Consensus Control For Multi-agent Systems With Impulsive Effects

Multi-agent systems are techniques of distributed artificial intelligence,wildly used in computer science,furnaces,space science,ecological management and so on.Consensus is one of the basic requirement for coordination of multi-agent systems,which has attracted tremendous interests of researchers.In many application scenarios,agents may suffer some abrupt jumps from the interference of external environment,frequency changing,system switching,man-made setting and etc.These abrupt jumps,which can be considered as impulsive effects,have assignable influence upon systems.This dissertation aims to investigate consensus control for multi-agent systems with impulsive effects,including the aspects of occurrence principles of impulse,different impulsive effects and their applications,as well as the properties of multi-agent systems such as: the dynamic behaviors of agents,the connected topologies among agents and the consensus rate of systems.Specifically,the main work and innovative results of this dissertation are listed below:1.Consensus control for nonlinear multi-agent systems with variable-time impulsive effects is investigated.The impulsive moments vary by the current states of agents.A classic distributed protocol is used for the target of consensus.Combining the effect of impulse and protocol on consensus convergence,we divide analysis mainly in three cases:the first one is stable impulse and convergent protocol,the second one is unstable impulse and convergent protocol,the third one is stable impulse and in-convergent protocol.For the first case,systems with impulsive effects can be weaker to systems without impulse,so consensus criteria are obtained without considering variable-time impulsive effects.For the second and third cases,to coping with the analyzing difficulty causing from the variable impulsive moments,B-equivalence method is used for constructing fixed-time impulsive systems as comparison systems of the variable-time impulsive ones.Some conditions are well selected for absence of beating phenomena,on this basis,the relationship of the fixed-time jump operator and the variable-time one is formed.By using the Lyapunov stable theory,we theoretically prove the stability properties of original systems are contained in the fixed-time impulsive comparison systems.Therefore,sufficient consensus criteria are established by analyzing the fixed-time impulsive systems.2.Taking advantage of stable impulsive effects,a variable-time impulsive coordinated tracking protocol is proposed for leader-following nonlinear multi-agent systems.According to the tracking task,the global error impulsive systems are formed,in which:impulsive moments vary with the current states,and jump operators are composed of graphs among agents and states of agents.In order to overcome the analyzing difficulty causing from variable-time impulses,B-equivalence method is introduced for constructing the fixed-time impulsive systems as comparison systems.Thus,some conditions are well selected for absence of beating phenomenon in systems,and then we express the jump operators of comparison systems as a function related to the original systems,and illustrate the two systems have the same stable property.Therefore,variable-time impulsive consensus criteria are established by analyzing the stability of the comparison systems.3.To specify the consensus convergence time,finite-time consensus of second-order nonlinear multi-agent systems with average impulsive effects is investigated.We propose a non-chattering finite-time consensus protocol,which is composed of neighboring and self feedbacks.The neighboring part is designed not only to guarantee consensus of systems,but also to exchange information among agents.The self feedback is designed to specify the consensus time.Different impulsive types: stability an divergence are divided based on strengths of impulse and controller,then by using Lyapunov stable theory and discontinuous analyzing method,we establish different sets of sufficient criteria and setting times respectively in the corresponding impulsive situations.Sufficient impulsive criteria of exponential consensus convergence for systems are additional established.4.More complex second-order nonlinear multi-agent systems are investigated with the aim of finite-time mean square consensus.In the investigated systems,connected typologies among agents are directed and dynamic,being governed by Markov switching,and agents suffer impulsive disturbances.A finite-time consensus protocol without sign function is proposed.Some graph-related matrices playing crucial roles in convergence analysis are constructed by employing graph theory.Using the stochastic and discontinuous stability theory,expectations of multi-agent systems energy evolution are bounded in both continuous and discontinuous time.Then,finite-time mean square sufficient consensus criteria are established and their setting times are obtained.Mean-square exponential consensus criteria are additional obtained by adjusting the control parameters.