基于LS—SVM逆系统方法的非线性动态矩阵控制

针对神经网络逆系统建模存在的诸多问题，提出了基于最小二乘法支持向量机的α阶逆系统方法的非线性动态矩阵控制新方法．将最小二乘支持向量机辨识非线性对象的α阶逆模型与原系统串联组成伪线性系统，根据线性动态矩阵预测控制方法对伪线性系统进行控制．仿真结果表明，系统存在扰动和模型参数发生变化时，依然具有很好的动、静态性能，且表现出很强的鲁棒性，证明了方法的有效性．该方法不依赖于系统的数学模型，简化了非线性对象动态矩阵控制器的设计，为非线性预测控制的研究提供了一种新途径．

Nonlinear Dynamic Matrix Control Based on LS-SVM Inverse System Method

For the problems of neural network for inverse system modeling,a novel method of nonlinear dynamic matrix control based on LS-SVM α th-order inverse system method is proposed. The method cascades the α th-order inverse model approximated by LS-SVM with the original system to get the composite pseudo-linear system. Then the linear dynamic matrix control method is introduced into the pseudo-liear system. The simulation result shows that both of the dynamic and static performance of the system is excellent even when there is disturbance or large change of model parameters. It is also shown that the system has strong robustness, which is a proof of the validity of the method. This method does not rely on the accurate mathematical model,therefore it not only simplifies the design of nonlinear DMC controller, but also supplies a new way for the research of nonlinear predctive control.