Modelo computacional de simulação para análise de crimes urbanos

Abstract

Crime is a complex problem that causes physical and psychological harm to victims and also causes significant financial losses to the State. Computer simulation can represent the functioning of a system and provide a dynamic view of it. When applied to the problem of crime, it has the potential to show trends and situations that are more or less prone to problems and to project scenarios for public administration and resource allocation. This work develops a computer simulation model with the objective of evaluating how variables such as population density, criminal motivation and police presence influence the occurrence of crimes. The model is agent-based and developed in NetLogo. Data provided by the State Secretariat of Public Security of Paraná (SESP-PR) were used to calibrate the model and represent the local reality. The construction of the urban environment was carried out with QGIS and georeferenced maps that represent strategic points and areas of highest crime incidence. The results show that the model reproduces patterns of crime concentration in areas of greatest risk, identified in the heat maps generated by the simulation. Furthermore, the variation of parameters such as criminals’ range of action and level of motivation allowed an analysis of the factors that influence crime concentration. Some limitations, such as the need for detailed data and the complexity of representing human behavior, indicate areas for improvement. Future work can integrate real-time data and routines closer to reality, expanding the applicability of the simulation in different urban contexts.