摘要
在科学计算和工程应用当中,很多问题最终转化为线性方程组问题,因此线性方程组的求解问题就显得尤为重要。针对此提出了一种以线性神经网络为基础,对具有唯一解的线性方程组进行求解的新方法。将原线性方程组求解问题转化为求神经网络的权值问题。通过运用此方法求解大量各种形式的线性方程组,验证了该方法的可行性和有效性。
In scientific computing and engineering application,many issues will eventually transformed into the problem of linear equations.Therefore,the question of linear equations system solving becomes particularly important.In this paper,it presents a novel method based on linear neural network to solve linear equations system which has a unique solution.The question of linear equations system solving is transformed into the question of seeking weights of neural network.The method is proved to be reliable and effective that has been used to solve a large number of various forms of linear equations.
出处
《山西电子技术》
2010年第2期9-10,共2页
Shanxi Electronic Technology
关键词
线性神经网络
线性方程组
算法
linear neural network
linear system of equations
studying algorithm
作者简介
曲婧（1985-），女，山西太原人，在读研究生，主要研究方向是通信与信息系统，智能信息系统及神经网络理论应用。
