基于Hopfield网络的多变量动态矩阵控制及其在精馏塔系统中的应用  

The Dynamic Matrix Control Based on Hopfield Neural Network and Its Application in a Distillation Column

作　　者：李树荣[1] 李峰[1] 

机构地区：[1]石油大学自动化系,山东东营257062

出　　处：《控制理论与应用》2001年第2期304-306,313,共4页Control Theory & Applications

基　　金：国家基础研究发展规划项目! (G19980 30 2 0 0 );山东省自然科学基金! (Q96G0 2 15 0 )资助项目

摘　　要：利用Hopfield网络的优化计算功能求解有约束多变量动态矩阵控制问题 ,算法简单 ,收敛性好 ,既可以用硬件实时实现 ,也可用数值积分求解 .对WoodIn this paper, we solve the constrained multivariable dynamic matrix control problem by using the Hopfield neural network. The optimization neural network designed is guaranteed to be globally convergent to solutions of the corresponding quadratic optimization problems and has a simple structure that can be easily realized using electric circuit. The solutions can be obtained by hardware circuit or numerical algorithm. The simulation on the Wood Berry distillation column demonstrates the effectiveness of the proposed method.

关 键 词：精馏塔系统 多变量动态矩阵控制 预测控制 HOPFIELD网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]