Room G.13, East Building
School of Geography
University of Leeds
Leeds LS2 9JT UK
+44 (0) 113 34 36759
Agent-Based Modelling of Crime
Start date: October 2012
Research into Crime and Crime Prevention is often based upon aggregated levels of human behaviour and offending patterns. However, such a level of focus will often overlook the individual-level behaviour of offenders that contribute to wider patterns of crime. Consequently, any subsequent Crime Reduction Strategies that result from such work are likely to be based predominantly on aggregated patterns of crime. This problem can be addressed using a relatively novel computer simulation technique known as 'Agent-Based Modelling', which allows for the modelling of behaviour at the individual level. Within the context of crime (and specifically, burglary), Agent-Based Modelling enables users to explore how the individual-level behaviour of burglars contributes to patterns at a wider, more aggregated level, whilst determining the extent to which wider patterns of burglary reflect the nature and extent of offending at an individual level.
The current project centres upon the development of an Agent-Based Model within the context of the residential burglary system in Leeds. The model will integrate real-life crime data and realistic offender behavioural profiles, as determined through interviews with offenders. Such 'real-world' data will help to establish the decision-making processes taken by offenders, enabling a more detailed exploration of burglars' motivations to offend. In addition, there remains scope to explore the behaviours and movement patterns of potential victims within the model. Information used within the model will help to develop different 'typologies' of offenders and victims based on their characteristics and behaviours, therefore contributing towards a greater understanding of the dynamic between these two groups at both a macro and micro-level. Outputs from the model may help to contribute towards future crime prevention/ reduction strategies, as well as awareness/ safety campaigns for local residents, whilst helping assess the effectiveness of existing schemes.
Through the development of this model, the research has the following main objectives;
- To gain a greater understanding of different offender (burglar) types and the factors that influence the decision-making process across each type of offender.
- To develop understanding of the daily/ seasonal commuting and movement patterns of potential victims, and the dynamics between different types of offenders and victims.
- To gain a greater understanding of crime patterns within the context of burglary and relating this to criminological theory. This will involve developing understanding into how individual-level behaviour contributes to wider patterns of crime, whilst determining the extent to which larger crime patterns reflect behaviour at an individual-level.
- To assess the impact and effectiveness of existing crime reduction policies/ strategies.
‘Agent-Based Modelling of Crime: Modelling the Residential Burglary System’. Paper presented at the Third Summer School of the European Social Simulation Association (ESSA), University of Toulouse, Toulouse, France, 2nd – 6th July 2012.
‘Exploring the impact and effectiveness of the ‘Project Optimal’ burglary reduction initiative in Leeds: a spatio-temporal approach’. Paper presented at the International Crime and Intelligence Analysis Conference, Manchester Central Convention Centre, Manchester, UK, 13th – 14th December 2012.