This project will investigate how data-informed agent-based models (ABM) might be used to increase understanding of police resource and demand dynamics. In this initial project phase, the research team (containing expertise in areas from agent-based modelling to policing practice) will undertake a range of activities including; consultation with police stakeholders, an audit of extant information about policing demand, identification of relevant data sources, and the development of proof-of-concept computational models.
Explaining the science
Understanding demand is a non-trivial task. An array of factors, internal and external to police organisations, are relevant. These factors are often highly interdependent (meaning that all choices have opportunity costs) and difficult to model using traditional analytical techniques.
Agent-based modelling is a powerful AI technique that enables researchers to build simulated versions of real-world systems. By modelling individual system entities (like people or organisations), the decisions they make, and the interactions that occur between them, ABM allow researchers to explore why certain emergent system outcomes might come about and how systems might be manipulated to produce a desired effect or remove an undesirable one.
To illustrate, there may be a pivotal point at which increasing policing personnel at a serious event results in diminishing returns or problems elsewhere in the supply system. Simulations also allow the hypothetical testing of different scenarios without the need for real world experimentation. Thus, they can explore ‘what if’ questions such as "what happens if I move 100 officers from proactive patrol to response?" or "what happens if I train another 50 officers to respond to public order events?"
The 21st century has seen increasingly diverse and novel responsibilities imposed upon the police service, while funding has declined. Limited policing resource has proved to be an enduring reality. For example, central government funding of the police in England and Wales fell by some 14% between 2010-11 and 2014-15. The need to do more with less makes an understanding of short-, medium- and long-term demand a key priority. Without an increased understanding of the factors that drive demand, optimising police responses will continue to be a challenge.
This project aims to assist by mapping the demand challenges and exploring the viability of technical solutions in the form of simulation models that could support evidence-based ethical decision making. There are many aspects to consider in this enterprise. Problems that the police address vary in terms of scale, harm, and whether responses are dealt with locally, regionally, or require national coordination. For example, some priority areas for which demand needs to be better understood (and met) include online crime, serious and organised crime, and high harm crimes against the most vulnerable. The current policing mantra of addressing ‘risk, threat and harm’ covers some of the issues, albeit imprecisely.
Furthermore, events such as the 2011 UK riots remind us of the need to plan for major incidents of public disorder or attacks on critical infrastructure that require a national response. Useful attempts by the police service to understand how to ‘manage’ demand across the board have been made, but no previous systematic and integrated quantitative efforts to model drivers of, and response to, demand has taken place. This project aims to do so.
The key aspiration is to assess the viability of models simulating policing demand, resource allocation, and the decision-making which connects them. The models must be flexible enough to simulate resourcing dynamics at both individual and organisational scales. A data-informed agent-based modelling approach will be taken, that begins with several example scenarios. Real world police data relevant to these scenarios will be required and utilised in ways which protect the anonymity of individuals.
Illustrating the national importance of these issues, the recent Home Affairs Committee report concluded that the police service is struggling to cope with recent rises in crime and the advent of new crime types. Resources available are deemed not to match current demand, still less anticipated demand. The authors of that report strongly recommend that police funding is prioritised in subsequent Comprehensive Spending Reviews.
However, the challenges to policing are not limited to there simply being too few resources. As noted in the Association of Police and Crime Commissioners and National Police Chiefs Council’s Policing Vision 2025, “Most forces do not have a thorough evidence-based understanding of demand, which makes it difficult for them to transform services intelligently and demonstrate they are achieving value for money”. The project directly aims to support police progress in this area, undertaking consultation with the police and other key stakeholders throughout the project.
Team activities that offer synergies with this research include: