论文部分内容阅读
分析了ELMAN网络的基本结构 ,并讨论了ELMAN神经网络的学习算法 ,针对BP算法的缺陷 ,提出了采用遗传算法修正网络的权值的学习算法。最后 ,利用仿真曲线说明了该方法在高阶次非线性系统中进行了辨识的可行性
The basic structure of ELMAN network is analyzed, and the ELMAN neural network learning algorithm is discussed. According to the flaw of BP algorithm, a learning algorithm that uses genetic algorithm to modify the weight of network is proposed. Finally, the simulation curve is used to demonstrate the feasibility of this method in high-order nonlinear systems