From modelling to experiment

Mathematical models applied to urban and regional planning have been widely developed during the sixties. Since that time the scientific and technologic developments have deeply transformed the field of spatial modelling. There has been a reaction against the idea that reality could be reduced to deterministic models. The paradigms of complexity, chaos, self-organisation, fractal geometry have made obvious the unpredictability of complex socio-economic systems. At the same time the progress of computation has led to the substitution of simulation methods to analytic solutions of mathematical models. In such a context, models are loosing in generality and reproducibility what they earn in adaptation to empirical situations. An important challenge is also to confirm the pertinence and specificity of the geographical approach. In that respect the spatial analysis programs must prove the evidence of a common methodology dealing either with physical or human and economic domain. We are working, for instance, on cellular automata programs applied to the historical evolution of an urban space and also to the run-off process in an elementary basin. The spatial structure of the models may be slightly different: rectangular or hexagonal tessellations in the “Human Geography” program, TIN structure, closer to the physical reality, in the other. The relations between the cells may also differ: they are often defined by a distance matrix for the socio-economic models, but a contiguity matrix is of course needed for the streaming process. But, beyond these technical differences, it appears that the geographical programs are developed on a macro-level, that is on aggregate statistical units. The elementary particle is always (or should be for a geographer...) a material, spatial unit, unlike the drop of water of the hydrologists, or the individual “agents” of the sociologists' multi-agents systems. The difference between the micro and macro level is not a question of scale, but a difference of logic. The simulation approach has a requisite, which is a need of systematic validation by a permanent comparison with the actual situation, but the objective is not prediction. The scientific concern is, before all, a precise understanding of the past and recent evolutions, more than a forecasting, which escapes to the specific field of scientific research. What is scientific is what can be measured. The possible prediction may rely on the scientific research, but belongs strictly to the domain of intellectual and personal thinking. This revised version was published online in August 2006 with corrections to the Cover Date.     

用格子波耳兹曼方法模拟双重孔隙介质中的流体迁移

作者在本文中介绍了基于格子波耳兹曼模型的双重孔隙介质中流体运移的数值模拟计算方法。我们从格子波耳兹曼碰撞模型出发，利用格子波耳兹曼方程、Chapman-Enskog展开，以及多尺度技术，得到了描述双重孔隙介质中流体迁移的二维扩散方程。利用格子气自动机方法计算该扩散方程，实现了对双重孔隙介质中流体运移过程的数值模拟仿真。数值实验表明，我们所使用的方法正确、有效。     

A minimum-volume oriented bounding box strategy for improving the performance of urban cellular automata based on vectorization and parallel computing technology

ABSTRACT

As an effective tool for simulating spatiotemporal urban processes in the real world, urban cellular automata (CA) models involve multiple data layers and complicated calibration algorithms, which make their computational capability become a bottleneck. Numerous approaches and techniques have been applied to the development of high-performance urban CA models, among which the integration of vectorization and parallel computing has broad application prospects due to its powerful computational ability and scalability. Unfortunately, this hybrid algorithm becomes inefficient when the axis-aligned bounding box (AABB) of study areas contains many unavailable cells. This paper presents a minimum-volume oriented bounding box (OBB) strategy to solve the above problem. Specifically, geometric transformation (i.e. translation and rotation) is applied to find the OBB of the study area before implementing the hybrid algorithm, and a set of functions are established to describe the spatial coordinate relationship between the AABB and OBB layers. Experiments conducted in this study demonstrate that the OBB strategy can further reduce the computational time of urban CA models after vectorization and parallelism. For example, when the cell size is 15 m and the neighborhood size is 3 × 3, an approximately 10-fold speedup in computational time can result from vectorization in the MATLAB environment, followed by an 18-fold speedup after implementing parallel computing in a quad-core processor and, finally, a speedup of 25-fold by further using an OBB strategy. We thus argue that OBB strategy can make the integration of vectorization and parallel computing more efficient and may provide scalable solutions for significantly improving the applicability of urban CA models.     

Exploring the magnitude–frequency distribution: a cellular automata model for landslides

Landslide magnitude–frequency curves allow for the probabilistic characterization of regional landslide hazard. There is evidence that landslides exhibit self-organized criticality including the tendency to follow a power law over part of the magnitude–frequency distribution. Landslide distributions, however, also typically exhibit poor agreement with the power law at smaller sizes in a flattening of the slope known as rollover. Understanding the basis for this difference is critical if we are to accurately predict landslide hazard, risk or landscape denudation over large areas. One possible argument is that the magnitude–frequency distribution is dominated by physiographic controls whereby landslides tend to a larger size, and larger landslides are landscape limited according to a power law. We explore the physiographic argument using first a simple deterministic model and then a cellular automata model for watersheds in coastal British Columbia. The results compare favorably to actual landslide data: modeled landslides bifurcate at local elevation highs, deposit mass preferentially where the local slopes decrease, find routes in confined valley or channel networks, and, when sufficiently large, overwhelm the local topography. The magnitude–frequency distribution of both the actual landslides and the cellular automata model follow a power law for magnitudes higher than 10,000–20,000 m² and show a flattening of the slope for smaller magnitudes. Based on the results of both models, we argue that magnitude–frequency distributions, including both the rollover and the power law components, are a result of actual physiographic limitations related to slope, slope distance, and the distribution of mass within landslides. The cellular automata model uses simple empirically based rules that can be gathered for regions worldwide.     

数字制图技术中元胞自动机的设计与规划

数字制图自执行框架为改善数字制图过程复杂、周期冗长的状况提供了途径,自执行框架的构成需要大量的数字制图元胞自动机作为主体。通过对一体化数字制图处理方法的分析,总结出数字制图处理方法的3种基本形态以及作用流程,并对其中的常见方法进行了分析规划,为数字制图元胞自动机方法库的建立提供了基础。     

建筑物出口的人员疏散研究

根据建筑物出口疏散的特点,提出了一种改进型的元胞自动机模型。研究了影响人员疏散速度的两个因素:人员之间的相互拥挤产生的冲突和人员移动到出口时由于转弯导致的移动速度减慢。使用数学方法推导得出了摩擦冲突函数和转弯函数。通过计算机仿真与实际疏散实验对比发现,在不考虑转弯因素的情况下,仿真结果与实验结果偏差较大,而在同时考虑相互冲突和转弯因素时的仿真结果与实验结果偏差很小,验证了这两个因素是疏散仿真不可忽略的因素。通过分析仿真结果,验证了本文所采用的模型和函数能够准确地反映实际疏散过程,具有较高的实用价值。     

基于应变率场的全球地震活动特征模拟探索

提出利用全球应变率资料模拟全球地震活动特征的基本思想,并作了初步探索。利用GSRM的全球应变率场结果,初步设计了模拟全球地震活动时空分布特征的细胞自动机模型。该模型把地球考虑为一个自组织的整体系统,制定了细胞自动机的演化规则,获得了模拟的人工地震目录。初步的模拟结果基本反映了全球地震活动的主要分布特征,体现了全球构造活动强弱的主要格局,初步达到了利用GPS等实测资料计算的应变率作为细胞自动机网格状态及其改变量来模拟复杂的全球地震活动特征的实验目的。     

Integration of neural networks and cellular automata for urban planning

This paper presents a new type of cellular automata (CA) model for the simulation of alternative land development using neural networks for urban planning. CA models can be regarded as a planning tool because they can generate alternative urban growth. Alternative development patterns can be formed by using different sets of parameter values in CA simulation. A critical issue is how to define parameter values for realistic and idealized simulation. This paper demonstrates that neural networks can simplify CA models but generate more plausible results. The simulation is based on a simple three-layer network with an output neuron to generate conversion probability. No transition rules are required for the simulation. Parameter values are automatically obtained from the training of network by using satellite remote sensing data. Original training data can be assessed and modified according to planning objectives. Alternative urban patterns can be easily formulated by using the modified training data sets rather than changing the model.     

基于耦合的地理模拟优化系统

尽管GIS在涉及空间信息的许多学科和行业有广泛的应用,但其在对过程进行模拟和优化方面存在严重的功能不足。本文提出地理模拟优化系统GeoSOS的概念与实现方法。进一步建立了GeoSOS1.0的模拟优化平台,作为GIS的重要补充工具。包含了三个重要部分:地理元胞自动机(CA)、多智能体系统(MAS)、生物智能(SI)。其核心内容就是根据微观个体的相互作用,达到模拟和优化的目的。根据Tobler地理学的第一定律,提出了GeoSOS的统一的相互作用规则。GeoSOS具备将模拟和优化耦合起来的功能。将动态模拟模型与空间优化模型耦合起来,使得优化的方案具有一定的前瞻性。对比实验结果发现,耦合模型产生的效用值比非耦合模型分别高出4.3%(点状优化)和4.1%(线状优化),表明GeoSOS能够改善优化的结果。