PROBLEMA DE ALOCAÇÃO DE BERÇOS EM PORTOS GRANELEIROS COM RESTRIÇÕES DE ESTOQUE E CONDIÇÕES FAVORÁVEIS DE MARÉ

• Ano de defesa: 2010
• Autor(a) principal: Barros, Victor Hugo
• Orientador(a): PAIVA, Anselmo Cardoso de
• Banca de defesa: Silva, Aristófanes Corrêa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do Maranhão
• Programa de Pós-Graduação: PROGRAMA DE PÓS-GRADUAÇÃO EM ENGENHARIA DE ELETRICIDADE/CCET
• Departamento: Engenharia
• País: BR
• Palavras-chave em Português: Programação linear; Problema de alocação de berços; Portos com restrições de maré; Controle de estoque
• Palavras-chave em Inglês: Linear programming; Berth allocation problem; Tidal ports; Stock control
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA::PROBABILIDADE E ESTATISTICA::PROBABILIDADE::ANALISE ESTOCASTICA
• Link de acesso: http://tedebc.ufma.br:8080/jspui/handle/tede/434

The problem of allocating berth positions for vessels in tidal grain port terminals is discussed in this work. A berth defines a specific location alongside a quay where a ship loader is available for loading or unloading vessels, accommodating only one vessel at time. In tidal ports, draft conditions depend on high tide conditions, since available depth under the low tide is not adequate to the movement of ships. Some port terminals at the port complex of São Luís, Maranhão, are associated to important transnational enterprises which maintain a strong control over the stock level of their goods. Since the stock level sometimes depends on a continuous process of consumption or production of grains, the decision making of loading or unloading vessels must contemplate the amount of the grain stored in the port yards. Therefore, a basic criterion for decision making is to give priority to the vessels related to the most critical grain stock level. This paper presents two integer linear programming models based on the transportation problem to represent the discussed problem. Some problem instances could be solved by a commercial solver. As an alternative to larger instances, which require large running time, an implementation of Simulated Annealing (SA) and the algorithm known as Population Training Algorithm for Linear Programming (PTA/LP) are used to solve the problem.