Positioning Drone Bases on Highways via MILP with Hybrid Coverage and Proximity Function

In Brazil, traffic accidents result in approximately 37 thousand deaths annually, corresponding to about 3.1% of global deaths, a proportion higher than its population representativeness. Drones emerge as a first response alternative and can speed up rescue by allowing a quick view of the scene and the delivery of supplies, especially in hard-to-reach places. This work proposes the Neighborhood Capacitated Maximum Covering Location Problem (N-CMCLP), a variant of the CMCLP that progressively penalizes accidents far from the base, favoring more central positions. The model is validated by discrete event simulation on 25,539 real records of accidents on federal highways in Pernambuco (PRF, 2017–2025). The N-CMCLP attends to 60% of accidents within 10 minutes (against 15% of the CMCLP), with a 44% reduction in the average response time and 70% in losses due to time.