切换拓扑下耦合复值神经网络的二分同步

Bipartite synchronization of coupled complex-valued neural networks with switching topologies

在切换有向符号图下,研究复值神经网络的二分同步问题.提出基于观测器的分布式控制协议,使用非分离方法获得耦合复值神经网络二分同步的判别准则.分别针对切换过程中部分子网络不存在有向生成树(frequent directed spanning tree, FDST)和每个子网络都存在有向生成树的情况(permanent directed spanning tree, PDST),基于多Lyapunov函数和平均驻留时间方法分析误差系统的收敛性.最后,通过数值仿真分析理论结果的有效性.

The bipartite synchronization problems of complex-valued neural networks are studied under switching directed signed graphs.By proposing a distributed observer-based protocol,some criteria of bipartite synchronization for coupled complex-valued neural networks are obtained based on non-separation method.Furthermore,the convergence of the error systems is analyzed by the methods of multi-Lyapunov functions and average dwell time.The configurations with frequent or persistent directed spanning trees Two network topologies with FDST and PDST are separately considered.Finally,the validity of the theoretical results is analyzed by a numerical simulation.