【摘要】 该文在分析神经网络辨识技术特点及现状的基础上,将BP神经网络结构和遗传算法相结合,设计了一种适用于非线性系统的辨识器模型。该辨识器模型首先建立初始的BP神经网络结构,再利用遗传算法对BP神经网络的权值和阈值进行优化,从而优化BP神经网络,通过迭代最终建立辨识器模型。最后,通过一个三阶非线性多输入单输出系统的仿真实验证明了所设计的辨识器具有辨识时间短、辨识精度高的特点,为神经网络辨识技术的研究提供了新的思路和方法。更多还原

【Abstract】 The paper analyzed the characteristic and the situation of neural network identification,and designed a suitable identificater’s model for nonlinear system.The model combined BP neural network with genetic algorithm.Firstly,an initial structure of BP neural network was established.Secondly,all weights and threshold in BP neural network was optimized by genetic algorithm,so as to optimize it’s structure.Finally,an identificater’s model was established by iteration.In the end,the paper proved through a simulation of third-order nonlinear MISO system,this identificater has such characteristics as short time and high precision.The paper provides a new idea and method for researching on neural network identification.更多还原