用神经网络进行非线性离散动态系统辨识的可行性

The Feasibility of Using Neural Networks for the Identification of Nonlinear Discrete Dynamic Systems

中文关键词  神经网络  非线性离散动态系统  Lyapunov函数

英文关键词  neural network  nonlinear discrete dynamic system  Lyapunov function

中文摘要

      本文通过证明由神经网络组成的动态系统具有可以逼近满足一定条件的非线性离散动态系统到任意精度的能力，以及讨论将其作为辨识模型对非线性离散动态系统进行实时辨识时系统的稳定性，对用神经网络进行非线性离散动态系统辨识的可行性进行了探讨，并提出了有待解决的问题。

英文摘要

      In this paper, by proving that dynamic systems constituted by neural networks can approximate a kind of nonlinear discrete dynamic systems with arbitrary accuracy, and analyzing the stability of applying neural networks to the real-time identification of nonlinear discrete dynamic systems, we discuss the feasibility of using neural networks for the identification of nonlinear discrete dynamic systems. We also address the problems remained to be solved.