Impulsive Consensus Of Multi-agent Systems With Communication Constraints

As a typical complex system,multi-agent system has rapidly become a hot research field due to its application prospects.As a fundamental problem,consensus problem of multi-agent systems has attracted wide attention,which can be considered as a control problem of networks.It is unavoidable in engineering applications to study the problem that the cost of control system under different control methods.In addition,due to the complexity of real worlds,the communication environment of multi-agent system cannot always be ideal.Considering these problems,this paper mainly focuses on the consensus of discrete-time nonlinear multi-agent systems with communication constraints via impulsive protocols.The main contents can be summarized as follows.Firstly,starting from the research background,this paper introduces the studies on the consensus of multi-agent systems and the development that impulsive control methods applied to complex networks and multi-agent systems.Combining with the researches in recent years,this paper investigates the solutions to the consensus problem of multi-agent with several communication constraints,especially focusing on the problems such as the communication instants subjected to the system state,the information for communication constrained by limited bandwidth and the communication adopts asynchronous eventtriggered strategy.Meanwhile,some fundamental preliminaries are introduced,including graph theory,Kronecker product,general model of consensus problem of discrete-time nonlinear multi-agent systems and discrete-time impulsive systems.Secondly,the consensus problem of multi-agent system with constraints on communication instants is studied,which means the instants when communication occurs are limited by internal or external factors.It can be also described by that only when the states of system intersect with an impulsive surface which is also state-dependent,communication during agents is permitted.Under some suitable assumptions,some constraints are added to the impulsive surface,with which an appropriate mapping can be constructed to transform the consensus problem of state-dependent impulsive multi-agent system into that of fixed-time impulsive multi-agent system,which can take the new system as the comparison system of the original system.Thus,the comparison system can ensure the exponential consensus of original system with its own exponential consensus.Thirdly,we study the consensus of multi-agent systems with constrained states for communication.This problem mainly focuses on the incomplete communication with constrained data transmitted by agents due to system or external constraints.Moreover,in order to reduce the resources used by communication,quantized data instead of real data transmitted through limited digital communication bandwidth can be considered,which must still guarantee the consensus of the multi-agent system.In this paper,quantizers with finite quantization levels are adopted and and different encoders and local decoders are designed for each pair of agents.With the consensus of multi-agent systems guaranteed,the relationship between the system design and the quantized method with encoding-decoding is analyzed in the situation that the maximal quantization levels are time-varying and fixed by pre-design.In addition,the convergence rate is further investigated.Fourthly,the problem of impulse consistency for multi-agent systems with limited communication load is studied.This problem is about that when the multi-agent system uses impulsive protocols,there would be a phenomena that the system will be under high load during communication but remain idle without communication,which is considered as unbalanced resource allocation.By adjusting this imbalance,the maximum load required by the system will be reduced,which can better cope with some situations where the system load during communication is limited.In this paper,a control strategy of asynchronous sampling for state of system in non-communication period and the sampled data is directly used for communication,which can still solve the consensus problem.For the two cases that multi-agent systems with and without leader,different event-based asynchronous sampling strategies are proposed.Moreover,for the case with leader,whether or not the leader adopt the sampling strategy is also discussed and studied.Finally,in view of the advantages and disadvantages in the results of such consensus problems of multi-agent systems with communication constraints,further discussion and a reasonable planning for future work are given.