Study On The Consensus Control For Multi-agent Networks

Consensus and cooperation problems have a long history of research,and the origin for the study of consensus problems can be traced back to De Groot's research in management science and statistics in 1960 s.At the same time,the researches have been found in various disciplines and fields,such as control theory,multiple sensors systems,distributed computing,management science and statistics,radar tracking,medical experts,flocking theory,rendezvous in space,sensor fusion,congestion control,smart grids,formation control and consensus-based belief propagation,etc.In recent years,the control problems of multi-agent networks have drawn enormous concern from researchers in diverse fields.A multi-agent network is a computerized network composed of numerous interactive intelligent agents in the network.The idea of the multi-agent network can be adopted to solve the tough problem that an individual agent is very difficult to complete,or even impossible to complete.This dissertation mainly focuses on the design of the consensus mechanisms and protocols of multi-agent networks under various situations,and the main works are as follows:1.The event-triggering consensus problem of multi-agent networks with time-varying topology under finite digital communication channel is studied.A novel communication framework smoothly combining the quantization scheme and the event-triggering strategy is designed,which manifestly improves the communication efficiency of the network.In pursuit of saving network resources and reducing the computational load of the microprocessors embedded in all the agents,a specially designed distributed event-triggering strategy is proposed to solve the consensus problem of the multi-agent network with time-varying topology.It is proved that the average consensus of the multi-agent network is asymptotically achived and the state of all the agents evolves well.Some numerical simulations are given to illustrate the correctness of the theoretical results and the feasibility of the proposed protocols.2.The pinning control consensus problem of multi-agent networks with general topology is studied.Based on the characteristics of M matrix,a novel consensus protocol of multi-agent network with general topology is established.The results show that the consensus problem of multi-agent networks under the pinning control scheme depends on the minimum real part of eigenvalues of the matrix,which is the sum of the Laplacian matrix and the control gain diagonal matrix.In this dissertation,the discrete multi-agent network is studied,and the condition that the consensus of the multi-agent network can finally be achieved is derived.The pinning control mechanisms in diverse situations are also studied,including discrete networks and continuous networks.Some numerical simulations are given to verify the correctness of the theoretical results.3.The consensus problem of multi-agent networks with time delay and stochastic perturbation via impulsive control is studied.Based on algebraic graph theory,Lyapunov stability theory and Halanay inequality theory,a novel consensus control protocol of multi-agent networks via impulsive control is designed.And,we also consider the pinning control mechanism,the internal delay and the transmission delay,so the consensus protocol is further more adaptable for the pratical multi-agent network.It is proved that the well designed mechanism can make the multi-agent network with time delay and stochastic perturbation finally converge,and some numerical simulations are given to verify the correctness and validity of the theoretical results.4.A novel intermittent impulsive control mechanism is proposed to make the multi-agent network with time-varying delay achive the consensus.The intermittent impulsive consensus control mechanism can break the upper bound of the impulsive interval in the general impulsive control,thus,it can replace the uninterrupted triggering mechanism.Integrating the intermittent impulsive control mechanism into the multi-agent network with time delay,its impulsive controller functions only in the control window.Based on the algebraic graph theory,Lyapunov stability theory and Halanay inequality theory,the consensus protocol of multi-agent networks with time delay is obtained by the pinning intermittent impulsive control.The relevant numerical simulations are given to verify the correctness and validity of the proposed consensus control mechanism.Finally,the research work during the doctoral period is summarized,and the prospect of the future research is also concerned.