Based On The Overall Multiagent Modeling And Simulation Method And Its Application

Complex science study has now become one of the main direction of system science,this study start at the middle of 1980's and has get many useful research outcome. Now the steps of this study become more and more quick since the appearance of swarm simulation platform. Our country's study on complex science was early,but the upsurge came after 1997. With the supportof the national nature science fund complex special item-((Study of Multiagent-BasedEnsemble Modeling and Simulation Method and it's Application))and college research item --((The study of military complex adaptive system theory ,method and it's application)),the author has studied the complex system's modeling and simulating method,optimizing technique and it's economic application in-depth.The main research contents and contributions of this paper are as fellows.1 Aim at the modeling difficult and problems on the study of complex adaptive system theory and it's application in economic system,based on the complex adaptive system theory,ensemble modeling idea and agent-orient simulation,the author suggest a multi-agent based ensemble modeling and simulation method,the author analyzed this method's structure carefully,gave a careful define and explain of this modeling and simulation step.2 The author suggested a multilayer genetic algorithm,gave an optimistic method of the complex system model on the model,agent and parameter levels.3 Aim at the agent's structure and character,the author suggested three methods to advance the agent's intelligence and adaptability and the ability of express complex model. The first method is adopt evolving technical to optimize agent's rule. In the second method,refer to the idea of thick granularity model and fine granularity model used in the parallel genetic algorithm,aim at the changed dynamic environment of complex adaptive system,the author suggested an environment-select based change granularity model. Through the evolving analyzing of agent's position and structure,studied the influence of the agent's position and structure change on the evolving of agent and model. The last method,through change the relation of agents,such as game relation,benefit matrix and function,gained the intelligence agent who can adapt the soft environment composed of neighbor agent,simulated the complexgame phenomenon of economy agent in economic system.4 Aim at the complex system's character that the parameter is too much and the relationship is too complexity,the author suggest adopting genetic algorithm based multiparameter optimize technique to solve the coordinate,control and optimize problem of parameters.5 Refer to the ideas of((artificial life)} and ASPEN model,using this methods and techniques suggested above,the author gave an evaluting example of aritifical society which develop from the single resource collection and autarkic society to a society with single market and trade ability,Studied and explained the typical phenomenon in life. Such as the move rule of agent under the driving of benefit,competition and conflict between agents,the influence of pollution on agent's moving,the burthen of environment,the population dynamic of agent with propagate ability,the agent's trade behavior and rules under the multi-resource condition,the function and ability of market,the reason and laws of the price fluctuating. This give the economy policy constitutor and decision-making a reference and useful suggestion,also it is a testifying of our methods.