Research On Bipartite Consensus Of Event-triggered Intermittent Control For Multi-agent Systems

Multi-agent system is a complex dynamic network composed of a series of agents interacting with each other with dynamic behaviors.Multi-agent can achieve consistent goals through internal information transmission in the system and complete complex tasks that a single agent cannot complete.In nature,the unified and coordinated group behaviors formed by the gathering of moving individuals,such as bird migration,fish migration,wolf cooperative hunting and other group behaviors,can be regarded as various multi-agent systems.These group behaviors have attracted the attention of researchers.In the past two decades,when the information transmission network is a simple cooperative network,various consensus studies on cooperative control of multi-agent systems emerge one after another and become increasingly mature.In the past decade,multi-agent systems with cooperative and competitive communication networks like social networks have gradually entered the public’s view.In order to explore the competition-cooperation network,this paper studies the competitive connection of adjacent nodes in the system.According to the network topology,this paper designs the cooperative control protocols on the signed graphs,and finally achieves the goal of bipartite consensus.The cooperative control of the system mainly includes time-triggered control and event-triggered control.In order to save resources and reduce the possibility of communication failure,a hybrid control protocol combining event-triggered and intermittent control is proposed in this paper.The intermittent control can significantly reduce the working time of the controller,and the event-triggered control can effectively reduce the information updating frequency.Combined with the merits of the two,the computational memory and energy consumption of the system are effectively saved.The main contents of this paper are as follows:(1)The bipartite leader-following consensus problem for first-order nonlinear multi-agent systems is studied.Algebraic graph theory and gauge transformation are used to deal with cooperative-competitive networks.By transforming the signed weight connection matrix into an unsigned directed graph,the consistency of nodes with positive and negative weight is unified.For first-order systems,a finite time hybrid control protocol is proposed.By using Lyapunov theory and finite-time convergence theorem,sufficient conditions for finite-time bipartite consensus of first-order multi-agent systems under hybrid control are given.(2)The bipartite consensus problem for second-order nonlinear multi-agent systems is studied.In the second-order systems,the multi-agent model considers two physical quantities,position and velocity,which has higher accuracy with the second-order integrator.In this paper,a finite time intermittent control protocol is proposed,and then event-triggered control is introduced.A mixed intermittent event-triggered control protocol of the second-order model is proposed.Based on Lyapunov theory and finite time convergence theorem,sufficient conditions for finite-time bipartite consensus of second-order systems under intermittent control,mixed control and event-triggered control in corollary are given.At the end of this paper,MATLAB numerical simulation experiments are conducted to compare the control effects of three control methods: hybrid control,intermittent control and event-triggered control,which confirms the advantages of hybrid control.