Research On Inconsistency In The Process Of Multi-agent Cluster

In recent years,with the rapid development of multi-agent system and cluster self-organization,the problem of multi-agent cluster inconsistency has become a hot topic in the control field.As the basic problem of cooperation and coordination among multi-agents,the problem of cluster inconsistency of multi-agents has attracted more and more attention from researchers in various fields.It is not only of great significance for the survival and reproduction of biological groups in nature,but also plays a very important role in controlling intelligent swarm robot systems.This paper makes a comparative study of the multi-agent communication topology interconnected cluster motion model and its improved model and the opinion cluster model based on opinion dynamics to explore the inconsistent interval in the clustering process,the causes and the advantages of the improved model.The main contents are as follows:(1)Based on the Lennard-Jones(L-J)model,a multi-agent communication topology interconnected cluster motion model is established to study the inconsistency in the process of multi-agent clustering,and the model is improved.The number of nearest associations K is randomly selected within a certain range,and the inconsistency in the process of multi-agent clustering is further studied.The results show that the random processing of the nearest association number K can effectively reduce the occurrence of inconsistent phenomena in the process of multi-agent clustering.(2)In order to further verify that the multi-agent communication topology interconnection cluster motion can reduce the occurrence of inconsistent phenomena in the cluster process,we compare the multi-agent communication topology interconnection cluster motion in which the nearest association number K is random in a certain range with the opinion cluster based on Hegselmann-Krause(H-K)model.The results show that,compared with the H-K model,the occurrence of non-consistent phenomena in the movement of random multi-agent communication topology interconnection clusters in the range of the nearest association number K is significantly reduced,and the whole cluster process shows three stages from micro-aggregation to inconsistent phenomena to tight aggregation,which is more in line with the law of social networks in reality.(3)Inspired by the opinion cluster Friedkin-Johnsen(F-J)model,all individuals in the multi-agent cluster are endowed with a certain degree of stubbornness,that is,the stubborn opinion,to explore the influence of stubbornness on the movement of interconnected clusters in multi-agent communication topology,and compared with F-J model.The research shows that when the number of individuals is the same,the degree of stubbornness is the same,and all individuals participate in the interconnection between the multi-agent communication topology interconnected cluster motion and the opinion cluster motion based on F-J model,the former can still keep the cluster in the final aggregation state.It is proved that the former is relatively stable than the latter.At the same time,it is proved from the side that the former can effectively reduce the occurrence of inconsistency.In this paper,through the comparative study of the multi-agent communication topology interconnected cluster motion model and its corresponding improved model and the opinion cluster model,it is found that the former can reduce the occurrence of inconsistent phenomena through parameter control in the clustering process.At the same time,the movement of multi-agent communication topology interconnected cluster is more stable.