Reducing police response times: Optimization and simulation of everyday police patrol

Police forces around the world are adapting to optimise their current practices through intelligence-led and evidence-based policing. This trend towards increasingly data-driven policing also affects daily police routines. Police patrol is a complex routing problem because of the combination of reactive and proactive tasks. Moreover, a trade-off exists between these two patrol tasks. In this paper, a police patrol algorithm that combines both policing strategies into one strategy and is applicable to everyday policing is developed. To this end, a discrete event simulation model is built that compares a p-median redeployment strategy with several benchmark strategies, i.e., p-median deployment, hotspot (re)deployment, and random redeployment. This p-median redeployment strategy considers the continuous alternation of idle and non-idle vehicles. The mean response time was lowest for the p-median deployment strategy, but the redeployment strategy results in better coverage of the area and low mean response times.