摘要
构造一个的结构可控的前向神经网络,并应用于结构位移法方程的求解.采用动态的网络学习算法,使该网络得以更快速度收敛.数值模拟证明,用这种加速的神经网络求解线性方程。
A structural controllable neural network, or SCNN for short, is constructed and applied to the solution of structural stiffness equations. This SCNN can be converged in faster speed by means of dynamic learning algorithm. As proved by numerical simulation, this accelerated SCNN is better than the traditional iteration of GaussSeidel for solving linear equations, it excels at the speed of convergence.
出处
《华侨大学学报(自然科学版)》
CAS
1998年第3期275-279,共5页
Journal of Huaqiao University(Natural Science)
基金
福建省自然科学基金
关键词
结构分析
神经网络
学习算法
structural analysis, neural network, learning algorithm