Security Group Consensus For Heterogeneous Multi-agent System With Cooperative-competitive Interactions Based On Differential Privacy

Recently,due to the advantages of high robustness,low energy consumption and high flexibility embodied in multi-agent systems,it has been favored by many researchers and has been widely used in the fields of smart power grid,smart city and UAV formation control.However,in the real application,the multi-agent systems may be damaged by malicious attack from outside or inside the system.Therefore,it is very important to study the security group consensus of multi-agent systems.At present,most of the research on the security group consensus is based on homogeneous multi-agent systems.Restricted by various factors,a heterogeneous system composed of agents with different dynamic behaviors can adapt to the ever-changing application scenarios.Meanwhile,an increasing number of scholars only consider a single cooperative relationship.But due to the restriction of resources,there will be competitive,even cooperative-competitive between agents.In addition,due to sensor aging and low bandwidth throughput,the existence of delay in the system is inevitable,and will affect the system stability.Inspired by the above related research works,the security group consensus problem of cooperative-competitive heterogeneous multi-agent systems is discussed under the premise of protecting the privacy information of agents.The main contents are summarized as follows:Firstly,the security group consensus problem of a class of cooperative-competitive heterogeneous multi-agent systems with input delay and communication delay is studied.Based on the differential privacy protection strategy,the cooperative-competitive interaction between agents is considered,and a novel security group consensus protocol is designed,which allows agents within a group to cooperate with each other and within different groups compete with each other.Based on algebraic graph theory and system stability theory,sufficient conditions for asymptotic group protocol of heterogeneous multi-agent systems are analyzed.The results show that the condition is related to the system parameters,input delay and the distribution of added noise.Finally,the precision analysis of the convergence values of the system and the differential privacy analysis of the agent are proposed in detail.The simulation results further verify the effectiveness of the security group consensus control and the correctness of the conclusion.Secondly,in order to further verify the correctness of the model and theory,based on the Java Swing technology,the coordinated control simulation system of multi-agent systems based on differential privacy mechanism is designed and completed,and the agents can achieve security group consensus on the premise of protecting its own privacy information.According to the characteristics of multi-agent systems,a formation control model of UAV is designed,in which each UAV can communicate with its neighbors to achieve secure group consensus.The system model fully embodies the autonomy and cooperation of the system,and the results show the feasibility of applying the theoretical results in the practical environment.