An Optimal Police Patrol Planning Strategy for Smart City Safety

The safety of citizens is an integral part of any smart city project. Police patrol provides an effective way to detect suspects and possible crimes. However, policing is a limited resource just like any other service that a Smart City provides. In order to efficiently consume this resource, the city has several aspects that can be controlled to make efficient use of Police patrolling: where (area), what (number), when (hour). In this paper, we utilize the LA County Sheriff's open crime dataset to study the police patrol planning problem. We propose a novel approach to build a network of clusters to efficiently assign patrols based on informational entropy. This minimizes Police time-to-arrival and lowers the overall numbers of police on patrol. Our algorithm relies upon the categories of crimes, and the locations of crimes. Since we use real-time traffic analysis to join crime clusters, our solution is extensible enough to be applied to any metropolitan area.