【作者】 罗平；

【导师】 杜清运；

【作者基本信息】 武汉大学 ， 地图学与地理信息系统， 2004， 硕士

【摘要】 元胞自动机(Cellular Automata，简称CA)充分体现了“复杂结构来自于简单子系统的相互作用”这一复杂性科学的精髓，非常适用于具有复杂时空特征的地理系统模拟。但CA模型并不起源于地理系统研究，其空间概念具有更广泛的外延，是一种基于几何特征和空间介质均匀的理想化模型，而地理系统具有典型的空间非均质、几何和非几何属性共同作用等重要特性，因此标准CA模拟复杂地理过程的能力受到了极大限制。为了更真实有效地模拟地理过程，标准CA模型必须扩展，一些学者做了很多有意义的理论尝试。这些研究极大地促进了CA的地理过程模拟能力，但是这些扩展大都是基于几何特性进行的，只是部分地解决了地理系统模拟的真实性问题。论文首先从经典地理过程分析的基本理论入手，从理论上分析和阐述了CA对于经典地理过程分析概念的表达程度，从而得出CA模拟地理过程的局限性实质；接着提出了基于地理特征概念的元胞自动机模型(GeoFeature-CA)，并从理论上分析了其解决元胞自动机模拟地理过程局限性的可能性。鉴于城市土地利用演化是元胞自动机在地学应用研究的一大热点和其本身具有的研究意义，论文选择其作为GeoFeature-CA理论模型的实证研究对象，构建了城市土地利用演化仿真模型(GFCA-Urban)，并以深圳特区为例进行了应用研究。最后，论文得出了一系列研究结论。 本研究的主要内容和成果如下： 1 元胞自动机地理过程模拟机制 对比CA与经典地理过程研究模型，模型表达的实质均是空间、梯度、流和空间关系。CA模型通过元胞、状态与元胞空间等概念实现了地理空间的基本表达，元胞空间与基于几何定位的地理空间概念有类似的意义，而元胞状态则是CA模型中唯一的地理变量。CA模型通过邻居概念实现了地理梯度的有效表达，邻居规则与地理梯度的实质均是确定空间相互作用对象，其差别在于邻居规则是一种基于几何特性的选择方式，而地理梯度是基于几何和属性综合来选择作用对象。CA模型通过元胞局部演化规则实现了地理流的有效表达，元胞流的方向由邻居规则确定，流的强度强烈地依附于元胞距离，这与空间相互作用模型中的距离衰减极为相似。需要注意的是，与地理流区别，元胞流的速度是均质的，不受元胞空间位置的限制，只与元胞空间划分和时间演化周期有关系，而且元胞流的组合方式极为复杂，理论上只要演化时间足够，元胞空间内任意两元胞之间都有流存在，这与地理流受交通方式和地理抗阻限制有很大区别。CA中的空间关系主要通过元胞构形来表达，构形隐含了几何和属性两种类型的空间关系，从几何上描述了元胞在元胞空间中的绝对位置、元胞间的相对位置、距离、方向等空间关系，从属性上描述了元胞状态的空间分布组合和空间目标之间的属性相关。 2 元胞自动机地理系统模拟的局限性实质 CA在地理系统模拟中的局限性实质是元胞地理特征表达不够，无法描述更深层次的地理属性信息。具体表现为四个方面:一是地理空间单元描述的限制，无法描述更高空间层次的位置，也无法描述属性意义上的其它地理位置;二是基于邻居概念的局部空间关系描述的限制，基于纯几何特性来定义邻居不能真实地反映局部的空间相互作用和空间关系;三是元胞流介质均匀的假设前提限制;四是局部演化规则的局限性。在CA中，元胞状态转换规则适用于所有元胞，这实际上是以地理空间各处遵循同样的规律为假设前提的。 3基于地理特征的元胞自动机概念模型(GeoFeature一CA) 地理特征元胞自动机的核心思想是将地理实体的几何和非几何属性综合引入元胞自动机，对其元胞、邻居模型和局部演化规则三个方面进行了扩展:(1)利用元胞实体概念代替元胞，丰富地理空间概念的描述。(2)基于几何和属性综合选择邻居。(3)引入属性控制的元胞局部演化规则。基于地理特征的以因引入属性控制，使在空间介质非均匀条件下研究流有了可能。 4基于地理特征CA的城市土地利用演化仿真模型(GFCA一Urban) 城市土地利用演化的实质是人为干预下，城市生态景观自组织机制的作用过程，从逻辑上看，模型较好地体现了城市土地利用演化系统的本质规律。模型具体构建时，以城市生态机制理论为基础，结合城市土地利用演化的具体情况，充分考虑了城市生态景观、自组织机制和认为干预调控等因素，提出了城市土地利用演化的“生命机制”和 “欲望”等概念，比较真实自然地反映了城市土地利用演化的实质。 基于地理特征CA的建模思想对于增强城市土地利用演化CA模型的实用性具有重要作用。虽然自下而上的模拟思路是城市复杂系统研究的基本趋势，但客观而言，单纯基于自组织机制的城市土地利用演化CA模型很难真实地模拟城市土地利用演化状况。主要原因在于城市系统是典型的非均质地理系统，其非均质性体现在城市演化影响因素的各个方面。因此，只有引进地理特征概念，才有可能比较完整地描述最基本的空间研究单元，自下而上的复杂系统研究思路在城市土地利用演化研究中才具有实用价值。鉴于城市中各土地单元的建设适宜性、时间、交通和规划等属性值存在差异性，为了在微观层次正确地描述各土地单更多还原

【Abstract】 Cellular automata show thoroughly the essence of complexity science that complicated construction comes from the interaction of the simple sub-system, so cellular automata is fit to study geographic system with spatial-temporal feature very much. But cellular automata did not originate the study of geographical system; its spatial concept has the more extensive meaning because it is a idealist model based on geometry feature and symmetrical spatial medium. But the geographical system has some typical characteristics including non-symmetrical space, geometry attribute, non-geometry attribute, etc. As a result, the ability of cellular automata of simulating complicated geographical process is limited relatively. In order to simulate geographical process effectively, standard cellular automata must be extended, some scholar did a lot of meaningful academic trial. These researches promoted the ability of cellular automata of simulating geographic process, but these expanding on cellular automata model were based on geometry feature, and did nothing but working out partial problems of simulating geographical system truly. At first the thesis analyzes and expounds expressive degree of cellular automata model to the concept of classical geographical process analyzing based on basic theory of classical geographical process analyzing, and finds out the limiting essence of cellular automata simulating geographical process; In succession the thesis puts forward the conceptual model of cellular automata based on geographical feature (GeoFeature-CA), and analyzes theoretically the ability of GeoFeature-CA in settling the limitation of standard cellular automata. Because urban land use evolvement is a important field in geography studying, the thesis selects it as demonstrative studying object of GeoFeature-CA model, and constructs simulative model of urban land use evolvement (GFCA-Urban), and take Shenzhen special area as an example to experience applicative study. Finally, the thesis draws a series of studying conclusions. The main contents and innovations of the thesis are as follows:1 The mechanism of simulating geography process of cellular automata The expressive essence of both cellular automata and classical geography process studying model is space, grads, flow and spatial relation. Cellular automata model express the concept of geographic space by cell, state, cellular space, etc. Cellular space has similar meaning with geographic space concept based on geometry, and cellular state is the only geographic variant. Cellular automata represent the concept of geographic grads by the concept of neighbor. The essence of both the rule of neighbor and geographic grads is selecting object of spatial interactions, their difference is that the rule of neighbor is a kind ofselective way based on geometry feature and geographic grads is a kind of selective way based on both geometry feature and non-geometry feature. Cellular Automata model express the concept of geographic flow effectively by local evolve rule of cellular state, and the direction of cellular flow depends on neighbor rule, and the strength of cellular flow depends on the distance between cells strongly, it is similar with the attenuation of strength of spatial interaction with distance. It must be noticed that there is discrepancy between cellular flow and geographic flow. The speed of cellular flow is symmetrical, and isn’t limited by cellular spatial position, and just is relative with the standard of dividing space and time. Moreover the way of combination of cellular flow is complicated very much; theoretically there always is flow between two cells as long as there is enough time. The spatial relation of cellular automata is expressed mainly by framework and figure of cell that hides and includes spatial relation of both geometry and non-geometry. The framework and figure of cell express the absolute position of cellular space, relative position, distance, direction and topologic relation by geometry feature, and express the spatial distributing and comb更多还原