Geosimulation Automata BasedModelingOfUrbanPhenomena TV pdf Geosimulation # 2004 John Wiley & Sons, Ltd ISBN 0 470 84349 7 Geosimulation Automata based Modeling of Urban Phenomena I Benenson and P M T[.]
Trang 2# 2004 John Wiley & Sons, Ltd ISBN: 0-470-84349-7
Geosimulation : Automata-based Modeling of Urban Phenomena I Benenson and P M To r r e n s
Trang 3Geosimulation Automata-based Modeling of Urban Phenomena
Trang 4Copyright # 2004 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester,
West Sussex PO19 8SQ, England Telephone (+44) 1243 779777 Email (for orders and customer service enquiries): cs-books@wiley.co.uk
Visit our Home Page on www.wileyeurope.com or www.wiley.com
All Rights Reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted
in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission
in writing of the Publisher Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to permreq@wiley.co.uk, or faxed to (+44) 1243 770620.
This publication is designed to provide accurate and authoritative information in regard to the subject matter covered It is sold on the understanding that the Publisher is not engaged in rendering professional services If professional advice or other expert assistance is required, the services of a competent professional should be sought.
Other Wiley Editorial Offices
John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA
Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA
Wiley-VCH Verlag GmbH, Boschstr 12, D-69469 Weinheim, Germany
John Wiley & Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia
John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809
John Wiley & Sons Canada Ltd, 22 Worcester Road, Etobicoke, Ontario, Canada M9W 1L1
Wiley also publishes its books in a variety of electronic formats Some of the content that appears in
print may not be available in electronic books.
Library of Congress Cataloguing-in-Publication Data
Benenson, Itzhak.
Geosimulation : automata-based modeling of urban phenomena / Itzhak Benenson, Paul M Torrens.
p cm.
Includes bibliographical references (p.).
ISBN 0-470-84349-7 (cloth : alk paper)
1 Urban geography – Simulation methods 2 Urban geography – Computer simulation.
I Torrens, Paul M II Title
GF125.B46 2004
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN 0-470-84349-7
Typeset in 10/12pt Times by Thomson Press (India) Limited, New Delhi
Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire
This book is printed on acid-free paper responsibly manufactured from sustainable forestry
in which at least two trees are planted for each one used for paper production.
Trang 5For my parents, Maya and Evsey, with love — Itzhak
Bert and Juicy, this is for you, for all thetimes you have rescued me and for makingthe good times so much better — Paul
Trang 6Preface xiii
Acknowledgements xvii
1 Introduction to Urban Geosimulation 1
1.1 A New Wave of Urban Geographic Models is Coming 1
1.2 Defining Urban Geosimulation 2
1.2.1 Geosimulation Reflects the Object Nature of Urban Systems 2
1.2.2 Characteristics of the Geosimulation Model 3
1.2.2.1 Management of Spatial Entities 3
1.2.2.2 Management of Spatial Relationships 3
1.2.2.3 Management of Time 3
1.2.2.4 Direct Modeling 4
1.3 Automata as a Basis for Geosimulation 4
1.3.1 Cellular Automata 5
1.3.2 Multiagent Systems 6
1.3.3 Automata Systems as a Basis for Urban Simulation 8
1.3.3.1 Decentralization 9
1.3.3.2 Specifying Necessary and Only Necessary Details 9
1.3.3.3 Diversity of Characteristics and Behavior 10
1.3.3.4 Form and Function Come Together 10
1.3.3.5 Simplicity and Intuition 10
1.3.4 Geosimulation versus Microsimulation and Artificial Life 11
1.4 High-resolution GIS as a Driving Force of Geosimulation 12
1.4.1 GI Science, Spatial Analysis, and GIS 12
1.4.2 Remote Sensing 12
1.4.3 Infrastructure GIS 13
1.4.4 GIS of Population Census 13
1.4.5 Generating Synthetic Data 16
1.5 The Origins of Support for Geosimulation 16
1.5.1 Developments in Mathematics 17
1.5.2 Developments in Computer Science 17
1.6 Geosimulation of Complex Adaptive Systems 18
1.7 Book Layout 18
2 Formalizing Geosimulation with Geographic Automata Systems (GAS) 21
2.1 Cellular Automata and Multiagent Systems—Unite! 21
2.1.1 The Limitations of CA and MAS for Urban Applications 21
2.1.2 The Need for Truly Geographic Representations in Automata Models 24
Trang 72.2 Geographic Automata Systems (GAS) 25
2.2.1 Definitions of Geographic Automata Systems 25
2.2.1.1 Geographic Automata Types 26
2.2.1.2 Geographic Automata States and State Transition Rules 27
2.2.1.3 Geographic Automata Spatial Referencing and Migration Rules 28
2.2.1.4 Geographic Automata Neighbors and Neighborhood Rules 30
2.2.2 GAS as an Extension of Geographic Information Systems 31
2.2.2.1 GAS as an Extension of the Vector Model 31
2.2.2.2 GAS and Raster Models 31
2.3 GAS as a Tool for Modeling Complex Adaptive Systems 32
2.4 From GAS to Software Environments for Urban Modeling 32
2.4.1 Object-Oriented Programming as a Computational Paradigm for GAS 32
2.4.2 From an Object-Based Paradigm for Geosimulation Software 33
2.4.3 GAS Simulation Environments as Temporally Enabled OODBMS 34
2.4.4 Temporal Dimension of GAS 34
2.5 Object-Based Environment for Urban Simulation (OBEUS)—A Minimal Implementation of GAS 35
2.5.1 Abstract Classes of OBEUS 35
2.5.2 Management of Time 37
2.5.3 Management of Relationships 38
2.5.4 Implementing System Theory Demands 39
2.5.5 Miscellaneous, but Important, Details 40
2.6 Verifying GAS Models 40
2.6.1 Establishing Initial and Boundary Conditions 41
2.6.2 Establishing the Parameters of a Geosimulation Model 42
2.6.3 Testing the Sensitivity of Geosimulation Models 44
2.7 Universality of GAS 44
3 System Theory, Geography, and Urban Modeling 47
3.1 The Basic Notions of System Theory 47
3.1.1 The Basics of System Dynamics 48
3.1.1.1 Differential and Difference Equations as Standard Tools for Presenting System Dynamics 48
3.1.1.2 General Solutions of Linear Differential or Difference Equations 49
3.1.1.3 Equilibrium Solutions of Nonlinear Systems, and Their Stability 51
3.1.1.4 Fast and Slow Processes and Variables 52
3.1.1.5 The Logistic Equation—The Simplest Nonlinear Dynamic System 53
3.1.1.6 Spatial Processes and Diffusion Equations 54
viii Contents
Trang 83.1.2 When a System Becomes a ‘‘Complex’’ System 57
3.1.2.1 How Nonlinearity Works 58
3.1.2.2 How Opennes Works 62
3.2 The 1960s, Geography Meets System Theory 73
3.2.1 Location Theory: Studies of the Equilibrium City 73
3.2.2 Pittsburgh as an Equilibrium Metropolis 74
3.2.3 The Moment Before Dynamic Modeling 77
3.2.4 Models of Innovation Diffusion—The Forerunner of Geosimulation 77
3.3 Stocks and Flows Urban Modeling 79
3.3.1 Forrester’s Model of Urban Dynamics 79
3.3.1.1 Computer Simulation as a Tool for Studying Complex Systems 79
3.3.1.2 Forrester’s Results and the Critique They Attracted 79 3.3.2 Regional Models: the Mainstream of the 1960s and 1970s 81
3.3.2.1 Aggregated Models of Urban Phenomena 82
3.3.2.2 Stocks and Flows Integrated Regional Models 83
3.4 Criticisms of Comprehensive Modeling 87
3.4.1 List of Sins 87
3.4.2 Keep it Simple! 88
3.5 What Next? Geosimulation of Collective Dynamics! 88
3.5.1 Following Trends of General Systems Science 88
3.5.2 Revolution in Urban Data 89
3.5.3 From General System Theory to Geosimulation 90
4 Modeling Urban Land-use with Cellular Automata 91
4.1 Introduction 91
4.2 Cellular Automata as a Framework for Modeling Complex Spatial Systems 93
4.2.1 The Invention of CA 93
4.2.1.1 Formal Definition of CA 93
4.2.1.2 Cellular Automata as a Model of the Computer 95
4.2.1.3 Turing Machine 95
4.2.1.4 Neuron Networks 96
4.2.1.5 Self-reproducing Machines and Computational Universality 97
4.2.1.6 Feedbacks in Neuron Networks and Excitable Media 97 4.2.1.7 Markov Processes and Markov Fields 98
4.2.1.8 Early Investigations of CA 99
4.2.2 CA and Complex Systems Theory 100
4.2.2.1 The Game of Life—A Complex System Governed by Simple Rules 100
4.2.2.2 Patterns of CA Dynamics 101
4.2.3 Variations of Classic CA 105
4.2.3.1 Variations in Grid Geometry and Neighborhood Relationships 105
4.2.3.2 Synchronous and Asynchronous CA 105
Contents ix
Trang 94.3 Urban Cellular Automata 106
4.3.1 Introduction 106
4.3.2 Raster but not Cellular Automata Models 107
4.3.3 The Beginning of Urban Cellular Automata 113
4.3.4 Constrained Cellular Automata 116
4.3.5 Fuzzy Urbanization 121
4.3.6 Urbanization Potential as a Self-existing Characteristic of a Cell 122
4.3.6.1 From Monocentric to Polycentric City Representations 123
4.3.6.2 Real-World Applications of Potential-Based Models 126
4.3.7 Urbanization as a Diffusion Process 131
4.3.7.1 Spatial Ecology of the Population of Urban Cells 132
4.3.7.2 Spread of Urban Spatial Patterns 133
4.3.8 From Fixed Cells to Varying Urban Entities 137
4.3.8.1 Infrastructure Objects as Self-existing Urban Entities 137
4.3.8.2 Changing Urban Partition 138
4.4 From Markov Models to Urban Cellular Automata 140
4.4.1 From Remotely Sensed Images to Markov Models of Land-use Change 142
4.4.2 The Link Between Markov and Cellular Automata Models 144
4.5 Integration of CA and Markov Approaches at a Regional Level 146
4.5.1 Flat Merging of Markov and CA Models 147
4.5.2 Hierarchy of Inter-regional Distribution and CA Allocation 150
4.6 Conclusions 150
5 Modeling Urban Dynamics with Multiagent Systems 153
5.1 Introduction 153
5.2 MAS as a Tool for Modeling Complex Human-driven Systems 154
5.2.1 Agents as ‘‘Intellectual’’ Automata 154
5.2.2 Multiagent Systems as Collections of Bounded Agents 154
5.2.3 Why do we Need Agents in Urban Models? 155
5.3 Interpreting Agency 155
5.4 Urban Agents, Urban Agency, and Multiagent Cities 158
5.4.1 Urban Agents as Entities in Space and Time 158
5.4.2 Cities and Multiagent System Geography 160
5.5 Agent Behavior in Urban Environments 160
5.5.1 Location and Migration Behavior 161
5.5.2 Utility Functions and Choice Heuristics 162
5.5.3 Rational Decision-making and Bounded Rationality 163
5.5.4 Formalization of Bounded Rationality 165
5.5.5 What we do Know About Behavior of Urban Agents—The Example of Households 170
x Contents
Trang 105.5.5.1 Factors that Influence Household Preferences 170
5.5.5.2 Householder Choice Behavior 172
5.5.5.3 Stress-resistance Hypotheses of Household Residential Behavior 172
5.5.5.4 From Householder Choice to Residential Dynamics 173
5.5.5.5 New Data Sources for Agent-Based Residential Models 175
5.6 General Models of Agents’ Collectives in Urban Interpretation 176
5.6.1 Diffusion-limited Aggregation of Developers’ Efforts 177
5.6.2 Percolation of the Developers’ Efforts 178
5.6.3 Intermittency of Local Development 180
5.6.4 Spatiodemographic Processes and Diffusion of Innovation 182
5.7 Abstract MAS Models of Urban Phenomena 184
5.7.1 Adaptive Fixed Agents as Voters or Adopters of Innovation 184 5.7.2 Locally Migrating Social Agents 190
5.7.2.1 Schelling Social Agents 190
5.7.2.2 Random Walkers and Externalization of Agents’ Influence 193
5.7.3 Agents That Utilize the Entire Urban Space 195
5.7.3.1 Residential Segregation in the City 195
5.7.3.2 Adapting Householder Agents 199
5.7.3.3 Patterns of Firms 205
5.7.4 Agents That Never Stop 205
5.7.4.1 Pedestrians on Pavements 208
5.7.4.2 Depopulating Rooms 213
5.7.4.3 Cars on Roads 216
5.7.5 Multi-type MAS—Firms and Customers 220
5.8 Real-world Agent-based Simulations of Urban Phenomena 224
5.8.1 Developers and Their Work in the City 224
5.8.2 Pedestrians Take a Walk 227
5.8.3 Cars in Urban Traffic 230
5.8.4 Citizens Vote for Land-use Change 233
5.8.5 In Search of an Apartment in the City 237
5.9 MAS Models as Planning and Assessment Tools 244
5.10 Conclusions 248
6 Finale: Epistemology of Geosimulation 251
6.1 Universal Questions 251
6.1.1 Social Phenomena are Repeatable 252
6.1.2 We are Interested in Urban Changes During Time Intervals Derived from Those of a Human Lifespan 252
6.1.3 Urban Systems are Unique because They are Driven by Social Forces 253
6.1.4 The Uniqueness of Urban Systems is not Necessarily Exhibited 253
6.1.5 Why do we Hope to Understand Urban Systems? 253
Contents xi
Trang 116.1.6 Tight-coupling between the Urban Theory and
Urban Data 254
6.1.7 Automata versus State Equations 255
6.2 The Future of Geosimulation 255
6.2.1 The Applied Power of Geosimulation 255
6.2.2 The Theoretical Focus of Geosimulation 256
6.2.3 From Modeling of Urban Phenomena to Models of a City: Integration Based on a Hierarchy of Models 256
6.2.4 From Stand-alone Models to Sharing Code and Geosimulation Language 257
Bibliography 259
Index 283
xii Contents
Trang 12Are we witnessing a revolution in urban geography? The answer to that question isalmost certainly that, yes, we are
That is a bold statement to make But, let’s consider the evidence
During the last four decades, a volume of research on topics of urban geography hasbeen conducted—everything from the geography of urban graveyards to the evolution
of world cities and massive Megalopoli Factual data has not always been available tosettle the arguments that discussion has generated either in print or in conversation,but data are, and always will be, in short supply Regardless of how much data wehave, we always thirst for more; it is a hallmark of life in an Information Age Butdata aside, the general tone of discussion and views on urban geography appear to becoming full circle All too often, the general impression is that of discussing the sameold issues, although they are often marketed in new forms This rebranding isundoubtedly important, but we are not so much interested in shifting units, marketingproducts, as we are in uncovering knowledge Put briefly, there is a strong sense thatall the good theoretical stuff has been said before
What might free us from ever-wandering around the same Mo¨bius strip; more data,better data?
Not so long ago, the data excuse was a pretty good one It is not any more Since thelast decade of the twentieth century, an enormous volume of data has becomeavailable to us, directly to our desktops and our libraries These data cover abewildering array of urban phenomena—information on urban infrastructure andpopulations at all levels of spatial and temporal resolutions have been generated andaccumulated We have not utilized most of them yet This is not because they areinaccessible; we simply shy away, for the most part, from getting stuck into thesehuge reservoirs of remotely sensed data and census databases Modern statistical andGIS environments enable combination of qualitative and quantitative methods, oftenfreely, and we are no longer critically constrained in terms of computing power
So, what then; data analysis?
The common sense view works something like this: let us take a theory, fitthe appropriate data, develop and evaluate a clear and lucid understanding of thephenomenon at hand, and then generate forecasts or what-if scenarios By thesemeans we might thin out all-embracing descriptions, perhaps even give birth to novelideas How has that worked out for us? Has it been successful thus far?
We must admit that most urban models do not work well enough when we deal withreal cities In some cases, the theory turns out to be ‘‘too general’’ to be of use in suchexercises; examining closely, we see that phenomenological ‘‘regression’’ betweenpotential factors and observed consequences, but not the theory, is applied In other,not less frequent, situations, application of theory demands so much ‘‘tuning’’ that the