ARTIFICIAL NEURAL NETWORKS: A NEW APPROACH TO MODELING INTERREGIONAL TELECOMMUNICATION FLOWS*

ABSTRACT. During the last thirty years there has been much research effort in regional science devoted to modeling interactions over geographic space. Theoretical approaches for studying these phenomena have been modified considerably. This paper suggests a new modeling approach, based upon a general nested sigmoid neural network model. Its feasibility is illustrated in the context of modeling interregional telecommunication traffic in Austria, and its performance is evaluated in comparison with the classical regression approach of the gravity type. The application of this neural network approach may be viewed as a three-stage process. The first stage refers to the identification of an appropriate network from the family of two-layered feedforward networks with 3 input nodes, one layer of (sigmoidal) intermediate nodes and one (sigmoidal) output node (logistic activation function). There is no general procedure to address this problem. We solved this issue experimentally. The input-output dimensions have been chosen in order to make the comparison with the gravity model as close as possible. The second stage involves the estimation of the network parameters of the selected neural network model. This is performed via the adaptive setting of the network parameters (training, estimation) by means of the application of a least mean squared error goal and the error back propagating technique, a recursive learning procedure using a gradient search to minimize the error goal. Particular emphasis is laid on the sensitivity of the network performance to the choice of the initial network parameters, as well as on the problem of overfitting. The final stage of applying the neural network approach refers to the testing of the interregional teletraffic flows predicted. Prediction quality is analyzed by means of two performance measures, average relative variance and the coefficient of determination, as well as by the use of residual analysis. The analysis shows that the neural network model approach outperforms the classical regression approach to modeling telecommunication traffic in Austria.     

Integration Processes in European Research and Development: A Comparative Spatial Interaction Approach Using Project Based Research and Development Networks,Co‐Patent Networks and Co‐Publication Networks

This study focuses on integration processes in European Research and Development (R&D) by analyzing the spatiotemporal dimension of three different R&D collaboration networks across Europe. The studied networks cover different types of knowledge creation, namely project‐based R&D networks within the European Union (EU) Framework Programmes (FPs), co‐patent networks, and co‐publication networks. Integration in European R&D—one of the main pillars of the EU Science Technology and Innovation policy—refers to the harmonization of fragmented national research systems across Europe and to the free movement of knowledge and researchers. The objective is to describe and compare spatiotemporal patterns at a regional level and to estimate the evolution of separation effects over the time period 1999–2006 that influence the probability of cross‐region collaborations in the distinct networks under consideration. The study adopts a spatial interaction modeling perspective, econometrically specifying a panel generalized linear model relationship, taking into account spatial autocorrelation among flows using eigenfunction spatial filtering methods. The results show that geographical factors are a lower hurdle for R&D collaborations in the FP networks than in co‐patent networks and co‐publication networks. Furthermore, it is shown that the geographical integration is higher in the FP network.     

中原城市群城镇网络结构特征分析:基于空间引力与客运联系

以中原城市群县、市区节点为研究单元,基于引力模型与复杂网络模型,根据测定的城镇间引力量与城际直接的客运班次两类数据,对中原城市群空间网络结构特征进行比较与归纳。结果表明:中原城市群空间联系突出了"双中心"与层次结构特征,引力与客运联系网络互为支撑,又各具特色,郑州与洛阳在城市群网络结构中的中心度较高,但对网络并没有绝对的控制力,城市群网络的集团化效应明显。城市群最大联系轴向有序,但空间网络结构并没有呈现较强的"长尾分布",城镇网络结构中心度分布有序性差。城市群引力网络子群体间的联系较弱,内部的联系密切,受行政区划的屏蔽或行政级别差异影响显著;而客运联系网络各子群体之间的联系较强,核心市区功能效应显著。     

Endogenous Fractal Dynamics at Alpine Treeline Ecotones

Many of the hypotheses proposed to explain ecotones are based on an individualistic paradigm and are essentially static. While they include local feedbacks, they ignore the interactions between pattern and process across scales. These feedbacks in ecotones are nonlinear in nature and complicate the relations of pattern and process in vegetation, which, combined with observed fractal patterns, suggests a complexity science approach to investigate ecotone dynamics. A cellular automaton of alpine treeline, including nonlinear, local, positive, and negative feedbacks in tree establishment and mortality, as reported in field studies, is used as a model system. Fourier analysis of simulated alpine treelines shows fractal patterns across the treeline landscapes, which are created by patch development. The temporal evolution of the spatial pattern is also fractal. Landscape scale linear correlations between spatial pattern and the rate of advance of trees into tundra arise from localized nonlinear interactions. A tree-patch-landscape scale explanation of pattern–process interaction is proposed in which the endogenous feedbacks determine the spatial and temporal fractal properties of the ecotone. The simulated treeline advance exhibits self-organized complexity and may indicate a potential strategy for monitoring change.     

广州批发市场的销售物流网络空间研究——基于基本—非基本经济活动视角

经济基础理论是解释城市增长的经典理论之一,但实地数据的缺乏使得相关实证研究效果不佳。结合实地调查数据,从基本—非基本经济活动的视角对广州市批发市场的销售物流网络空间进行研究,得出基本经济活动与非基本经济活动的比率高达4.27,反映出城市经济活动外向性的特点;基本经济活动网络总体上形成了以广州为中心,珠三角和省外城市为重要支点的星型流动格局;非基本经济活动网络呈现出城市中部强、南北弱的货流格局。最后,采用R型和Q型因子分析方法,对销售物流网络进行流场识别,深入刻画网络的空间特征。结果显示销售物流网络中形成了一个综合活动场、五个基本活动场以及两个非基本活动场。     

The Functional Resilience of an Innovative Cluster in the Montpellier Urban Area (South of France)

The aim of this article is to evaluate the resilience of territories through a study of the morphology of local firms' networks: in this case, in the Montpellier urban area. Two types of data collection were used for the analysis: a statistical survey of a sample of firms and a list of shared patents pending. The first step of the approach enabled us to analyse the motives that led firms to develop innovative partnership. In the second step, we modelled the networks, using graph theory, over different periods of time. Ultimately, the analysis revealed pivotal firms within the network. These firms played an important role in the resilience and spatial organization of the territory. We found that the geographical concentration or dispersal of the firms thus continued under certain economic conditions.     

Spatial Filtering for Identifying a Shortest Path Around Obstacles

The shortest path between two locations is crucial for location modeling, spatial analysis, and wayfinding in complex environments. When no transportation system or network exists, continuous space movement adds substantial complexity to identifying a best path as there are increased travel options as well as barriers inhibiting potential movement. To derive the shortest path, various methods have been developed. Recent work has attempted to exploit spatial knowledge and geographic information system functionality, representing significant advantages over existing methods. However, a high density of obstacles increases computational complexity making real‐time solution difficult in some situations. This article presents a spatial filtering method to enhance Euclidean shortest path derivation in complex environments. The new approach offers substantial computational improvement while still guaranteeing an optimal path is found. Application results demonstrate the effectiveness of the approach and its comparative superiority.     

Modeling distance with time in ancient Mediterranean seafaring: a GIS application for the interpretation of maritime connectivity

Investigations of Mediterranean connectivity have increasingly turned toward maritime landscape models to frame questions of seaborne exploration, marine resource exploitation, trade and exchange, and seafaring culture. Environmental and technological parameters are consistently acknowledged as crucial for understanding when and why different relationships developed across the sea, but their formal employment in the modeling and interpretation of maritime space remains quite limited. The methodology outlined here utilizes Geographic Information Systems (GIS) to integrate environment and technology as analytical tools for exploring the complexity of seaborne connectivity. Focusing on sailing days as practical units of distance and using an Archaic Greek shipwreck off Turkey as a case study, this preliminary model demonstrates how a more nuanced spatial approach can inform the human geography and socioeconomic structures of ancient maritime interaction.     

Multiple pathways across past landscapes: circuit theory as a complementary geospatial method to least cost path for modeling past movement

The incorporation of geospatial technologies in archaeology has resulted in productive advances in the analysis of past behavioral processes. Archaeologists have relied on the Geographic Information Systems (GIS) application of cost surface analysis and the computation of least cost paths (LCP) to study movement, one key social process. Recent research has identified limitations with LCP for modeling nonhuman species movement, notably the inability of LCP to accommodate multiple pathways. Archaeology must consider the implications of these critiques for models of past human movement. In this paper, I apply a different approach, circuit theory modeling, enacted through the program Circuitscape, to an archaeological case study previously analyzed with LCP modeling, travel to a regionally significant ceremonial earthwork center during Late Prehistory (ca. AD 1200–1600) in the Northern Great Lakes. Through this comparative analysis, I evaluate the advantages and disadvantages of these methods for modeling past movement. The results suggest LCP modeling and circuit modeling offer archaeology complementary geospatial methods for conceptualizing past mobility. Combining circuit theory and LCP allows archaeologists to produce richer models of past movement by appreciating scenarios where multiple pathways are important as well as scenarios where optimum single travel routes have priority.     

ART‐P‐MAP Neural Networks Modeling of Land‐Use Change: Accounting for Spatial Heterogeneity and Uncertainty

Spatial land‐use models over large geographic areas and at fine spatial resolutions face the challenges of spatial heterogeneity, model predictability, data quality, and of the ensuing uncertainty. We propose an improved neural network model, ART‐Probability‐Map (ART‐P‐MAP), tailored to address these issues in the context of spatial modeling of land‐use change. First, it adaptively forms its own network structure to account for spatial heterogeneity. Second, it explicitly infers posterior probabilities of land conversion that facilitates the quantification of prediction uncertainty. Extensive calibration under various test settings is conducted on the proposed model to optimize its utility in seeking useful information within a spatially heterogeneous environment. The calibration strategy involves building a bagging ensemble for training and stratified sampling with varying category proportions for experimentation. Through a temporal validation approach, we examine models’ performance within a systematic assessment framework consisting of global metrics and cell‐level uncertainty measurement. Compared with two baselines, ART‐P‐MAP achieves consistently good and stable performance across experiments and exhibits superior capability to handle the spatial heterogeneity and uncertainty involved in the land‐use change problem. Finally, we conclude that, as a general probabilistic regression model, ART‐P‐MAP is applicable to a broad range of land‐use change modeling approaches, which deserves future research.     

Nonlinear Multivariate Spatial Modeling Using NLPCA and Pair‐Copulas

A novel geostatistical modeling approach is developed to model nonlinear multivariate spatial dependence using nonlinear principal component analysis (NLPCA) and pair‐copulas. In spatial studies, multivariate measurements are frequently collected at each location. The dependence between such measurements can be complex. In this article, a multivariate geostatistical model is developed that can capture both nonlinear spatial dependence across locations and nonlinear dependence between measurements at a particular location. Nonlinear multivariate dependence between spatial variables is removed using NLPCA. Subsequently, a pair‐copula based model is fitted to each transformed variable to model the univariate nonlinear spatial dependencies. NLPCA and pair‐copulas, within the proposed model, are compared with stepwise conditional transformation (SCT) and conventional kriging. The results show that, for the two case studies presented, the proposed model that utilizes NLPCA and pair‐copulas reproduces nonlinear multivariate structures and univariate distributions better than existing methods based on SCT and kriging.     

空间联系视角的沈阳市多中心城市结构研究

以沈阳市为研究案例,借鉴国际著名团队POLYNET的研究经验并加以改良,进行了空间联系视角的多中心城市结构的实证分析。分别从人口密度分布、就业-居住空间联系网络与企业空间组织网络三个维度来识别沈阳市的人口中心、就业中心与产业中心,综合分析这些"中心"得出四点结论:①城市核心区就业与居住中心构成十字交叉的内循环结构;②浑南新区次中心特征颇具雏形;③外围人口中心、就业中心与产业中心错位发展;④沈阳市呈现不均衡的多中心结构特征。     

Polycentricity and the Multiplexity of Urban Networks

Abstract

Empirical studies on polycentric urban regions (PURs) tend to analyse their spatial organization by examining only one type of functional linkage between cities. However, it has generally been accepted that urban networks are multiplex phenomena and that spatial interactions between cities can take many different forms, for example, commuting, shopping trips, and inter-firm trade. The spatial organization of each of these functional linkages is not necessarily identical, and, therefore, a region can appear to be polycentric and spatially integrated based on the analysis of one type of functional linkage but monocentric and loosely connected based on the analysis of another type of functional linkage. The aim of this paper is to stimulate further discussion on the multiplexity of urban networks with regard to the relational complexity of urban regions. Focusing on one PUR (Randstad Holland), we compare the geographical scope and spatial structure of different functional networks within it. Our results indicate that the spatial organization of the urban network depends on the lens through which it is assessed.     

Measuring Spatial Correspondence among Network Paths

Measuring the spatial correspondence among geographic features is an important component of many analyses such as those seeking to identify similar or dissimilar features, explore the extent to which change has taken place, and explain the processes influencing spatial change. In networked systems, the paths providing connectivity between pairs of nodes are often the geographic features among which spatial correspondence is to be assessed. That is, given a set of paths between a pair of nodes, the extent to which they share portions of the network and spatially deviate from one another can provide insight on the factors underlying the use of a system. While methods for measuring the spatial correspondence among specific types of network paths have been devised, this article proposes a methodology for measuring the spatial correspondence among the topology of network paths of any type, including those involving redundant use of network space. By basing comparisons upon the topologic relationships among the paths, the proposed approach better accounts for average spatial deviation as well as the asymmetric spatial relationship between pairs of paths, enabling greater stability and consistency in the analysis of their spatial characteristics. The developed methodology is applied to evaluate variability in spatial correspondence among a set of network paths to demonstrate its utility.     

A Combined Cluster and Interaction Model: The Hierarchical Assignment Problem

This article presents a new spatial modeling approach that deals with interactions between individual geographic entities. The developed model represents a generalization of the transportation problem and the classical assignment problem and is termed the hierarchical assignment problem (HAP). The HAP optimizes the spatial flow pattern between individual origin and destination locations, given that some grouping, or aggregation of individual origins and destinations is permitted to occur. The level of aggregation is user specified, and the aggregation step is endogenous to the model itself. This allows for the direct accounting of aggregation costs in pursuit of optimal problem solutions. The HAP is formulated and solved with several sample data sets using commercial optimization software. Trials illustrate how HAP solutions respond to changes in levels of aggregation, as well as reveal the diverse network designs and allocation schemes obtainable with the HAP. Connections between the HAP and the literature on the p-median problem, cluster analysis, and hub-and-spoke networks are discussed and suggestions for future research are made.     

Reconceptualizing the Policy Subsystem: Integration with Complexity Theory and Social Network Analysis

The concept of the policy subsystem is an essential building block for several of the basic frameworks of policy process studies. Over time issues have become more complex, crossing subsystem boundaries, and so subsystems have escalated in their complexity as well. It is increasingly insufficient to study just one policy subsystem and so scholars have turned to studying boundary‐spanning regimes or policy networks. In this essay, we review the major contributions to developing the concept of a policy subsystem and trace its evolution into broader conceptualizations like issue and policy networks. We argue that the future for theories of the policy process is in more explicit integration of complexity theory and more effective modeling of subsystems with the utilization of social network analysis. In closing, we discuss the enduring nature of the concept of policy subsystems and highlight studies that continue using it in innovative ways.     

SOME OBSERVATIONS ON THE STRUCTURE OF SOCIO-ECONOMIC CELLULAR NETWORKS

The paper has presented an approach to cellular network pattern analysis which is novel in several respects. First, the approach emphasizes the consideration of the complete geometric structure of the pattern rather than the examination of summary statistics for individual variables or correlations between pairs or other combinations of variables. By comparing empirical structures with model-generated ones it is suggested that the approach offers a way of evaluating the relative impact of the three major forces thought to be responsible for the geometric structure of a network, i.e., the generating process, the phenomena expressed in the network, and the function of the network. It was demonstrated how different processes can lead to diverse geometric structures. Examination of two empirical networks composed of bus service centre hinterlands suggested that the generating process was the dominant determinant of physical structure in this particular instance. However, more research on other cellular networks consisting of different phenomena, and on different generative processes, must be undertaken before any firm conclusions can be reached. The approach taken here would seem to offer one way of undertaking such research.     

Measuring the Structure of Road Networks

Spatial networks display both topologic and geometric variations in their structure. This study investigates the measurement of a road network structure. Existing measures of heterogeneity, connectivity, accessibility, and interconnectivity are reviewed and three supplemental measures are proposed, including measures of entropy, connection patterns, and continuity. The proposed measures were applied to 16 test networks, which were derived from four idealized base networks: 90°, 45°, 30°, and completely connected. The results show that the differentiated structures of road networks can be evaluated by the measure of entropy; predefined connection patterns of arterial roads can be identified and quantified by the measures of ringness, webness, beltness, circuitness, and treeness. A measure of continuity evaluates the quality of a network from the perspective of travelers. Proposed measures could be used to describe the structural attributes of complicated road networks quantitatively, to compare different network structures, and to explore the structural evolution of networks in the spatial and temporal context. These measures can find application in urban planning and transportation practice.     

信息时代区域空间结构构成要素分析

进入20世纪90年代以来,信息技术对区域与城市社会、经济及文化系统影响迅速加深,从而使区域与城市空间结构正经历着巨大的变革。本文首先回顾了经典的空间结构构成要素分析,然后对信息技术影响下的空间结构的点、线、面在内涵及形态上的相应变化进行了解释。对于点要素而言,其规模、功能、地位及不同点之间的联系都发生了变化。而线要素的变化主要表现在信息技术促使了信息流对空间结构作用的强化,不同网络之间的互动日益重要。这些又导致面要素出现了如智能区域等新的空间组合模式。最后,指出了空间结构的成长与变化是与特定社会经济环境相联系的,需要针对特定地区进行具体分析。     

国外创新网络研究述评与区域共生创新战略

创新网络是企业创新在地理空间和战略关系层面上的综合体现。在经济地理视角下,本文重点从企业层面的联盟创新网络,产业层面的集群创新网络,以及区域层面的共生创新网络加以具体评述与分析。在企业层面,强调企业在网络当中的核心地位,以及如何通过企业战略联盟构建创新网络;在产业层面,强调产业集群要素对创新网络影响,更倾向刻画创新网络是一种以产业地理空间为边界的整体性创新网络;在区域层面,跨越企业所处的产业边界,基于生态共生思想提出区域共生创新网络的新范式。最后,提出创新网络未来的研究应重点关注企业、产业与区域三个层面的协同演化,突破传统区域创新网络的空间特征和地理格局,探讨如何构架一个具有层次性、协同性和互动开放性的区域生态共生创新网络。