Autor: MATHEUS OLIVEIRA MEIRIM
Orientadores: Rafael Martinelli
Data e Hora: 17/08/2023, 10h
Link/ Sala:
Banca Examinadora: Rafael Martinelli – orientador – PUC-Rio; Claudio Contardo Vera – Concordia University; Olivier Gallay – HEC Lausanne; Davi Michel Valladão – PUC-Rio.
Resumo:
In recent years, e-commerce has been spreading in society, and the logistics of product delivery is one of the pillars for this market to maintain a high level of service and remain advantageous for consumers to purchase online. This study aims to investigate the last-mile delivery vehicle routing problem for e-commerce and apply the metaheuristic Iterated Local Search (ILS) to optimize the routing of last-mile shipments in a Brazilian e-commerce company. In order to find routes with a higher number of served customers at lower costs for the deliveries to be made, this study proposes an extension to the Vehicle Routing Problem With Occasional Drivers (VRPOD), considering heterogeneous fleets, multiple depots, and occasional drivers delivering more than one package. For the application of the method, data provided by an e-commerce company were used, and they were adequately anonymized to prevent the identification of the company and its customers, respecting ethical principles. A total of 39 instances were used, ranging from 3 to 344 vertices. The results of the proposed model are presented in two scenarios. First, considering that the routing for each depot is independently performed without the use of occasional drivers, and the second scenario considers the availability of occasional drivers to be used in some routes. Both scenarios were compared with the routes generated by the existing company router algorithm, and preliminary results indicate that, for all instances, the number of served customers is higher in 35% of the cases and at least equal in the other instances.