A hybrid approach for the dynamic routing of police patrols

Efficient allocation of police resources is critical to public safety, particularly in urban environments with limited resources and high crime rates. This study proposes a dynamic approach to police patrolling that integrates predictive analytics, Ant Colony Optimization (ACO), and discrete-event simulation to optimize hotspot coverage and incident response times. The method combines historical Calls for service data and urban structures to predict high-risk areas (hotspots) and guide patrol deployment. The model dynamically adjusts patrol routes based on real-time hotspot conditions and incident priorities, employing Bayesian decision-making and a dynamic pheromone mechanism. A case study in Porto Alegre, Brazil, demonstrates the effectiveness of the proposed method in reducing idle time at crime hotspots and improving incident response times across priority levels. The results highlight the potential of data-driven policing to enhance operational efficiency and balance resource allocation and public safety.