Heurísticas para o roteamento de fluxos na coleta e entregade recursos entre canteiros de obras

• Ano de defesa: 2017
• Autor(a) principal: Lucas Dietrich Silva Barbosa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Problema de roteamento de veículos; Movimentação de recursos entre canteiros de obras; Heurísticas construtivas; Veículos; Programação heuristica; Engenharia de produção
• Link de acesso: http://hdl.handle.net/1843/BUOS-AUTK8V

This paper deals with the problem MRCO (Movimentação de Recursos entre Canteiros de Obras) which is a generalization of the Vehicle Routing Problem, whose motivation is to treat the transport stream in a company in the business of construction of Belo Horizonte. This problem considers the existence of a set of resources that will be shared by a number of customers (construction sites) and belonging to a group in common. The distribution of these resources is via a set of trained vehicles. A number of depots (residence of drivers) and are home to just a single vehicle. The vehicle must leave and return at the end of a route necessarily to their respective deposit. A resource collected in a construction of a cluster can be used to meet any other site in the same grouping. Each site can be played more than once. Resources can be discretized in items and so both collection and delivery can be divisible. The objective of this problem is to build k vehicle routes that meet the resource allocation requests from the construction sites and minimize the sum of the costs of transport of vehicle. We propose with the objective of treating the problem two constructive heuristics that ensure all restrictions present in the problem.Such developed heuristics are tested in instances close to the real ones and in test instances.The results are compared with company practice and with the method available in the MRCO literature. The results found are superior to those of the company and are better, mainly related to the computational execution time, compared to the method of literature