基于混沌神经网络预测模型的最优控制决策及应用  

Optimal Control Based on Chaos Neural Network Forecast Model and Its Application

作　　者：窦春霞[1] 张淑清[1] 

机构地区：[1]燕山大学电气工程学院,秦皇岛066004

出　　处：《动力工程》2004年第1期68-72,共5页Power Engineering

基　　金：国家自然基金项目(No.60102002);河北省基金项目(No.6011224);霍英东基金项目(No.81057)

摘　　要：为了实现非线性、大时滞系统的自适应控制,首先根据具有混沌特性的非线性、大时滞系统的时序列重构相空间,计算出相空间的饱和嵌入维数和最大Lyapunov指数,并以此为指导,建立混沌神经网络预测模型,该模型即便在网络输入不完整或发生变异的情况下,仍能对系统作高精度的短期预测;在此基础上,将预测模型的输出通过反馈校正,再将校正误差和控制增量引入性能函数最优,最后得到最优控制决策,实现了对非线性、大时滞系统高精度的自适应控制。最后将预测控制决策应用到非线性、大时滞的锅炉过热汽温控制中,仿真结果表明了该控制的有效性、快速性和鲁棒性。In order to realize adaptive control of nonlinear big-lagged system, a chaotic attractors space is reconstructed, systemic embed dimension and maximal Lyapunov exponent are calculated by the chaotic time series of the nonlinear big-lagged system in this paper. By above all, a chaos neural network model is constructed, which can make high precisin short-term forecast for the nonlinear big-lagged system even by imperfect and variation inputs. On the basis of this, a feedback rectification term is achieved by the forecast model outputs, a optimal controller is designed by the feedback rectification term and control input error being introduced into a performance functin, and robust adaptive forecase control to the nonlinear big-lagged system is realized. Last, the controller is applied to nonlinear big-lagged overheat temperature system of drum boiler, the validity, the high-speed and the robustness are proved by simulate results. Figs 6 and refs 6.

关 键 词：自动控制理论 最优预测控制决策 混沌神经网络预测模型 非线性大时滞系统 适应控制 鲁棒性 

分 类 号：O211.61[理学—概率论与数理统计] TP273[理学—数学]