Problema de composição de frota: uma abordagem sequencial de simulação-otimização com horizontes rolantes

• Ano de defesa: 2019
• Autor(a) principal: Figueira de Faria, Luis Filipe
• Orientador(a): Silva, João Eduardo Azevedo Ramos da
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus Sorocaba
• Programa de Pós-Graduação: Programa de Pós-Graduação em Engenharia de Produção - PPGEP-So
• Palavras-chave em Português: Simulação; Otimização; Horizontes Rolantes; Pesquisa Operacional; Composição de frota
• Área do conhecimento CNPq: ENGENHARIAS::ENGENHARIA DE PRODUCAO::PESQUISA OPERACIONAL
• Link de acesso: https://repositorio.ufscar.br/handle/20.500.14289/11443

To physically distribute finished goods, shippers can hire third-party logistics (3PL) in order to meet the forecasted demand by the customers. Nevertheless, the contract made between shippers and 3PL is based on decisions that need to be considered over several time horizons: strategic, tactical and operational. This question, known as fleet dimensioning or fleet composition problem, has been vastly approached in literature, divided mainly into two Operational Research’s techniques: Systems Simulation and Optimization. However, it is known that these techniques when applied in a pure or isolated form are not capable of delivering the adequate solution for the problem, once they need several simplifications, in the case of Optimization, or require an exhaustive analysis process, in the case of Simulation. The aim of this study is to propose a sequential simulation-optimization method with rolling horizons to integrate the solutions given by both techniques and provide a much more adequate result for the decision maker. As result, the more realistic dynamical and stochastic characteristics of the simulation model and the optimized fleet provided by the optimization model are successfully united in a rolling horizon environment. This combination has been analyzed over three different scenarios of demand variation, from 10 to 30%. When comparing to the isolated optimization, it can be highlighted the quantity of SPOT vehicles needed to overcome the unpredictability of transport solicitations and the relative costs of not considering simulation on decision making, which can increase the final costs of dimensioning implementation by up to 30%. Therefore, it becomes possible for the shipper to periodically re-evaluate his fleet dimensioning, providing a much safer decision-making process and a much better use of resources.