Optimization Models for Locating Last Mile Lockers

This work analyzes models for the location of last-mile lockers in the city of São Paulo, with the objective of determining the ideal locations that minimize the distance between lockers and potential customers. Classic models in the literature are explored, such as Coverage, Maximum Coverage, p-Median, and p-Center. In the computer tests, public data from the city of São Paulo are used, including population, area and crime indicators, in order to avoid the allocation of lockers in high-risk regions.
The analysis is conducted in two phases: first, the coverage models are optimized; subsequently, restrictions related to crime are incorporated, limiting the installation of lockers in high-risk areas. In addition to crime, the evaluation is based on the models’ metrics: sum of distances, maximum distance, and customer coverage. Thus, the study contributes to the debate on urban logistics by representing a network of lockers that seeks both efficiency and security.