Urban Simulation: Methods, Models and Planning Applications

A Series of Nine Lectures with Workshops presented by Michael Batty

School of Geography at Ritsumeikan University, Kyoto, Japan

October 12th – October 15th, 2010

This course presents the basic elements of urban simulation, that is, methods and models which embody the science of cities as predictive computer models. The models included cover a wide range of types, each of which represents different aspects of the city systems that are codified in the form of simulation models. The first models developed were Land Use Transportation Models (LUTM), sometimes called Land Use Transportation Interaction (LUTI) models and these were first applied in North America from the late 1950s onwards. Further development and applications in other parts of the western world significantly in Britain and Australia began in the 1960s and 1970s. The models essentially represented the city as a set of sectors dealing with employment, population, education and so on and tied these sectors together through various kinds of flows and interactions which were conceived in terms of travel demand. They invariably define the city as a set of small zones, at a cross section in time in which activities such as population and employment are attributes of each zones and the flows or travel between zones represent the ways in which these various sectors and activities are connected.

Modellers gradually began to grapple with problems of making these models more real by disaggregating their spatial representation to ever finer zones and their demographic and economic sectors in more sectoral detail. They were also expanded to consider explicit supply side aspects of the urban economy, simulating rudimentary markets, while they also have been extended to embrace more formal temporal dynamics than the original types of such model which assumed the city to be in equilibrium. In this sense, they were/are referred to as comparative static models. These lectures will begin mainly with land use transportation models. However in parallel, other models of city systems came to be developed, particularly those that pick up on the explicit dynamics of urban change. These models are based much more on representing the system using finer spatial cells closer to representing real development than the land use and traffic zones that are intrinsic to LUTI models. They also pick up on the representation of physical layers of cities using GIS and they have been extended into representing not only fine spatial layers of cells but also individuals or agents – usually populations of individuals – who act to change the development in these cells. These cellular automata (CA) and agent based models (ABM) as they are called, have barely been linked to mainstream LUTI models as yet. In the last three lectures in this series, we will sketch their rudiments.

We thus define Urban Simulation as subsuming Land Use Transportation, Cellular Automata and Agent-Based Models, but these also extend to related modelling approaches such as systems dynamics and micro simulation which we will also note but will not cover.

In nine lectures and the related workshops, we will cover as many of these model types as we can. We will also introduce in the first lectures ideas about what a model is and how it can be seen as essential to the way we develop a science of cities that is useful to our understanding of the city and how it can be planned.

The lectures that are planned are titled as follows. After they have been presented, you will be able to click on each one and download the relevant pdf which it taken from the PowerPoint that contains the relevant material.

The first pdf you can look at is a short summary of the course material as presented at the start of the course - it repeats the salient material about the course outline and the list of lectures.

The list of lectures follows: note that some of these may run into each other and but there will be 9 distinct slots.

Lecture 1: What are Models: The Scientific Context: Definitions of Model and Theory: The Model- Building Process, Data Analysis to Calibration to Prediction

Lecture 2: Modelling Histories: Types and Styles: Urban Models defined, The Urban Modelling Timeline, What Kind of Cities, Examples of Three Model Types

Lecture 3: Basic Theories of Space, Social Physics and the Urban Economy: The Role of Distance in London: Examples: Von Thunen, Population Density, Gravitation

Lecture 4: Land Use Transportation Models: Gravitation and Spatial Interaction, Derivation Methods

Lecture 5: Integrated Urban Models: Some of the previous Lecture 4 is repeated with better visuals

Lecture 6: Integrated Assessment: The Tyndall Models for Greater London: Topical Issues of Climate Change and Energy Predictions in Models

Lecture 7: Cellular Automata Modelling: Principles of Cell Space Simulation

Lecture 8: Modelling Urban Morphology: Fractal Geometry, Relations to CA, and Urban Form

Lecture 9: Agent-Based Urban Models: Individual Based Models: Ideas about Dynamics and Movement: Pedestrian Models

There two additional lectures which support the course – the first is a lecture on new methods of visualisation that I will give to the Colloquium on the first day 12th October at 18-00 hours; the second is part of the Annual Japan GIS meeting on 23rd October. This second lecture is quite similar to the first but has more material on web scraping. These lectures are on:

Colloquium Lecture: Visualising Geographies: Map Mashups, GIS, and Multimedia in a Web 2.0 World

Conference Lecture: Representation and Simulation in GIS: Emerging Trends, New Methods

Background Reading

In this web page, as well as access to pdfs of the lectures themselves, there are some pdfs of the key readings that underpin the course

There is a general historical summary of the development of urban modelling by myself which is entitled:

Batty, M. (2008) Fifty Years of Urban Modelling: Macro Statics to Micro Dynamics, in S. Albeverio, D. Andrey, P. Giordano, and A. Vancheri (Editors) The Dynamics of Complex Urban Systems: An Interdisciplinary Approach, Physica-Verlag, Heidelberg, DE, 1-20

The best reviews of the Land Use Transportation class of Urban Models is by Wegener and this is available in several publications, one of which is

Wegener, M. (2005) Urban Land Use Transportation Models, in D. J. Maguire, M. F. Goodchild, and M. Batty (Editors) GIS, Spatial Analysis, and Modeling, ESRI Press, Redlands, CA, 203-220

I am afraid I do not have pdfs of these chapters but the books from which they come are quite widely available particularly the ESRI book. In the first two lectures, there are five key references to basic ideas about models and these are as follows. You can download them by clicking on them in the order below. The second reference isn't scanned either but I expect there are various people around here who have the Kemp book

1. Batty, M. (2009) Urban Modeling, in R. Kitchin and N. Thrift (Eds) International Encyclopaedia of Human Geography, Volume 12, Elsevier, Oxford, 51–58.

2. Batty, M. (2008) Spatial Interaction, in K. K. Kemp (Editor) Encyclopaedia of Geographic Information Science, Sage. Los Angeles, CA, 416-418 (not yet scanned)

3. Batty, M. and Torrens, P. (2005) Modelling and Prediction in a Complex World, Futures, 37 (7), 745-766.

4. Lowry, I. S. (1965) A Short Course in Model Design, Journal of the American Institute of Planners, 31, 158-165.

5. van der Leeuw, S. E. (2004) Why Model? Cybernetics and Systems: An International Journal, 35, 117-128.

An essential early reference is Ira S. Lowry’s Model of Metropolis that that was his report to the RAND Corporation when he worked on the model for Pittsburgh in the early 1960s. This was at the beginning of developments in urban modelling and basically was the first example of a model that coupled sectors together using the same submodel. You can download the report from this site as it is freely available from RAND on the web.

And if you want some more old background on these models, you can download my book Urban Modelling: Algorithms, Calibrations, Predictions (1976) from our web site by clicking here left. If you want a softcopy of the original, it has been reissued by CUP and you can get it from Amazon.com by clicking here right.

One Key Reference for Each Lecture

The background reading pertains to the first few lectures. I have identified one key reference for each of the lecture blocks and these are listed below

For Lectures 3-5 which deal with generic principles of Spatial Interaction, John Roy’s (2004) book Spatial Interaction Modelling: A Regional Science Context (Springer, Berlin) is a good summary. Get it from the library if your library has it. Here is the Amazon.com entry

For Lectures 4-5 on Land Use Transportation Models and Integrated Models, then the key handbook Button, K. J., Haynes, K. E., Stopher, P., and Hensher, D. A. (Editors) (2004) Handbook of Transport Geography and Spatial Systems, Volume 5 (Handbooks in Transport), Elsevier Science, New York, has several good article on land use transport models - see those by Horowitz, Echenique and Miller amongst others. Also if you log onto this link, it will take you to Google Books which has a copy of the entire handbook

For Lectures 5 and 6 on Integrated Urban Models and Integrated Assessment, the handbook edited by Button et al. which we referenced in Lecture 3 above has a good paper in it by Miller, but the two other key references are Iacono, M., Levinson, D., and El-Geneidy, A. (2008) Models of Transportation and Land Use Change: A Guide to the Territory, Journal of Planning Literature 22, 323-340, and Hunt, J. D. , Kriger, D. S. and Miller, E. J.(2005) Current Operational Urban Land-Use-Transport Modelling Frameworks: A Review, Transport Reviews, 25, 3, 329-376. Click the titles below to get access the copies of the articles

For Lecture 7 on Cellular Automata Models there is a relatively simple and I hope intelligible article I wrote in 1997 in the Journal of the American Planning Association: Batty, M. (1997) Cellular Automata and Urban Form: A Primer, Journal of the American Planning Association, 63, 266-274. But there is a lot to read on CA. My book Cities and Complexity (MIT Press, Cambridge, MA, 2005) has a lot of general material and covers the DUEM model. A more focussed book is Liu, Y. (2008) Modelling Urban Development with Geographical Information Systems and Cellular Automata (CRC Press, Boca Raton, FL). Click on the icons below to get to the article online and info on the books from Amazon.

For Lecture 8 on Modelling Urban Morphology: Fractal Geometry, Relations to CA, and Urban Form , have a look at our book Fractal Cities (Academic Press, 1994) which you can download from our web site. Click the link below. In the edited book, GIS, Spatial Analysis, and Modeling(ESRI Press, Redlands, CA,co-edited with myself, D. J. Maguire and M. F. Goodchild), there are several good articles on various aspects of ABM, CA and LUTI models which are relevant to many of the topics introduced in these six lectures. ENJOY.

For lecture 9 on Agent-Based Models, my Cities and Complexity book suffices. and I would have liked to present a lecture on planning support systems but I will not have time so I have put up my article on PSS from Brail’s recent edited book Planning Support Systems: Progress, Predictions, and Speculations on the Shape of Things to Come, in R. K. Brail (Editor) Planning Support Systems for Cities and Regions, Lincoln Institute of Land Policy, Cambridge, MA, 3-30, which you can get by clicking on the link below

The Assignments

We have four assignments which we will demo and explain in the workshop sessions. These are the basis for the coursework and it is assumed you will choose one of these topics and work it up for data on appropriate examples in Japan.

1 Population Density: Working out, plotting, examining and commenting on population density variation with respect to distance in a large city using a small number of zones (~10 or so): We have a spreadsheet that computes population densities for 33 London boroughs, provides a list of centroid coordinate data for each borough, computes the airline distance from the City to all boroughs and then develops a negative exponential population density model for this data, plotting it and examining its performance.

2 Population size distributions: this program in a spreadsheet examine the rank size frequency distribution of population and employment counts and densities using he London borough data and fits an inverse power law to the data. This is a model dealing with the competition between different zones of different size based on the hypothesis that there are fewer zones of large size than zones of small size. It is the basis of the rank size rule.

3 Residential or retail location modelling: we provide a spreadsheet similar to that for population density but this time, for working out all airline distances between the 33 London boroughs and then working out a simple singly constrained gravity model with a given parameter value to predict trips between employment origins and residential destinations. The program examines the performance of this model. It also contains the observed matrix of work trips between all origins and destinations so you can check how good the model is.

The three spreadsheets for these three examples can be downloaded below:

4 Exploring Agent-Based Models: using one model from the Netlogo suite of models and commenting on its structure with respect to how the model works; and then making a model proposal using ABM. You are asked to download the software from the web site below, then load the Models Library, select the Social Science subdirectory, and load the Segregation model and then explore this. This is Schelling’s model that you should find out about and there is a description in my Cities and Complexity book. Then you need to consider an agent based model proposal which relates to the course material and you need to write a proposal for how you might build such a model using Netlogo; if you want to you can implement it but in 24 hours this will be hard although not impossible

Last Updated 15 October 2010, by Mike Batty