Consensus Research Of Multi-agent Systems Under Random Switching Topologies

MASs are widely considered in the engineering field.The coupling of agents can be modeled as the graph.The relationship between nodes can be named as the topology.Each node represents a MAS and edges represent the relationship between nodes.Graphs can be connected or unconnected,undirected or directed.According to the properties of multi-agent systems,graph theory or Lyapunov method is usually used to solve the problem of multiagent.In practical applications,the switching between the topologies of multiagent systems can be either fixed or random.The consensus problem of the MAS with stochastic topoloies in this paper is studied.To solve this problem,we use the Lyapunov function,which is more general.The main research work done and the research results obtained are as follows:Firstly,we considered the MASs with the pinning controller.There are protocol failures in alternate time intervals.Then,the consensus is guaranteed.And two sufficient conditions for a pinning controller with fault design are established.By controlling some nodes,the consensus of the whole multi-agent system can be achieved.Finally,the dwell time of the system in the case of protocol invalidation and protocol validity is fully considered.Then,we considered the consensus problem of linear multi-agent systems with random switching topologies.The dwell time of the system consists of fixed and random parts.Different from the traditional study of topological switching signals,the rapid topological switching can avoid certain pressure on the system in practical application.Then,three control protocols are used by using the mode-dependent Lyapunov function method.Finally,we considered MASs with semi-Markov switching topologies.The dwell time of its topologies is an arbitrary probability distribution.In does not directly consider the continuous time under the condition of multi-agent system,construct a discrete markov jump system to ensure the consensus of the original system.Two kinds of consensus solvability conditions is gotten.This reduces the computational complexity.And solvability conditions does not contain the kronecker product of matrix calculation conditions,so the condition has less computational complexity.