摘要
本文对具有投影学习算法的Hopfield型二值神经网络(HBNN)在同步工作方式下的收敛性能进行了比较全面的讨论。证明了网络具有演化收敛到网络平衡点的特性,推导了计算网络收敛所需瞬态工作步数上界的估算公式,分析并给出了计算平衡点(记忆模式)收敛半径的算法。
In this paper, the convergence properties of Hopfield-type associative neural network with projection learning rule under synchronous update operation have been studied. Using energy function method, we have proved the network having the property converging to one of its equilibrium points. A formula for estimating the upper bound of update steps required for convergence and an algorithm for calculating the convergent radius of the equilibrium point have been derived.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1993年第5期91-95,共5页
Acta Electronica Sinica
基金
国家攀登计划资助项目
关键词
神经网络
收敛性
联想神经网络
Associative neural network, Convergence, Convergence time, Convergent radius