离散时滞动力学网络的拓扑辨识

The Topology Identification of Discrete-Time Dynamical Network with Delay

考虑离散时滞动力学网络的拓扑辨识问题.通过选取合适的参数自适应更新法则,设计有效的响应网络估计器来辨识网络中未知的或不确定的耦合矩阵,并利用Frobenius矩阵范数和Lasalle不变原理证明了该方法的可行性.值得注意的是,网络的耦合矩阵可以是有向加权的. 2个数值例子验证了该方法的有效性.

Topology identification of discrete-time dynamical network with delay is considered. The effective response network estimator through choosing proper parameter adaptive updating laws is designed to identify unknown or uncertain coupling matrix. The method is proved to be effective according to Frobenius matrix norm and Lasalle’s invariance principle. It’s worth noting that the coupling matrix of network can be directed and weighted. The method is verified by two examples.