Analysis Of Aggregation And Decentralized Evolution Behavior Of Collective Dynamic Systems

The Collective motion is currently a hotpot research direction.The aggregation behavior generated by inter-agent interactions is considered as the most common collective motions in nature.All kinds of aggregation phenomena,ranging from bacterial colonies to animal groups and even in human society,are emerged with various dynamic mechanism and purpose.In recent years,researchers in the fields of physics,biology,control and so on have devoted to abstracting the inherent control mechanism from collective motion.They developed analytical model from theoretical analysis,and established numerical simulation models,to perform in-depth investigation and exploration on the collective behavior of complex systems.We analyze the aggregation and dispersion phenomena in the evolution of several typical flocking dynamic models with various aggregation phenomena quantitatively,including the Vicsek model and the Boid model.And introduce the concept of clustering coefficient in complex systems and distribution coefficient to measure the degree of aggregation and spatial distribution of the system,respectively.Through detailed numerical simulation experiments,the effects of field radius,velocity,population size on the aggregation degree of the system in noise-free case are studied,and the optimal parameter combinations are revealed to reappear the most concentrated/dispersed motion pattern.Next,we investigate the effect of noise on group aggregation motion and found that,moderate noise can actually promote the aggregation behavior.Besides,we show that the optimal view vanishes and the relation between aggregation and view angle turns into positive correlation.The results in this paper not only brings about a deeper understanding of aggregation behavior in nature,but also have promising potential applications in guiding the control of aggregation and dispersion behavior in multi-agent system.The academic significance of this paper is that in the field of biology,it can help us explore the complexity of biology and reveal the spontaneous aggregation and dispersion of bird flocks and fish school.In engineering field,by applying the enlightenment from natural world to practical application,the distributed strategy will be applied to the coordinated control of autonomous multi-agent systems(such as multi-robot or autonomous unmanned aerial vehicle systems).By designing certain control laws and methods,the whole system will present desired group behavior,which has important application value for tasks such as unmanned system cooperation and disaster relief.