基于地理加权和元胞自动机的建设用地发展模拟

Simulation of Construction Land Development Based on Geographic Weighted Regression and Cellular Automata

为了防止建设用地的肆意扩张,缓解经济和生态平衡之间的矛盾,为政府部门的前期规划工作提供更合理的借鉴。为此以大连市某区域为研究区,将地理加权回归的思想融入到元胞自动机中,实现了城市空间发展的模拟,并与二次logistic回归模型作了对比。结果表明,在土地利用格局与影响因素建模过程中,存在着不同程度的空间自相关性,并在空间结构上不是单一趋势分布的,体现着空间不稳定性。将地理加权回归和元胞自动机的结合,不仅体现了空间分异特征,而且模拟精度也有所提高。

For the purpose of preventing unlimited expansion of construction land, alleviating the contradiction between economy and ecological balance, and providing reasonable reference for the preliminary planning of the government, this paper selects certain areas of Dalian City as a case study to build geographical weighted regression thoughts into the cellular automata, thus realizing the simulation of urban spatial development, and contrasts the simulation result with logistic regression model. The result indicates that in the process of land use pattern and influence factor modeling, there are different degrees of spatial autoeorrelation and that the spatial structure is not a single trend of distribution, reflecting spatial instability. Combining geographical weighted regression with cellular automata not only embodies the features of spatial distribution, but also improves the simulation accuracy.