欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
考虑了一类鞍点问题.基于其系数矩阵的结构特点,将原问题转化为低维线性系统,提出了求解这类问题的一个新神经网络.运用Lyapunov稳定性理论和LaSalle不变原理,证明了所提出的模型是Lyapunov稳定的,且收敛于原问题的一个精确解,并在适当的条件下指数收敛到原问题的唯一解.最后通过数值实例说明了该模型的可行性和有效性.
英文摘要：
This article studies a class of saddle point problems. Based on the construction of coefficient matrix, the author converts the original problems into a linear system with lower dimension, and then presents a new neural network for solving the saddle point problems. The network is shown to be Lyapunov stable, and converge to one of its exact solutions by the Lyapunov theorem and LaSalle invariant set principle. Meanwhile, the network is shown to converge exponentially to its unique solution under some mild conditions. The validity and efficiency of the proposed neural network are demonstrated by several numerical examples.
同期刊论文项目
同项目期刊论文
期刊信息
位置：