Some Agent-Based Model About Spatial Population Transfer

Recent several decades,with the development of nonlinear science and complexity science,another important and universal distribution was discovered. That is the power-law.Power-law consists power-law distribution Zipf's Law and Pareto distribution.Certainly finding out the property of power-law is important, but more important is finding out the underlying dynamics mechanism of power-law.That is the hot problem in academia all the time.This study builds some agent-based model about spatial population transfer in order to do some try in this field and got some meaningful conclusions.Before introduce the new model,this paper reviews the research history and research findings of city formation and rank-size distribution of cities firstly.After that,this paper briefly introduces the concept of power-law,power-law distribution, Zipf's law and Pareto distribution and their relationship.Recently,someone queried the Zipf's law in explaining city-size distributions.In order to find out the true,I give my analysis and judgment by experiment.The result proves that Zipf's Law isn't a simple and widespread existence of statistics phenomenon,but a power-law which have the special dynamics mechanism.For the sake of finding out the mechanism,this paper establishes a spatial and gradual model to simulate population transfer based on sum up past research. So-called gradual model is that the behavior rule of agents is gradual complication in the models.Just in the change process of behavior rule,we find two important conditions of generating Zipf's Law,which are "difference" of agents and "preferential attachment" of behavior rule."Difference" has more concrete requirement.In the meantime,this text also introduces the disagglomerating effect in models.Furthermore,the "city competitive power",which has important influence to population transfer,is abstracted into the model by adding grade into models.We find that,in the process of procedure circulating,there will be some small cities will expand in short time.However,the Rank-Size Distribution of cities obeys Zipf's Law with the existence of small cities expanding quickly all the same.Moreover,other amelioration in this model is joining new agents.First,we want to enrich model series and make the model close to reality by this way. Second,we want to simulate "migrate into province" by this way.Subsequently, we do the empirical research citing Zhejiang province as an example.The conclusion proves that the result of the theories and practice are uniform,the result of empirical research indicate that inflation speed of big city,such as Hangzhou, Ningbo,Wenzhou,is obviously higher than small county.Therefore,we should not only pay attention to the cities with strong competitive power,but also concern to these towns with weak competitive power,which are limited to nature condition. The dissimilarity of the population inflation speed,at a certain degree,reflects the difference of residents' income.This has instruction meaning to make economy development policy of Zhejiang province in future.