"What Ya Gonna Do When They (Don’t) Come For You": Scheduling Algorithms for Geographically Equitable Police Dispatch

Abstract

In the year 2019, the debate over equity in Policing has gripped American attention. However while most research works to analyze race and gender bias in police law enforcement, few studies seek to understand equity as it applies police emergency response. In this micro-thesis I utilize 911 Call and Response Data from the Detroit Police Department, to prove that significant relationships exist between race and emergency police response time. Specifically, I demonstrate that 911 calls which originate from zip codes with higher percentages of African Americans, experience longer wait times in the police dispatch queue. Furthermore, this micro-thesis serves to formalize a scheduling model of police dispatch, such as to secure an equitable delivery of service to all groups across the city. The main conceptual contribution is the formalization of a scheduling model which seeks not to minimize flow-time, but minimize variance between the wait time of jobs split between unrelated machines. Such research bears striking implications on study of Algorithmic Fairness and Scheduling Theory application.