一类离散时间神经网络模型的状态估计  

State Estimation of Discrete-two Neural Networks Model

作　　者：朱彩兰[1] 

机构地区：[1]江海职业技术学院基础部,江苏扬州225101

出　　处：《辽宁工业大学学报（自然科学版）》2014年第5期343-350,共8页Journal of Liaoning University of Technology(Natural Science Edition)

摘　　要：研究一类具有分布时滞和含马尔可夫参数切换的离散时间的神经网络的状态估计问题。所考虑的神经网络有若干有限的态式,并且这些态式的切换服从某个马尔可夫链。通过构造新的李雅普诺夫—克拉索夫斯基泛函设计神经网络的估计系统,导出了状态估计器存在的充分条件。所导出的状态估计器的存在性条件为线性矩阵不等式的形式,可通过使用一些标准的数值方法对所获得的判据进行有效的检验.最后我们给出了数值例子说明所提出的判定条件的有效性和应用性。The problem on a class of discrete-time stochastic neural networks with distributed time-delay and Markovian jumping parameters, was discussed. The neural network under consideration processes finite modes, however corresponding parameter switching is subject to a Markov process. By employing new Lyapunov-Krasovskii functionals and constructing a state estimate system for the neural network, the sufficient conditions on existence of the state estimator was derived, and the explicit expression for state estimate gain matrix was obtained. The sufficient conditions obtained for the existence of state estimator can be expressed in the form of LMI and can be readily checked by using some standard numerical packages. Finally, numerical example is given to demonstrate the effectiveness and applicability of the proposed testing criteria.

关 键 词：离散时间神经网络 分布时滞 马尔可夫链 状态估计 

分 类 号：O177.92[理学—数学]