Chapter 1

Agent-based Modelling and Geographical

Information Systems
Learning Objectives
• Introduce the textbook, structure and online resources.
• Provide contextual background and general overview of the major
developments in geographical modelling for the simulation of the
individual.
• What is a model?
• Introduce complexity theory.
• Set the scene for the upcoming chapters:
• Provide a brief overview of the main topics covered in each of the chapters.
Introduction
• Geographers have always been interested in the role and influence of
the individual within geographical systems.
• Specifically what are the consequences of individual behaviors and decision-
making over space and time?
• In the last 20 years we have seen an explosion in computer processing
power and storage and micro-level data sets.
• Coinciding with this, ideas from complexity science now influence our
thinking geographical systems.
• How individuals interact and shape geographical systems from the bottom up.
Complexity and Geographical Systems
• Complexity arises when a small number of rules or laws, applied at a
local level and among many entities, are capable of generating
complex global phenomena:
• Collective behaviors, extensive spatial patterns, hierarchies, etc.
Hierarchies and Interactions

A simple hierarchical structure of a city composed of multiple neighborhoods which form a hierarchy at the
more macro level but also have interactions (e.g. commuter flows) amongst each other.
Complexity and Geographical Systems
Term Explanation Example(s)
In economics, national and global markets evolve
The system’s ability to self-organise –
Self-organisation from locally interacting agents all pursuing what
without higher-level direction.
they want.
Increasing education and employment
Outputs do not have to be
Non-linear opportunities for the people does not stop rioting
proportional to their inputs.
(Appendix A13).
Self-organisation results from Negative: thermostat that attempts to regulate
feedback mechanisms (positive & the temperature in a room.
Feedback
negative) as a result of interactions Positive: a run on a bank which can cause it and
between individual entities. other banks to fail.
How history dictates how systems Rank size distribution of cities or residential
Path Dependence
evolve and restructure. decisions impact land markets.

Key terms used in complexity theory

Models
• In seeking to understand the complexity of the world around us, we
require methods that can capture and regenerate the characteristics
of complex systems.
• We do this through modelling.
• A model is a simplified representation of a complex reality, to the point that
the reality is understandable and analytically manageable.
• A model can be constructed as a computer program that uses a
simplified digital representation of one or more aspects of the real
world, transforming them to create a new representation.
• There are many reasons and purposes for modelling, from description
and explanation to prediction.
Data
• While models provide one route towards understanding geographical
complexity, we are also rapidly gaining a richer understanding of the world
through the vast amount of new data becoming available and enabling new
veins of research.
• By the year 2020, many predict the global universe of accessible data to be of the
order of 44 zettabytes or 44 trillion gigabytes
• Data now not only comes from government and commercial organizations (e.g. US
Census Bureau) but also from individual users.
• However, data alone does not give us the ‘why’, only the ‘what’, in the
sense that it gives us the patterns but not the processes that cause them to
emerge. For this, we need to model social interaction and the individuals
involved in such interactions.
Individuals
• There has been a marked change over the past sixty years in how we simulate the
individual within geographical systems.
• Previous generations of researchers did not have either the computational power or
access to data that today’s researchers do.
• Lack of data and tools for simulating individuals meant that models examined the
systems as a whole, based on aggregate analysis offering a limited way to examine the
dynamics taking place within such systems
• The increase in computation, individual-level data sources and the infusion of ideas from
complexity theory allow us to look at how geographical systems emerge from the
“bottom up”
• E.g. , it is the local-scale interactive behavior (commuting, moving) of many individual objects
(vehicles, people) from which structured and ordered patterns emerge, such as peak-hour traffic
congestion and the large-scale spatial clustering of socio-economic groups by residence.
• Understanding how geographical systems grow and evolve from the bottom up allows us
to understand how uncoordinated local decision-making gives rise to coordinated global
patterns which define the shape of the world around us and how these global patterns
then influence the local decisions of individuals.
Agent-Based Modelling and Geographical
Information Systems
• While data can take us some of the way towards understanding the
world around us, only models allow us to build a comprehensive
understanding of the mechanisms and behaviors that shape complex
geographical phenomena.
• In the remaining chapters we will explore how we can model
geographical systems from the “bottom up” using agent-based
models
• The overarching aim of the book is to facilitate the in-depth learning
of both the fundamental knowledge and practical skills required for
building and running agent-based models for simulating geographical
systems.
Book Outline
 Chapter 2:
Introduction
to Agent-
based
Modelling

• Introduce key concepts behind agent-based modeling

(ABM):
• What is an agent?
• Advantages and limitations of ABM
• Applications of ABM:
• E.g. Residential movement, transportation
• How ABM has been used in decision making?
 • This chapter will introduce the Preparing to
Model
core concepts and frameworks Design and
that can be used to plan, Preparation

implement and disseminate Model Design

geographical agent-based
Chapter 3: models. Model Building the
Designing and • What are the questions that
Implementation Model

Developing an social scientists and

Agent-based geographers need to consider Verification
Model when designing and building an
agent-based model?
• What design frameworks and Calibration Model
software toolkits are available Evaluation

to use? What are their relative

pros and cons? Validation

• What methods are available for

documenting design concepts
and why are they useful to Explanation Prediction

modelers?
Running the Model and Understanding Insights
 Chapter 4:
Building
Agent-Based
Models with
NetLogo

• Provides an overview of the programming language and

concepts that are used within NetLogo (Wilensky, 1999):
• How to create a simple environment.
• Step-by-step instructions for creating a simple model.
• Subsequent chapters build upon the basics presented
here.
 Chapter 5:
Fundamentals
of
Geographical
Information
Systems
• Presents the main concepts and terminology that are required
to understand geographical information systems. E.g.:
• Data types (i.e. raster & vector).
• Accuracy and precision.
• Provide a brief overview of the development of GIS.
• Using QGIS, we demonstrate how to prepare and manipulate
example GIS data and how to integrate it into an agent-based
models.
 Chapter 6:
Integrating
Agent-Based
Modelling and
GIS

• Chapter explores the benefits to linking GIS and Agent-

based models. Specifically:
• Loose and tight coupling approaches:
• Critiquing the relative advantages and disadvantages of both.
• An overview of open source toolkits (e.g. MASON, NetLogo,
GAMA) that can be used for the creation of geographically
explicit agent-based models.
• Examples of how geographical (raster & vector) data and agent-
based models can be combined.
Chapter 7:
Modelling
Human
Behavior

• Chapter explores the most common approaches by which

researchers incorporate human behavior into agent-based
models.
• E.g. mathematical and conceptual cognitive models
• Provides examples of how these approaches can be
implemented:
• Consumer behavior using probabilistic rules
• Simulating behavior in riots using a cognitive model
Chapter 8:
Networks

• Networks play a critical role in our lives in terms of physical networks

we use to navigate upon, our social networks and more recently how
we communicate via cyber networks (e.g. social media).
• This chapter provides a brief introduction to such networks and
shows how they can be integrated into agent-based models.
• Models are introduced that demonstrate how to navigate agents
along a physical road network (a common requirement for spatially-
explicit agent-based models) and how social networks evolve over
time.
 Description of Point
Visual Comparisons
Data
Key

Choropleth Nearest Point data

KDE Point maps Ripley's K
maps Neighbour Point data

Aggregate
data

Local Indicators of
Chapter 9: Global Statistics Spatial Association
(LISA)
Spatial
Statistics
RSS 2 (S)RMSE Dual KDE GI*
R

• This chapter presents a range of statistics and

algorithms that can be used to compare two spatial
data sets.
• These are important for modelling because, at some
point, it will be necessary to compare a model
outcome to some real-world data in order to assess
how reliable the model is.
 Simulated Observed
Data Data

Generate
Data

Compare Data and Error

Start Run Model Calculate Error sufficiently Yes Stop
(aka fitness) low?
Chapter 10:
No
Evaluating
Models: Adjust model
parameters
Verification,
Calibration,
Validation • Model evaluation is one of the central challenges associated with
agent-based models.
• A key question that all modelers face is “how well does this model
simulate the phenomenon of interest?”.
• Researchers often adopt the same three stage process of
verification, calibration and validation.
• This chapter presents an overview of the methods that are
commonly used within each of these stages.
• The overarching aim of this chapter is to provide the reader with the
knowledge to design their own approach to evaluating agent-based
models.
Chapter 11:
Alternative
Modelling
Approaches • ABM is one of the most popular approaches used in social and
spatial simulation.
• However, there are several other alternative approaches that are
available to the researcher including Cellular Automata,
Microsimulation, Discreet Event Simulation, System Dynamics and
Spatial Interaction models.
• This chapter presents an overview of these other approaches giving
simple examples on how they can be used and summarizing the
main differences between them.
• To compare these models, they are applied to the same issue, the
spread of a disease using a Susceptible-Infected-Recovered (SIR)
epidemic model.
• This shows that while the same general patterns emerge, the reasons for
this are very different.
 Chapter 12:
Summary and
Outlook
• This chapter reflects on the current state of the art of agent-
based models and factors that may shape the future of this
discipline.
• Specifically the key challenges for developing robust agent-
based models of geographical systems are discussed as well as
potential solutions.
• E.g. , the relationship between theory and models, inter-model
comparison, replication, verification and validation, behavior, sharing
and dissemination, along with data challenges.
• We need to address these issues if these models are to be
used to offer insight into key societal challenges,
• E.g., climate change, urban growth and migration.
 Online Resources

• Visit:
https://www.abmgis.org/
for models, tutorials and
data used in this book.
• Each chapter has a specific
page with more
information. See:
https://github.com/abmgis/
abmgis