基于BP网络的三值动态矩阵控制算法的实现  

The realization of three dynamic matrix control algorithm based on the BP neural network

作　　者：慈宇红[1] 孙新蕊[2] 

机构地区：[1]秦皇岛职业技术学院基础部,河北秦皇岛066100 [2]燕山大学电气工程学院,河北秦皇岛066004

出　　处：《高师理科学刊》2007年第4期39-43,共5页Journal of Science of Teachers＇College and University

摘　　要：通过具体分析动态预测控制算法在工业应用中存在的实际问题,提出了一种三值动态矩阵控制算法,该算法主要在优化性能指标函数式中真正地建立了"头、中、尾"3点约束.仿真结果表明,该算法具有系统跟踪精度高、鲁棒性好等优点.同时,针对慢时变、大时延和不确定性系统的跟踪问题,提出了一种基于BP网络的三值动态矩阵控制算法.该算法在三值动态矩阵控制算法的基础上,利用一个3层BP神经网络对被控对象进行辨识.通过仿真,验证了所提算法的有效性.Through the concrete analysis the actual problem of performance predictive control algorithm which exists in the industrial application,proposed a kind of three values dynamic matrices control algorithm,this algorithm has established three constraints truly in the optimization performance index function in the formula.Simulation results show this algorithm has well tracking performance and robustness.Subject to the tracking performance of slow time-variable, a long time-delay and uncertainty systems, proposed one kind of three dynamic matrix control algorithm based on the BP neural network. This algorithm uses a three BP neural network to recognizes the object. The simulated results have shown this algorithm has well.

关 键 词：预测控制 三值动态矩阵控制 热力网系统 BP神经网络 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]