Some Novel Control Schemes For Consensus Of Multi-Agent Systems With Delays

The thesis studies the novel control strategies for consensus of multi-agent systems with delays.Both inherent dynamical delays of the agents and the coupling delays in the agents' network topology are considered in the model formulation.We consider two problems,namely,the synchronization of delayed neural networks and synchronization of multiplex networks.In these problems,we address a scenario where the coupling strength is time-varying and intermittent.One topic of our study focuses on the synchronization of delayed neural networks with general intermittent.By representing the intermittent controller as a controller with time-varying control strength,the synchronization problem is recast into the stability problem of a delayed system with a time-varying coefficient.The main theorem is established on the basis of a novel Halanay-type inequality,from which the largest admissible can be determined.The proposed framework encompasses many existing intermittent policies and provides good generality.It covers intermittent control with an arbitrary number of switching in each control period,occasional control failure,and pulse-modulated intermittent control,and useful corollaries are deduced.The other topic is the synchronization of the multiplex network from a control theoretic perspective,where pinning synchronization of multiplex network with intermittent control is considered.By representing the intermittent controller with on-off control strength,the delayed intra-coupling and non-delay intra-coupling networks are driven to synchronous state.The synchronization condition is established in terms of the average of the control strength on the basis of Halanay-type inequality.Thereafter,the framework was extended to intermittent control with multiple switch control strength,and the corresponding corollary is deduced.The control schemes proposed in this work are more flexible and general,and thus have potential applications in studies on multi-agent systems.