标题:
随机Hopfield神经网络渐进行为的变互式凸组合方法
Reciprocally Convex Approach to
Asymptotic Behavior of Stochastic
Hopfield Neural Networks
作者:
陈无及, 李青松
关键字:
Hopfield神经网络, 有界性, 均方指数稳定, 变互式凸组合方法Hopfield Neural Networks, Boundedness, Mean Square Exponential Stability, Reciprocally Convex Approach
期刊名称:
《Advances in Applied Mathematics》, Vol.3 No.2, 2014-05-16
摘要:
本文通过构造合适Lyapunov-Krasovskii泛函,采用变互式凸组合方法、自由权矩阵方法、随机分析技巧、线性矩阵不等式(LMI)方法,研究了一类随机Hopfield神经网络的随机一致最终有界性、随机吸引子的存在性以及均方指数的稳定性。我们得到了随机Hopfield神经网络的新的结果和时滞依靠的充分条件。最后,应用MATLAB数值模拟检验来得到了理论的有效性。
In this paper, the
problems of stochastic uniformly ultimate boundedness, the existence of stochastic
attractor and mean square exponential stability for stochastic Hopfield neural
networks are studied by constructing proper Lyapunov-Krasovskii functional,
utilizing the reciprocally con- vex approach, the free-weighting matrix method,
some stochastic analysis techniques and linear matrix inequalities technique (LMI). We deduce novel results and
delay-dependent sufficient conditions for stochastic Hopfield neural
networks. Finally, numerical simulations are given on MATLAB to verify the
effectiveness of the gained results.