Robustness analysis of Wang neural network for online linear equation solving
Robustness analysis of Wang neural network for online linear equation solving
复制标题
DOI：
10.1049/el.2012.1940
复制
发表时间：
2012-10
影响因子：
1.1
通讯作者：
Ke Chen
中科院分区：
文献类型：
--
作者：
Ke Chen
在理想的无噪声条件下，Wang神经网络可以有效地在线求解线性方程组，并可实现为电路。在考虑系数噪声和较大的实现误差的情况下，对该模型进行了鲁棒性分析，并估计了稳态解误差的上界。说明性的例子证明了理论结果的鲁棒性的王神经网络求解在线线性方程组。
The Wang neural network, which could be implemented into circuits, can solve online linear equations effectively and efficiently under the ideal noise-free condition. With the consideration of noises on coefficients and large implementation error, robustness analysis of such a model is presented with the upper boundary of the steady-state solution error also estimated. Illustrative examples demonstrate theoretical results about the robustness properties of the Wang neural network for solving online linear equations.