论文部分内容阅读
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 network and the back propagation learning algorithm of feedforward network are expounded, and the simulation analysis of structural analysis is carried out by using single hidden layer network.Examples show that the large-scale parallelism consisting of nonlinear simulated neurons, Interconnected networks are very effective in structural analysis.