Agent-Based Modelling

This training session is open to all ESRC and non-ESRC funded PhD and MA Social Research students within the WRDTP’s seven partner universities. PGRs from all seven interdisciplinary Pathways are welcome to attend.

Agent-based models (ABMs) have become a popular method for simulating complex social systems in a range of fields. The technique allows researchers to create artificial environments inhabited with virtual populations of heterogenous, autonomous decision makers – ‘agents’. The agents are able to perceive their surroundings, reason, and act according to rules that are created on the basis of empirical observations and theoretical knowledge.

Unlike many other quantitative methods that can be executed relatively easily using well-known software or programming libraries, an ABM is often unique to a particular question or study. This means that most ABMs are bespoke and must be developed from scratch prior to their application to address a particular problem. This sets a significant barrier to entry.

This course will introduce the technique of agent-based modelling, discussing its foundations as well as a range of advantages and disadvantages associated with its application in the social sciences. It will then work through a number of practical modelling examples, gradually increasing in difficulty and complexity. Throughout we will use the software platform NetLogo, a freely available and commonly-used package for the development of agent-based models that is designed to be easy-to-use even for people with no prior programming experience.

Please note that attendees will need to bring their own laptops and ensure they have NetLogo (free) – see https://ccl.northwestern.edu/netlogo/

Nick Malleson
Professor of Spatial Science, School of Geography, University of Leeds
I am a Professor of Spatial Science at the Institute for Spatial Data Science (ISDS) at the School of Geography, University of Leeds, UK. My research leverages techniques developed in computer science, statistics and machine learning, and applies them to critical social problems that have a strong geographical context. I am best known for my work on agent-based modelling (an individual-level simulation approach) and in the development of machine-learning and geographical information science techniques to solve problems in the domains of criminal justice, mobility and health.
Dan Birks
Associate Professor of Quantitative Policing & Crime Data Analytics, School of Law, University of Leeds
I am a computational social scientist primarily focusing on urban analytics and the role computational methods can play in better understanding, predicting and disrupting crime problems and improving well-being. I hold degrees in Artificial intelligence and Computer science, Cognitive science, and Criminology, and have almost 20 years’ experience working closely with criminal justice practitioners and policy makers in the UK and Australia to deliver high impact research outcomes.