Meta-heurística para planejamento da logística reversa de pneus inservíveis para atender uma planta geradora de energia com base no modelo Two-Echelon Capacitated Vehicle Routing Problem

• Ano de defesa: 2021
• Autor(a) principal: Barbosa, João Henrique Brunow
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Espírito Santo; BR; Mestrado em Engenharia Civil; Centro Tecnológico; UFES; Programa de Pós-Graduação em Engenharia Civil
• Palavras-chave em Português: Logística reversa de pneus; 2E-CVRP; meta-heurística; simulated annealing; subject.br-rjbn; Engenharia Civil
• Link de acesso: http://repositorio.ufes.br/handle/10/15335

Approximately 800 million tires are discarded annually in the world. This residue causes a great environmental impact when improperly disposed of, considering that they have an unknown decomposition time. In urban centers, one of the collection strategies is carried out in two stages. In the first stage, collection points such as tire stores and workshops receive waste tires directly from users, where vehicles collect and transport them to intermediate warehouses. In the second stage, vehicles collect the cargo leased from the intermediary warehouses and deliver the cargo to the recycling company so that it can be properly disposed of. The problem in the literature that most resembles the problem described above is the Twoechelon Capacitated Vehicle Routing Problem (2E-CVRP), which is a two-level vehicle routing problem. On the first level, transport is carried out by large vehicles and departs from a central warehouse to cargo distribution and consolidation centers. On the second level, transport is carried out from distribution centers to end customers using smaller vehicles. In view of their similarities, one can consider the problem of reverse tire logistics as a 2E-CVRP. This thesis proposes an algorithm inspired by Simulated Annealing for the problem. The development of the algorithm proved to be necessary, since the mathematical model proposed above for this problem was not capable of finding solutions in a reasonable execution time for a practical application. The algorithm was developed using C language, using the Dev-C++ development tool.