Data-Driven Optimal Police Patrol Zone Districting and Staffing

点击翻译按钮获取中文摘要

This award contributes to the nation’s security and welfare by developing data-driven decision tools designed to improve police and first-responder operations. The project simultaneously addresses improved patrol zone design and efficient officer staffing. Police zone designs in large U.S. cities can become outdated if not effectively revised to keep up with population growth, changing transportation patterns, and urban development, leaving vulnerable communities at risk. In addition, police staffing nationwide continues to be constrained by limited budgets and high attrition rates. This project will leverage the vast amounts of data that are available today to create new analytical models and algorithms that promise significant improvements over traditional approaches. The research outcomes in this project will be informed by data from the City of Atlanta, Georgia but will be of value to police agencies nationwide. In particular, the project will develop computational decision support tools with interactive graphical interfaces to visualize police zone reconfiguration and police workload change. The research team will make available computer codes for data analysis, zone design and staff planning free of charge to other interested agencies. The research methods employed in this project will bridge several fields in operations research and data analytics, including spatial-temporal models in statistics, queueing theory in applied probability, and discrete stochastic optimization. The project will create high fidelity models for police emergency service systems using large-scale police reports and census and transportation data. By leveraging modern statistical and queueing methods, the service system will be modeled and analyzed using stochastic methods in order to estimate intensity, location, and categories of emergency calls, as well as the service capacity and travel time of police officers and first-responders. The project uses optimization methods to develop efficient algorithms for zone design and staffing based on these stochastic models. The model’s effectiveness in preventive policing and equitable coverage will be back tested through empirical analysis of policing data. The project will involve graduate students who will be learn how modern optimization methods can be used to improve design in a service sector critical to the nation's security.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.