基于案例推理的元胞自动机及大区域城市用地演变模拟

发布时间：2018-05-04 16:02

本文选题：城市用地演变 + 案例推理　；参考：《河南大学》2015年硕士论文

【摘要】：近年来,我国城镇化和工业化进程不断加快,并存在城市用地过度扩张、基础设施重复建设、空间开发布局无序等问题。为了使我国国土空间资源得到充分利用,使各生产要素能够在城市群、省域乃至更大范围的区域层面集约高效地流通分配,国家日益重视区域土地利用规划,不断加强对全国及省域土地利用规划的编制工作,不断加快对区域规划法及相关法律体系的构建,以促进我国区域政策的科学化与规范化,强化区域规划对我国发展的战略指导作用。为满足区域土地利用规划编制需求,我国学者对区域层面的城市用地演变展开研究。按照我国城乡一体化土地分类系统对土地的分类方法[1](划分为农用地、建设用地、未利用地三类),城市用地是指市域范围内的建设用地。目前国内外对城市用地演变的研究大多以单个城市为主,主要关注单个城市内部的城市用地演变机制,通过构建一组统一的用地转换规则来模拟单个城市内的城市用地演变。而区域内部因自然资源禀赋、经济人口本底和基础设施要素等差别迥异,各个城市的用地演变驱动因子和转换规则呈现显著的空间差异特征。因此针对区域层次的城市用地演变分析时,须重视区域内部各个城市的城市用地演变规则的空间差异。这就需要对传统的城市用地演变模拟方法进行改进,以提高区域层面城市用地空间格局模拟的精度。本文研究适用于区域城市用地演变研究的地理模拟模型,对于模拟大区域城市演变、制定区域土地利用规划等具有重要的理论意义和应用价值。本文针对传统元胞自动机(Cellular Automaton简称CA)在区域城市用地模拟中的局限,提出基于案例推理的元胞自动机模型,并以湖南省为研究区对模型的模拟结果进行检验。基于案例推理的元胞自动机模型(CBR-CA)是将案例推理法(Case-Based Reasoning简称CBR)与元胞自动机相结合的一种改进CA模型。主要研究内容包括:1、将模拟单元内的土地类型、特征属性以及空间关系等影响用地转化的因素以地理案例的形式进行组织与描述;2、在全省范围内采样构建案例库,通过案例库中收集的案例深入挖掘学习区域内不同区位的城市用地转化规律并以此指导模型对待测案例的用地状态预测;3、将案例推理方法分析得出的城市用地演变模拟结果引入CA中,最终通过CA迭代来模拟未来城市用地的空间格局。4、以湖南省作为研究区域,研究湖南省城市用地演变规则在空间上的分异规律,运用基于案例推理的CA模型对湖南省1990年到2010年间的城市用地演变进行模拟。5、对CBR-CA模型的模拟精度进行评价,将模拟结果与传统CA模型对湖南省城市用地的模拟结果作对比,以验证CBR-CA模型对于区域城市用地演变模拟的适用性。本文通过分析、讨论得到以下主要结论:(1)采用基于CBR-CA模型对区域层面城市用地演变进行模拟是合理的。传统的CA模型通过构建具体的转化规则模拟城市用地的演变,而CBR-CA模型将区域城市用地演变规律以地理案例的形式表现出来,以案例为载体可以隐含的表达区域内部用地演变规律的空间差异,且免去了构造繁琐复杂的用地转化规则的工作;此外,相比传统的CA模型,CBR-CA模型能够将地理单元的方位、空间距离和用地类型拓扑结构等影响城市用地演变的空间关系应用于用地演变的模拟,从而更加全面深入地表达、提炼城市用地演变规律,更真实地模拟城市用地演变这一复杂地理现象的发展机制。(2)湖南省内不同地区影响城市用地演变的驱动因子不同,用地转化规律空间差异明显,但各驱动因子与城市演变的相关系数存在一定的空间分异规律。对城市用地演变规律的空间差异特征进行研究有助于指导CBR-CA模型合理科学的采样构建案例库,也有助于深入的城市化机制与特征。(3)CBR-CA对区域城市用地模拟结果较好,模拟精度高于传统CA的模拟结果。由模型检验结果表明,从全省尺度来看,基于logistic的CA模型(Logistic-CA)模拟精度为74.2%,而CBR-CA的模拟精度达到82.4%;从地区尺度和县市尺度来看,基于案例推理的模拟结果也都优于Logistic-CA的模拟结果。
[Abstract]:In recent years, the process of urbanization and industrialization in China has been accelerating, and there are many problems, such as excessive expansion of urban land, repeated construction of infrastructure, and disorderly layout of space development. In order to make our country's land space resources fully utilized, the production factors can be circulated intensive and efficient in the regional level of urban agglomeration, provincial and even wider range. Distribution, the state attaches great importance to regional land use planning, constantly strengthens the work of the national and provincial land use planning, and continuously speeds up the construction of Regional Planning Law and relevant legal system, in order to promote the scientific and standardized regional policy, and strengthen the strategic guiding role of regional planning for the development of our country. The demand of land use planning is made by Chinese scholars. The classification of land is [1] (divided into three categories: agricultural land, construction land, unused land) according to the integrated land classification system of urban and rural areas in China (divided into three types of land, construction land, unused land), urban land use is the construction land within the scope of the city. Most of the changes are focused on a single city, mainly focusing on the urban land evolution mechanism within a single city. Through the construction of a set of unified land conversion rules to simulate the urban land use evolution in a single city, the differences in the natural resource endowment, the base of the economic population and the basic elements of the infrastructure are very different in the region, and the use of each city in each city. In order to improve the urban land evolution, it is necessary to improve the traditional urban land evolution simulation method in order to improve the regional level of urban use. The accuracy of the spatial pattern simulation. This paper studies the geographical simulation model suitable for the study of regional urban land use evolution. It has important theoretical significance and applied value for the simulation of large regional urban evolution and the formulation of regional land use planning. This paper aims at the traditional cellular automata (Cellular Automaton for short) in the regional urban land use model. A cellular automaton model based on case-based reasoning is proposed and the simulation results of the model are tested in the study area of Hunan province. The cellular automaton model based on case based reasoning (CBR-CA) is an improved CA model combining case reasoning (Case-Based Reasoning abbreviated CBR) with cellular automata. The contents include: 1, we organize and describe the factors affecting land conversion in the simulation unit, such as land type, characteristic attribute and spatial relationship, and also organize and describe the factors that affect land conversion in the form of geographical cases. 2, the case base is constructed in the whole province, and the law of urban land conversion in different locations in the learning area is deeply excavated through the cases collected in the case base. The guidance model treats the prediction of the land use status of the case. 3, the urban land use evolution simulation results obtained by CBR are introduced into CA, and the spatial pattern.4 of the future urban land use is simulated by CA iteration, and Hunan province is used as the research area to study the spatial differentiation rules of the urban land use evolution rules in Hunan province. The CA model based on case-based reasoning is used to simulate the urban land use evolution of Hunan province from 1990 to 2010. The simulation accuracy of the CBR-CA model is evaluated. The simulation results are compared with the traditional CA model on the simulation results of the urban land use in Hunan province to verify the applicability of the CBR-CA model to the regional urban land evolution simulation. Through the analysis, the following main conclusions are discussed: (1) it is reasonable to simulate the evolution of urban land use based on CBR-CA model. The traditional CA model simulates the evolution of urban land by constructing specific transformation rules, and the CBR-CA model shows the evolution law of regional urban land use in the form of geographical case. The case is the carrier that can imply the spatial difference of the land evolution law in the region, and avoids the work of constructing the complicated and complex land conversion rules. In addition, compared with the traditional CA model, the CBR-CA model can influence the spatial relationship of the geographical location, space distance and the topology structure of the land type. It is applied to the simulation of land use evolution, so as to express more comprehensively and deeply, to refine the law of urban land evolution, and to simulate the development mechanism of the complex geographical phenomenon of urban land use evolution. (2) the different driving factors affecting the evolution of urban land use in different regions in Hunan province are different, and the spatial difference of land conversion rules is obvious, but the driving causes are different. The relationship between the correlation coefficient of the urban evolution and the suburban evolution has a certain spatial differentiation law. The study of the spatial difference characteristics of the urban land use evolution law helps to guide the rational and scientific sampling construction of the CBR-CA model, and also helps to deepen the urbanization mechanism and characteristics. (3) CBR-CA has better simulation results for the regional urban land use, and the simulation precision is good. It is higher than the traditional CA simulation results. The model test results show that the simulation precision of the logistic based CA model (Logistic-CA) is 74.2%, and the simulation precision of CBR-CA reaches 82.4%. From the regional scale and the county and city scale, the simulation results based on case-based reasoning are also superior to the Logistic-CA simulation results.

【学位授予单位】：河南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：F299.23

【相似文献】