Aplicação de Metaheurísticas para o problema de roteamento de veículos dinâmico para transporte reativo a demanda

• Ano de defesa: 2012
• Autor(a) principal: Miranda, Dângelo Silva
• Orientador(a): Santos, André Gustavo dos
• Banca de defesa: Raggi, Luiz Aurélio
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Viçosa
• Programa de Pós-Graduação: Mestrado em Ciência da Computação
• Departamento: Metodologias e técnicas da Computação; Sistemas de Computação
• País: BR
• Palavras-chave em Português: Roteamento de veículos; Estático; Dinâmico; Transporte reativo a demanda; Metaheurísticas
• Palavras-chave em Inglês: Vehicle routing; Static; Dynamic; Demand responsive transportation; Metaheuristics
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://locus.ufv.br/handle/123456789/2643

Vehicle routing problems have great interest in operations research, as they are logistics problems related to the delivery of goods and people transportation, and because they have applications in several fields of economy. This work deals with static and dynamic vehicle routing problem for demand responsive transportation. Being a responsive transportation, passengers' needs must be considered also along the route planning, in order to assure the quality or service, route flexibility and vehicle usage effectiveness, with minimum impact in the costs for the company. Since the problem has combinatorial characteristics, a Reactive GRASP heuristic is proposed for the static case. The randomness parameter used in construction phase has its probability self-adjusted. To find the approach best suited for the set of instances, different combinations of strategies and neighborhoods are proposed and tested. Besides that, it is proposed an hybrid algorithm, substituting the local search of the Reactive GRASP by an Tabu Search method. Furthermore, dynamic replanning algorithms were proposed and tested for the dynamic scenario. The results of the proposed methods are compared one to each other and to results of algorithms from the literature. The experimental analisys shows that the proposed methods give satisfactory results, as the metaheuristics obtained in average solutions 10,27% better than those from the literature in the static scenario, and the dynamic replanning methods improve the results even further.