摘要
阐述了人工神经网络的基本原理和前馈网络的反向传播学习算法,并应用单隐层网络进行了结构分析的仿真研究.实例计算结果表明,由非线性模拟神经元组成的大规模并行、互相连接的网络在结构分析中是非常有效的。
The basic principle of artificial neural networks and the back-propagation learning algorithm for feedforward networks are described,and the simulation of structural analysis is carried out by using neural networks with one hidden layer.It is shown by two illustrative numerical examples that massively parallel,interconnected networks of nonlinear analog neurons are most effective in structural analysis.
关键词
人工神经网络
仿真
结构分析
structural analysis
artificial neural network
simulation
