School Bus Routing: an integrated approach of clustering and optimization applied to CEFET-RJ Nova Iguaçu Campus

This study addresses the School Bus Routing Problem (SBRP) in the context of CEFETRJ
Campus Nova Iguaçu, aiming to quantitatively assess the logistical feasibility of a student
transportation system. We propose an integrated framework that combines geospatial data processing,
clustering via the K-means algorithm with the number of clusters determined by the Kneedle
method, and optimization through a Mixed-Integer Linear Programming model for the Vehicle
Routing Problem with TimeWindows (VRPTW). The model is solved using the CPLEX optimizer,
followed by a heuristic procedure for route sequencing. Results indicate a reduction in fleet size
from 20 to 12 vehicles in peak scenarios, highlighting relevant logistical gains. The study also points
out computational limitations of exact methods for large-scale instances and sensitivity to data
quality. As a contribution, it provides an applied case study to support decision-making in school
transportation systems.