摘要
针对一类能够由中立型变延迟非线性微分方程描述的神经网络模型,通过构造一个Lyapunov-Krasovskii泛函,给出了平衡点惟一性和全局渐近稳定的充分条件.所得到的稳定判据不依赖于时间延迟大小,不要求神经元激励函数的有界性、严格单调性和可微性,只与连接矩阵和延迟的导数项有关,且该稳定判据能够表示成线性矩阵不等式形式,易于求解验证.
A new sufficient condition is given to provide the uniqueness and global asymptotical stability of the equilibrium point for a class of delayed neural network models of neutral type by constructing a Lyapunov-Krasovskii functional without assuming the boundedness, strict monotonicity and differentiability of neuron excitation function. This stability criterion imposes constraints on the interconnected matrices and derivative of time delays and is independent of the amplitude of time delays, which can be conveniently expressed in the form of linear matrix inequality (LMI) and easy to check.
出处
《东北大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2006年第2期123-126,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(60244017
60325311)
关键词
延迟神经网络
中立型
全局渐近稳定
线性矩阵不等式
时变延迟
delayed neural networks
neutral type
globally asymptotical stability
linear matrix inequality (LMI)
time varying delay