基于动态约束的元胞自动机与复杂城市系统的模拟

为获得复杂城市系统更理想的模拟效果,提出时空动态约束的城市元胞自动机(CA)模型。用不同区域、不同时间新增加的城市用地总量作为CA模型的约束条件,形成时空动态约束的CA模型,并利用该模型模拟1988—2010年东莞市和深圳市城市扩张过程。结果表明,利用CA模型模拟的1993年城市用地总精度比静态CA模型提高了5.86%,而且模型中的动态约束条件可以反映城市发展的时空差异性。

Dynamic Constrained CA Model for Simulating Complex Urban Systems

This paper develops urban cellular automata model by taking dynamic spatial and temporal constraints into account.Recently,cellular automata(CA) have been increasingly used to simulate urban growth and land use dynamics.Traditional CA models simulate urban growth with explicit transition rules,which are invariant within a region.The static transition rule CA models have limitations in simulating dynamic urban growth in a large complex region.It is considered that the development probability of cells with the same attributes may vary in the spatio-temporal dimensions because of the complexity of urban systems.The constraints are based on the amounts of land consumptions for different subregions and time periods.These amounts of land consumptions were obtained by using remote sensing data.This proposed model was used to simulate the urban growth of the Shenzhen and Dongguan city from 1988 to 2004.The prediction of urban land use in 2010 was also carried out by using this model based on the development trajectory.Overall accuracies and Moran I values were calculated by comparing the simulated patterns with the actual ones from remote sensing data.The analysis indicates that this model can produce better simulation results than the non-constrained CA model.