论文部分内容阅读
从理论上研究将神经网络用于非线性系统控制，通过对神经网络的训练，实现一类非线性系统的定点跟踪。证明了神经网络学习算法的收敛性不仅与系统的动态特性有关，而且与网络的初始条件有关。仿真结果表明，适当选取网络的初值和加权的调节速率，可以实现非线性系统的定点跟踪。
In theory, the neural network is applied to the nonlinear system control. By training the neural network, the fixed-point tracking of a class of nonlinear system is realized. It is proved that the convergence of neural network learning algorithm is not only related to the dynamic characteristics of the system, but also to the initial conditions of the network. The simulation results show that the proper selection of the initial value of the network and the weighted adjustment rate can achieve the fixed-point tracking of the nonlinear system.