基于B-P神经网络的非线性系统预测控制的研究

RESEARCH OF PREDICTIVE CONTROL BASED ON B P NETWORKS FOR NONLINEAR SYSTEM

预测控制是以计算机为手段基于模型预测进行控制的方法，但是已有的预测控制算法通常是针对线性渐进稳定对象的，或者即使针对非线性使用了非线性模型，但由于算法过于复杂不能适用于快速系统．本文对复杂非线性系统提出了一种基于Ｂ－Ｐ神经网络的预测控制方法，仿真和实际结果表明该方法的有效性和快速性，能够实现对非线性系统的实时智能优化控制．

Predictive control is a control method based on model prediction by computers.The existing algorithms are usually applicable to linear approximate steady objects.Even if the nonlinear model is used for the nonlinear system,the too complex algorithm is not suitable for quick systems.This paper presents a predictive control method based on B P neural networks for the complex nonlinear indefinite system.The results obtained from simulation and practice demonstrate the effectiveness and quickness of the method.It can realize the real time intelligent optimum control for nonlinear systems.