种模糊神经网络模型,由非线性函数的输出值可以方便地求出输入向量的某一区域,使得对此区域中的每一向量,非线性函数的输出与模糊神经网络的输出值之差可以达到任意要求的精度.将这种模糊神经网络用于热轧产品的质量控制,获得了好的效果.

A kind of FNN and learning algorithm is presented in this paper. Using the FNN, by non\|line function's output value, the region of input vector may be get expediency, such that for any vector in the region, the error between the non\|line function and the FNN output value may be small arbitrary. The new FNN is used in the control of hot rolling mill production and a satisfying result is obtained.