基于改进遗传算法的递归神经网络非线性系统辨识

Nonlinear System Identification with Recurrent Neural Network Based on Genetic Algorithm

将递归内时延神经网络应用于非线性动力学系统辨识中 ,描述了其动力学方程 ,并引入改进遗传算法作为其学习算法 ,通过非线性动力学SISO和MIMO系统的辨识仿真研究 。

An internal time delay recurrent neural network (RNN) is used for identification of nonlinear dynamic systems, and its dynamic equations are described. As a learning algorithm, an improved genetic algorithm is applied to train the RNN. Two identification simulation examples for nonlinear dynamic SISO and MIMO plants validate the efficiency of the internal time delay RNN and the improved GA.