Research On Multi-agent Consistency Under Complex Conditions

With the development of Multi-agent system,the scale of its system increasing gradually,control difficulty and communication traffic and so on increase exponentially.Traditional design thought for single agent is difficult to solve such problems.Reflected groups coordination by the biological groups on the nature such as ant colony,bee swarm and so on,which brings inspiration for Multi-agent research,thus Multi-agent arises at the moment.Comparing with the intelligent individual,Multi-agent system have the advantages of simple structure,good expansibility,good flexibility and good economy.So it has a good prospect for development.Consistency refers to the change of the number of intelligent objects with time,all of the behavior movements tend to be consistent eventually.In this paper,a key problem in Multi-agent cooperative control,namely the problem to improve related consistency algorithm has been discussed,The main work has several aspects:(1)The information exchange expression means of Multi-agent was introduced in this paper by using the Graph Theory and the Matrix Theory,on the basis of interaction matrix,Multi-agent first-order system model?the second order system model and its corresponding consistency algorithm were given,the algorithm makes the Multi-agent constraint between each other,finally the Multi-agent behavior is consistent.(2)On the basis of Multi-agent system consistency algorithm theory,respectively discuss the adaptability of Multi-agent consistent algorithm of the first and second order system to the noise interference and the system disturbance.Research shows the algorithm can't make the Multi-agent form consistency eventually.In order to solve this problem,put forward the corresponding improved algorithm.Numerical simulations verify that the improved algorithm enables the system to be consistent in case of noise interference and System disturbance.A typical example is given that when five agents were angle error of 50 degrees,motion direction of 5 agents are consistent after 4 seconds by using this improved algorithm,the Angle error between each agent is not more than 5 degrees.(3)First,Multi-agent formation first-order system model,second order system model were given in the chapter,then consistency algorithm of the first-order system and the second-order system with noise interference and system disturbance were applied to the formation control of Multi-agent.After analysis,Multi-agent formation trajectories have large volatility,bumpy motion,it can't reach agreement.Then the corresponding improved algorithm was proposed,the improved algorithm was used in the formation behavior again,comparing simulation diagram with original algorithm,illustrates the rationality of the improved algorithm effectively.