| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.11 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A Base Naval de Lisboa é um órgão de base da Marinha que tem por missão assegurar as atividades relacionadas com o apoio logístico às unidades navais atracadas/estacionadas na sua área de responsabilidade. Decorrente da elevada requisição de cais por parte dos navios, fruto das restrições orçamentais impostas ao Ramo Marinha, a gestão dos cais e das atividades logísticas reveste-se de grande complexidade. Nessa continuidade e, como solução àquela problemática, propõe-se elaborar/implementar um protótipo de um sistema de apoio à decisão (SADEN), que possibilite ao decisor criar instâncias do problema de estacionamento de navios assim como encontrar as soluções admissíveis, visualizá-las e, dessa forma, permitir propor automaticamente a distribuição de navios, pelos cais disponíveis, conduzindo à resolução de um problema da Base (definido na literatura como berth allocation problem) e simultaneamente, a médio/longo prazo, permitir a redução de custos e ganhos de tempo. O problema de estacionamento é modelado como um problema de Programação Linear Binário (PLB) e foram consideradas duas variantes: a primeira consiste em afetar navios aos cais em função da adequabilidade expressa através de uma função de utilidade e a segunda variante considera o problema de estacionamento ao longo de um período de tempo onde são conhecidas as saídas e chegadas nesse período e adequabilidade entre navio e cais e se procura minimizar mudanças de cais para o mesmo navio, que ocorreria para permitir a saída de outros navios. No final são apresentados resultados numéricos dos modelos implementados e as principais conclusões verificadas. Destas, salienta-se a técnica de modelação que sustenta a criação das instâncias dos problemas onde a dimensão destas é inversamente proporcional às restrições impostas pelo decisor. Esta técnica de modelação, através de uma tabela de variáveis, baseia-se na eliminação de variáveis que decorre dos requisitos impostos pelo decisor. Como resultado final, torna-se possível resolver as instâncias recorrendo a métodos exactos como sucede com a utilização de algoritmos de Branch & Bound.
The Lisbon Naval Base (BNL) is one of the base organs of the Portuguese Navy whose mission is to support all activities related to the logistic support to naval forces within its area of responsibility. Imposed budget constraints have placed a highly demand for pier availability, since vessels have to spend more time in the dock rather than in missions, ledding to an increase effort to manage available piers and logistic activities within the BNL. Therefore, to solve this situation, the present dissertation proposes to elaborate/implement a prototype of a Decision Support System (DSS -SADEN) that will help decision makers to create instances of the ship allocation problem, obtain their respective solution, visualize them and thus, allow the automatic distribution of ships by available berths, leading to the resolution of a problem of the Base (defined in the literature as a berth allocation problem), and simultaneously, in the medium / long term, to allow the reduction of costs and gains of time. The BAP variant is modelled as a Binary Linear Programming (BLP) problem where two variants are considered: the first variant consists in a simple allocation problem where vessels should be affected to piers, in accordance with the suitability expressed through a utility function The second variants considers a time period where arrivals and departures are known and the purpose is to allocate vessels to piers minimizing movements of vessels between piers that are not departing during that period. Finally, numerical results for the implemented models are presented as well as the main conclusions. The one that is enphasized the most is the modeling technique that supports the creation of instances of problems where the dimension of these is inversely proportional to the constraints imposed by the decision maker. This modeling technique, uses a table of variables and is based on the elimination of variables that results from the requirements imposed by the decision maker. As a final result, it is possible to solve instances using exact methods such as using Branch & Bound algorithms.
The Lisbon Naval Base (BNL) is one of the base organs of the Portuguese Navy whose mission is to support all activities related to the logistic support to naval forces within its area of responsibility. Imposed budget constraints have placed a highly demand for pier availability, since vessels have to spend more time in the dock rather than in missions, ledding to an increase effort to manage available piers and logistic activities within the BNL. Therefore, to solve this situation, the present dissertation proposes to elaborate/implement a prototype of a Decision Support System (DSS -SADEN) that will help decision makers to create instances of the ship allocation problem, obtain their respective solution, visualize them and thus, allow the automatic distribution of ships by available berths, leading to the resolution of a problem of the Base (defined in the literature as a berth allocation problem), and simultaneously, in the medium / long term, to allow the reduction of costs and gains of time. The BAP variant is modelled as a Binary Linear Programming (BLP) problem where two variants are considered: the first variant consists in a simple allocation problem where vessels should be affected to piers, in accordance with the suitability expressed through a utility function The second variants considers a time period where arrivals and departures are known and the purpose is to allocate vessels to piers minimizing movements of vessels between piers that are not departing during that period. Finally, numerical results for the implemented models are presented as well as the main conclusions. The one that is enphasized the most is the modeling technique that supports the creation of instances of problems where the dimension of these is inversely proportional to the constraints imposed by the decision maker. This modeling technique, uses a table of variables and is based on the elimination of variables that results from the requirements imposed by the decision maker. As a final result, it is possible to solve instances using exact methods such as using Branch & Bound algorithms.
Description
Keywords
Berth allocation problem, modelos de programação linear inteira, Base Naval de Lisboa Berth allocation problem; Integer Linear Programming Models, Base Naval de Lisboa