摘要
通过引入非线性测度的概念,不要求激励函数连续可微、单调非降或有界,并假设自反馈项为非线性,研究了一类Hopfield神经网络平衡点的存在性、唯一性.得到了全局指数稳定性的判据,并给出了收敛率估计式.
The existence, uniqueness of the equilibrium point of a class of Hopfield neural networks were studied by introducing nonlinear measures. The activation function is nonlinear and not necessary to be continuous, monotone and bounded. Some sufficient conditions and convergence estimation for global exponential stability of neural networks were obtained.
出处
《西南交通大学学报》
EI
CSCD
北大核心
2005年第3期338-342,共5页
Journal of Southwest Jiaotong University
基金
国家自然科学基金资助项目(10272091)
关键词
神经网络
指数稳定性
收敛率
非线性测度
neural networks
stability
global stability
nonlinear measures