Quantitative methods for ENEM 2024

The 2024 Brazilian National High School Exam (ENEM) registered over 4.3 million confirmed registrations across approximately 1,800 municipalities, constituting a large-scale logistical challenge. In this work, we developed a candidate geolocation module using IBGE data and a mixed-integer programming allocation model tailored to Enem rules. We also present a performance comparison against a greedy heuristic and detail the computational strategies employed. The results show a 70% reduction in the average travel distance of participants compared to the previous edition, without increasing allocation costs. This experience demonstrates the effectiveness of modular and auditable solutions for managing large educational examinations.