Uma abordagem de computação aplicada para otimização espacial de unidades policiais

Abstract

The growth of crime in large urban centers intensifies challenges related to territorial planning and the efficient allocation of public safety resources. In this context, this work proposes a methodological framework to support the strategic location of police facilities by integrating geospatial databases, multicriteria decision analysis (MCDA), and spatial clustering algorithms. The approach combines data engineering processes, weighting of urban and crime-related variables through the Analytic Hierarchy Process (AHP), and spatial clustering techniques. As an applied case study, K-means was applied to 554,832 crime records from the Greater London metropolitan area, associated with spatial and urban infrastructure data. Validation was performed using the Elbow, Silhouette, and Calinski-Harabasz metrics, enabling the identification of spatially coherent clusters to support decision-making.