Method Explor ation of Geogr aphical Spatial Differ entiation Based on Multi-agent: A Case Study of Urban Residential Simulations
基于多智能体的地理空间分异现象模拟 ———以城市居住空间演变为例

Many geographical phenomena have some spatial differentiation rules. So it is of theoretical significance to research the formation and development of the geographical spatial differentiation. But a top-down approach has difficulty in analyzing evolvement of these geographical phenomena. This paper represents the bottom-up approach based on multi-agent system and reproduces the differentiation of complex geographic phenomena. With technologies of java-swarm simulation, this paper simulates urban system from chaos to order and provides the analysis approach to develop and verify the urban theories. In order to implement a segregate urban spatial structure at macro level emerging from simple rules established at micro level, the model based on multi-agent is constructed as follows: simulated inhabitants wander through the universe to find the optimal place of residence according to constraints as their incomes and perception of the environment. And the model consists of four related components, i.e., environment controller, the city's housing infrastructure, inhabitant agents and rules controller. Negative exponential function (E = e-ax) isintroduced to represent income distribution, where a is defined as a factor of income gap,changing of x can represent the dynamics of the income of agent. The housing infrastructure of the model is a M×M square lattice cells, which symbolize houses. The migration decision of agent depends on the residential pressure. It is suggested that the probability of leaving its current location increases monotonically with an increase in the residential pressure. And the residential pressure of an agent is estimated by its economic tension and environment pressure. On the Swarm platform, Java language is used to realize the model. The spatial differentiation model exhibits the progress of residents' aggregating and the influence of houses' values. Under the different factors of income gap, the influences of spatial differential are compared. Results of simulation show that spatial differentiation is an inevitable result. And widening of income gap makes poorer passivity in choosing house location.