论文部分内容阅读
本文以双输入双输出的多变量系统为例,讨论了多变量系统的神经网络动态解耦问题,给出了网络训练的具体算法。并以电加热炉为对象,经解耦后应用广义预测控制器进行了实时控制,获得了满意的结果。
In this paper, we take the dual-input and dual-output multivariable system as an example to discuss the dynamic decoupling problem of multivariable neural network, and give the concrete algorithm of network training. The electric heating furnace is taken as an object and the real-time control is realized by using the generalized predictive controller after decoupling. The satisfactory result is obtained.