带有混合时滞的中立型随机神经网络的状态估计

State Estimation for Neutral Stochastic Neural Networks with Mixed Delays

研究了一类具有Markov跳变参数的中立型神经网络的状态估计问题,所考虑的神经网络含有Markov模态依赖的离散时滞和无穷分布时滞。通过构造模态依赖的Lyapunov-Krasovskii泛函,得到状态估计误差系统全局渐近稳定的充分条件,从而保证了所考虑神经网络的全阶状态估计器的存在。通过将非线性耦合的矩阵不等式转化为线性矩阵不等式,利用线性矩阵不等式计算状态估计器的增益矩阵,由此解决了神经网络的状态估计问题。最后,通过数值例子说明所提出设计结果的有效性。

This thesis is devoted to studying state estimation problems for a class of the neural network of neutral-type with Markovian parameters.The neural network under consideration involves discrete time delays and infinite distributed delays,which are dependent of Markov chain.The objective of this thesis is to obtain a sufficient condition for the global asymptotic stability of the state estimation error system by constructing a mode-dependent Lyapunov Krasovskii functional,which guarantees the existence of the full order state estimator of the neural network under consideration,having estimated the state of the neutral-type neural network by analyzing the globally asymptotic stability of estimate error system in mean square,designing the state estimator.Finally,numerical examples are presented to demonstrate the effectiveness of the theoretical results developed in this thesis.