Consensus Problem Of Multi-Agent Systems Based On No Relative State Information

The problem of multi-agent consensus is one of the hot topics in the field of multi-agent systems research.At present,most methods for achieving consensus in this field involve agents communicating with each other,updating their states by obtaining absolute or relative state information about themselves and their neighbors.However,in practical applications,it may be difficult to obtain absolute or relative state information about neighboring agents,or communication between agents may not be possible and communication between neighbors may not contribute to achieving system goals.In such cases,existing methods for multi-agent consensus will no longer be applicable.To address these issues,this thesis proposes two methods for achieving multi-agent consensus based on the absence of relative state information.Firstly,this thesis applies the incremental control method to the multi-agent consensus problem.Under certain initial conditions,the control variables of neighboring agents are used as control inputs to update the state of each agent,which can transform the system into a linear consensus problem.The proposed algorithm is theoretically proven and applied to a supply chain inventory system.Specifically,due to the complex stochastic process,accurate inventory levels in the system will be difficult to obtain.Therefore,using incremental control inputs can achieve consistent inventory levels for the corresponding factories,stabilizing the overall market supply.The effectiveness of the theoretical method is verified through simulation experiments.Secondly,a multi-agent consensus method based on third-party information feedback is proposed.In the case where agents cannot obtain absolute or relative state information from neighboring agents due to the lack of communication between them,a third-party centralized unit is introduced to provide information feedback.Agents only update their own states based on their own information and the comprehensive information feedback from the third party,thus achieving system consensus.This thesis first proposes two algorithms for achieving consensus based on third-party information feedback between two agents,and experimentally analyzes the convergence of the system.Based on this,several algorithms for achieving consensus based on third-party information feedback between multiple agents are proposed,and the influence of the third-party feedback function and related parameters on the convergence of the system is analyzed.In addition,the algorithm is applied to the beamforming problem to set appropriate phase control for each sensor to maximize the received signal power.The experimental simulation results show that the proposed algorithm can achieve multi-agent consensus without relative information between agents.