自调整动态神经网络模型及其在带材板形预测中的应用

 建立了一个自调整动态神经网络预测模型，它在BP网络模型基础上，对网络的自身结构及学习规则进行了动态优化。网络能自组织和自学习自己的结构，即在学习过程中，网络可根据具体问题自动调整本身的结构，从而使结构达到最优。在此基础上，建立了板形神经网络预测模型，经某带钢厂四辊冷轧实测数据仿真验证表明，该模型具有很高的预测精度。

Self-Adjusting Dynamic Neural Network Model and Its Application in Strip Shape Prediction

A prediction model based on self-adjusting dynamic neural network is built, and on the basis of BP network, the structure and the studying rules of the network are optimized dynamically. The network has the self-organizing capability and the self-studying capability for its structure, and in the course of studying, the network can adjust automatically its structure according to specific questions, and make its structure reach the optimal state. The neural network prediction model of shape control is built. Computer simulations using data obtained from 4- high cold rolled some strip steelworks indicate that the model has high prediction precision.