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.     

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 netowrks 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.     

城市扩展元胞自动机多结构卷积神经网络模型

传统的城市扩展元胞自动机(CA)模型是基于单个元胞的变量信息挖掘来构建转换规则的。针对这一问题,本文基于多结构卷积神经网络提出从区域特征出发且顾及区域多尺度特征挖掘转换规则的城市扩展元胞自动机模型(MSCNN-CA),并以武汉主城区和上海浦东新区为例,模拟了两个试验区2005—2015年期间城市扩展过程。模型验证表明:与逻辑回归和神经网络相比,本文构建的3个单一结构的卷积神经网络元胞自动机(CNN-CA)模型在4个指标(Kappa系数、FoM(figure of merit)值、命中率(h)和错误率(m))上都有不同程度的提高。特别是FoM指数,在武汉主城区提高了23.3%~29.4%,在上海浦东新区提高了20.3%~28.5%。此外,MSCNN-CA模型与3个单一结构的CNN-CA模型相比,在各个指标上也有所改善,FoM指数在武汉主城区提高了0.8%~4.8%,上海浦东新区提高了2.8%~7.8%。两个试验区的模拟结果表明:相比传统CA模型,基于多结构卷积神经网络的城市扩展元胞自动机模型(MSCNN-CA)能够有效提高城市扩展模拟的精度,更真实地反映城市扩展空间演变过程。相比单结构的卷积神经网络CA模型,多结构卷积神经网络CA模型的稳定性和模拟结果准确性有所提升。     

一种优化的基于神经网络的经验ZTD模型

目前,经验对流层天顶延迟（ZTD）模型已经有了飞速的发展,因为它们在使用时无需任何测量的实时地面气象数据,这给GNSS用户提供了极大方便。神经网络技术在实测参数型的ZTD建模中已经取得了一定的成果。与此同时,国内虽然有学者构建了神经外网络的经验ZTD模型,其最大的缺点是忽略了ZTD时间变化且只能单独预报ZTD。本文针对这些缺点构建了优化的神经网络经验ZTD模型。试验结果表明,本文提出的神经网络模型可以分别预报天顶干延迟ZHD和天顶湿延迟ZWD,且具有良好的精度：ZHD的Bias和RMSE分别为-3.7和19.8 mm;ZWD的Bias和RMSE分别为-0.6和34.2 mm。本文的神经网络模型预报的ZHD和ZWD的精度均与目前世界著名的GPT2w格网模型相当。另外,与GPT2w模型相比较,神经网络模型最大的优点就是无需庞大的预存格网数据作为输入,在使用时仅需要知道一个训练好的神经网络即可,该特点为GNSS用户提供了极大的方便。     

Spatial Differences in Multi-Resolution Urban Automata Modeling

The last decade has seen a renaissance in spatial modeling. Increased computational power and the greater availability of spatial data have aided in the creation of new modeling techniques for studying and predicting the growth of cities and urban areas. Cellular automata is one modeling technique that has become widely used and cited in the literature; yet there are still some very basic questions that need to be answered with regards to the use of these models, specifically relating to the spatial resolution during calibration and how it can impact model forecasts. Using the SLEUTH urban growth model ( Clarke et al. 1997 ), urban growth for San Joaquin County (CA) is projected using three different spatial grains, based on four calibration routines, and the spatial differences between the model outputs are examined. Model outputs show that calibration at finer scaled data results in different parameter sets, and forecasting of urban growth in areas that was not captured through the use of more coarse data.     

Simulating urban growth in a megalopolitan area using a patch‐based cellular automata

Although traditional cellular automata (CA)‐based models can effectively simulate urban land‐use changes, they typically ignore the spatial evolution of urban patches, due to their use of cell‐based simulation strategies. This research proposes a new patch‐based CA model to incorporate a spatial constraint based on the growth patterns of urban patches into the conventional CA model for reducing the uncertainty of the distribution of simulated new urban patches. In this model, the growth pattern of urban patches is first estimated using a developed indicator that is based on the local variations in existing urban patches. The urban growth is then simulated by integrating the estimated growth pattern and land suitability using a pattern‐calibrated method. In this method, the pattern of new urban patches is gradually calibrated toward the dominant growth pattern through the steps of the CA model. The proposed model is applied to simulate urban growth in the Tehran megalopolitan area during 2000–2006–2012. The results from this model were compared with two common models: cell‐based CA and logistic‐patch CA. The proposed model yields a degree of patch‐level agreement that is 23.4 and 7.5% higher than those of these pre‐existing models, respectively. This reveals that the patch‐based CA model simulates actual development patterns much better than the two other models.     

Assessing the effect of temporal dynamics on urban growth simulation: Towards an asynchronous cellular automata

Time is a fundamental dimension in urban dynamics, but the effect of various definitions of time on urban growth models has rarely been evaluated. In urban growth models such as cellular automata (CA), time has typically been defined as a sequence of discrete time steps. However, most urban growth processes such as land‐use changes are asynchronous. The aim of this study is to examine the effect of various temporal dynamics scenarios on urban growth simulation, in terms of urban land‐use planning, and to introduce an asynchronous parcel‐based cellular automata (AParCA) model. In this study, eight different scenarios were generated to investigate the impact of temporal dynamics on CA‐based urban growth models, and their outputs were evaluated using various urban planning indicators. The obtained results show that different degrees of temporal dynamics lead to various patterns appearing in urban growth CA models, and the application of asynchronous (event‐driven) CA models achieves better simulation results than synchronous models.     

Integrating cellular automata,artificial neural network,and fuzzy set theory to simulate threatened orchards: application to Maragheh,Iran

Urbanization processes challenge the growth of orchards in many cities in Iran. In Maragheh, orchards are crucial ecological, economical, and tourist sources. To explore orchards threatened by urban expansion, this study first aims to develop a new model by coupling cellular automata (CA) and artificial neural network with fuzzy set theory (CA–ANN–Fuzzy). While fuzzy set theory captures the uncertainty associated with transition rules, the ANN considers spatial and temporal nonlinearities of the driving forces underlying the urban growth processes. Second, the CA–ANN–Fuzzy model is compared with two existing approaches, namely a basic CA and a CA coupled with an ANN (CA–ANN). Third, we quantify the amount of orchard loss during the last three decades as well as for the upcoming years up to 2025. Results show that CA–ANN–Fuzzy with 83% kappa coefficient performs significantly better than conventional CA (with 51% kappa coefficient) and CA–ANN (with 79% kappa coefficient) models in simulating orchard loss. The historical data shows a considerable loss of 26% during the last three decades, while the CA–ANN–Fuzzy simulation reveals a considerable future loss of 7% of Maragheh’s orchards in 2025 due to urbanization. These areas require special attention and must be protected by the local government and decision-makers.     

元胞自动机城市增长模型的空间尺度特征分析

基于元胞自动机模拟城市系统的复杂行为时,空间尺度是一个非常重要的概念,模型的模拟结果往往会随着输入数据的空间尺度变化而发生变化。然而,目前的元胞自动机城市增长模型大多没考虑数据的空间尺度特征,本文拟通过改变模型中输入数据的空间尺度来验证元胞自动机城市增长模型对尺度的敏感性及其空间尺度特征,并以长沙市为例进行实证研究。研究结果表明:元胞自动机城市增长模型只有在一定的尺度范围内才具有较高的模拟精度,并且模型对尺度具有一定的敏感性,因此为了使模型能够具有较高的模拟精度,并较好地反映城市形态特征,应认真选择模型中输入数据的空间尺度。     

Cellular Automata Based Model of Urban Spatial Growth

In the study reported in this paper an attempt has been made to develop a Cellular Automata (CA) model for simulating future urban growth of an Indian city. In the model remote sensing data and GIS were used to provide the empirical data about urban growth while Markov chain process was used to predict the amount of land required for future urban use based on the empirical data. Multi-criteria evaluation (MCE) technique was used to reveal the relationships between future urban growth potential and site attributes of a site. Finally using the CA model, land for future urban development was spatially allocated based on the urban suitability image provided by MCE, neighbourhood information of a site and the amount of land predicted by Markov chain process. The model results were evaluated using Kappa Coefficient and future urban growth was simulated using the calibrated model     

基于数据同化的元胞自动机

提出基于集合卡尔曼滤波(EnKF)的元胞自动机(CA)模型。在CA模型中,由于不同的样本会训练出不同参数值 的转换规则,且获取的转换规则在整个模拟过程中不能改变等原因,误差在模拟过程中会不断累积。本文在CA模型中 引入集合卡尔曼滤波的数据同化方法,建立了基于集合卡尔曼滤波的数据同化CA模型,同化遥感观测数据,根据得出 的同化值修正模拟结果使之向真实情况逼近。利用该模型模拟了广东省东莞市的发展情景(1995年—2005年),实验表 明,与传统CA模型相比,基于集合卡尔曼滤波的CA模型能够融合遥感观测数据,并能更有效地模拟城市扩张过程,达 到良好的模拟效果。     

基于灰色预测和神经网络的城市建设用地量预测

采用灰色预测和NARMA(p，q)递归网络模型预测相结合的方法，对城市建设用地量预测值进行神经网络组合预测，在杭州市的实际应用中得到了较好的结果。     

Modelling urban growth in the Indo-Gangetic plain using nighttime OLS data and cellular automata

The present study demonstrates the applicability of the Operational Linescan System (OLS) sensor in modelling urban growth at regional level. The nighttime OLS data provides an easy, inexpensive way to map urban areas at a regional scale, requiring a very small volume of data. A cellular automata (CA) model was developed for simulating urban growth in the Indo-Gangetic plain; using OLS data derived maps as input. In the proposed CA model, urban growth was expressed in terms of causative factors like economy, topography, accessibility and urban infrastructure. The model was calibrated and validated based on OLS data of year 2003 and 2008 respectively using spatial metrics measures and subsequently the urban growth was predicted for the year 2020. The model predicted high urban growth in North Western part of the study area, in south eastern part growth would be concentrated around two cities, Kolkata and Howrah. While in the middle portion of the study area, i.e., Jharkhand, Bihar and Eastern Uttar Pradesh, urban growth has been predicted in form of clusters, mostly around the present big cities. These results will not only provide an input to urban planning but can also be utilized in hydrological and ecological modelling which require an estimate of future built up areas especially at regional level.     

城市扩展极限学习机模型

城市空间结构及其扩展的模拟是城市科学管理与规划的重要前提,本文基于极限学习机提出了顾及不同非城市用地转化为城市用地差异与强度的城市扩展元胞自动机模型（ELM-CA）。模型验证表明：①ELM-CA模型的模拟精度达到70.30%,相比于逻辑回归和神经网络分别提高了2.21%和1.54%,FoM系数分别提高了0.025 9和0.017 9,Kappa系数分别提高了0.024 7和0.016 9,且Moran I指数接近于实际值,说明极限学习机模型较逻辑回归和神经网络能更有效模拟城市扩展的空间形态及其变化;②ELM模型的训练时间仅为神经网络的1/3左右,体现了ELM学习速度的优势;③在小样本情况下,逻辑回归和神经网络都受到明显的影响,而极限学习机还能保持良好的性能,这个特点使其在样本难以获取的情况下具有明显的优势。两个时相的城市扩展模拟与真实数据的比较表明：基于极限学习机的城市扩展元胞自动机模型（ELM-CA）,简化了CA模型的复杂度,并在小样本情况下能有效提高模拟精度,适合于复杂土地利用条件下城市扩展模拟与预测。     

利用CA模型进行城市空间增长动力的研究--以武汉市主城空间增长过程分析为例

提出了基于扩散生长的状态型CA模型、基于轴线生长的交通型CA模型和基于优势驱动的环境型CA模型等三种CA模型，设定了包围填充、扩散生长、交通延伸、交通连接、交通吸引、优势生长等6种转换规则。模拟结果表明，武汉市主城整体上呈现出“摊大饼”的发展态势，并且扩散到城市地区，近郊优势增长十分明显，导致了大规模的郊区化，呈现出城乡一体化的发展态势。     

GM-BP 组合预测模型在基坑沉降分析中的应用

为提高变形预测的精度,采用GM(1,1)与BP神经网络组合模型进行预测。灰色GM(1,1)模型使用方便,在样本数据较少的情况下能够取得不错的预测效果,但对预测序列存在规律性波动或突变时的预测能力不强;而神经网络模型建模过程相对复杂,需要较多的训练样本,但对于数据存在规律性波动和突变时有很好的预测能力。组合模型融合两者优点,将其应用于基坑沉降数据预测,结果表明,该模型预测精度优于传统的单一预测模型。     

Improving the capability of an integrated CA-Markov model to simulate spatio-temporal urban growth trends using an Analytical Hierarchy Process and Frequency Ratio

The creation of an accurate simulation of future urban growth is considered one of the most important challenges in urban studies that involve spatial modeling. The purpose of this study is to improve the simulation capability of an integrated CA-Markov Chain (CA-MC) model using CA-MC based on the Analytical Hierarchy Process (AHP) and CA-MC based on Frequency Ratio (FR), both applied in Seremban, Malaysia, as well as to compare the performance and accuracy between the traditional and hybrid models. Various physical, socio-economic, utilities, and environmental criteria were used as predictors, including elevation, slope, soil texture, population density, distance to commercial area, distance to educational area, distance to residential area, distance to industrial area, distance to roads, distance to highway, distance to railway, distance to power line, distance to stream, and land cover. For calibration, three models were applied to simulate urban growth trends in 2010; the actual data of 2010 were used for model validation utilizing the Relative Operating Characteristic (ROC) and Kappa coefficient methods Consequently, future urban growth maps of 2020 and 2030 were created. The validation findings confirm that the integration of the CA-MC model with the FR model and employing the significant driving force of urban growth in the simulation process have resulted in the improved simulation capability of the CA-MC model. This study has provided a novel approach for improving the CA-MC model based on FR, which will provide powerful support to planners and decision-makers in the development of future sustainable urban planning.     

耦合遥感观测和元胞自动机的城市扩张模拟

在传统元胞自动机(CA)模型中,静态的模型参数和模型误差不能释放是影响城市扩张模拟效果的两个重要原因。文中引入集合卡尔曼滤波方法到CA模型中,提出了基于联合状态矩阵的地理元胞自动机。该模型在模拟过程中可以通过同化遥感观测数据,动态地调整模型参数和纠正模拟结果,使模型参数能够反映转换规则的时空变化,同时也能较好地释放积累的模型误差。将模型应用于东莞市的城市扩张模拟中,实验结果表明,模型能够准确地调整模型参数使之符合城市发展模式,同时也能有效地控制模型误差,其模拟的空间格局与真实情况吻合。     

变形监测数据的RBF神经网络预测方法

研究了RBF神经网络的变形预测模型及其训练准则和算法，分析了基于RBF神经网络和BP网络的盾构施工变形预测结果，得出了很好的预测效果。     

基于神经网络方法的极化雷达地表参数反演

人工神经网络（Artificial Neural Network)是一个由独立处理单元以一定拓扑结构高度连接而成的并行分布式信息处理结构，适于解决各种非线性问题，积分方程（Integrated Equation Model)单散射模型可模拟各种地表参数条件下裸露地表后向散射系数，以IEM为基础生成训练数据，用L波段的C波段SIR-CHH，VV极化单散射后向散射系数数据为神经网络输入，通过后向反馈（BP）神经网络模型可同时反演得到裸露地表条件下地表介电常数，地表相关长度和均方根高度等地表参数。