项目名称: 基于智能空间聚类和GPU计算的大尺度城市扩张模拟研究
项目编号: No.41301443
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 天文学、地球科学
项目作者: 李丹
作者单位: 南阳师范学院
项目金额: 25万元
中文摘要: 我国正处于快速城市化时期,在地理空间上主要表现为急剧的城市扩张现象,对大尺度区域城市扩张现象进行地理模拟和预测能够了解和掌握大尺度区域的城市扩张规律,及其发展的格局、过程和趋势,对土地资源集约利用、粮食安全、碳循环等方面的研究都具有重要的意义。然而大尺度城市扩张模拟在分区模拟策略、数据粒度效应影响和模拟运行效率等方面的研究还尚待完善。本研究拟针对上述不足,提出建立大尺度城市扩张模拟的理论框架和方法技术体系,为有效模拟该时空过程提供支持。研究内容包括:(1)引入智能算法及社会经济因素到空间聚类算法中,提出合理的空间分区策略。(2)研究空间数据粒度对大尺度城市扩张模拟的模型参数、模拟精度、误差和不确定性等方面的影响。(3)引入GPU高性能计算技术,构建基于数据分块的大尺度并行元胞自动机模型。本项目的研究成果能够有力支持大尺度区域的城市扩张模拟,同时为相关的大尺度地理模拟提供理论和方法技术支撑。
中文关键词: 元胞自动机;城市扩张;空间分区;MAUP;CUDA
英文摘要: Nowadays, China is in a period of rapid urbanization, it mainly shows the fast urban expansion phenomenon in geospatial. Modeling, Simulating and predicting these urban expansion phenomenon in large-scale areas is a significant method to understand and grasp the laws, patterns, processes and trends in such expansion, also it has important significance in the researches of intensive use of land resources, food security, carbon cycle and other related fields. However, there needs more deeply researches in spatial partition strategy, scale effects and the computation efficiency. This study intends to improve above shortages, establish theoretical framework and technology system of large-scale urban expansion simulations, which could support such complex space-time process researches. The study includes: (1) propose a better spatial partitioning strategy, which introduces intelligent algorithms and socio-economic factors to spatial clustering algorithm. (2) research the spatial scale effects of model parameters, simulation accuracy, error and uncertainty in large-scale urban expansion simulations. (3) introducing the GPU high-performance computing technology, it could construct a parallel cellular automata model based on large-scale data block. The results of this research could effectively support large-scale urban
英文关键词: cellular automata;urban expansion;spatial partition;MAUP;CUDA