Strategic Fleet Renewal for Offshore Support Vessels - A maritime fleet size and mix problem
MetadataShow full item record
- Institutt for marin teknikk 
In this thesis, a model is proposed for the maritime fleet renewal problem (MFRP) with applications for offshore support vessels (OSVs). This topic of strategic fleet renewal is important in order to ensure a cost efficient deployment of the future fleet. The central problem to be addressed within this work is to study if it is possible to develop a suitable and relevant model to determine the different aspects of renewing a fleet of OSVs. The MFRP consists of deciding how many ships of each type to use in order to meet future demand. The MFRP is suitable for planning a fleet for a long time horizon, and it finds the best modification of the current fleet of ships in order to adapt to changes in the future market. The proposed model is developed for the MFRP, and contains decision variables that state how many and of what type of ship that should be sold or bought. In addition, tactical decisions are included, such as chartering in or chartering out and fleet operations, while at the same time maximizing profit. The model is a mixed integer programming (MIP) model, developed as a two-stage scenario based model with a stochastic approach. The demand, costs, and revenues are dependent on the uncertainty for this problem, and a scenario is a possible development of the market status for the offshore industry. This scenario-based technique provides advantages in handling the uncertainty of the future market, when modeling this problem. By using stochastic programming, the problem gets a realistic approach on the uncertainty aspects of programming. The model is implemented in commercial software with input data for three test instances, with three scenarios. The test instances are chosen to reflect shipping companies of different sizes. The computational study shows that the model is able to solve all test instances. The main results of the computational study show that the model gives results which indicate that the model works well with a fleet of OSVs. In addition, the results show that the deterministic solution can be sufficient in many of the test cases. The deterministic solution captures the right fleet mix in order to meet future demand, and this can be useful information to reduce the complexity of the problem.When performing a sensitivity analysis, the model structure did not show much sensitivity about changes. This gives an indication that the model is developed in a robust manner, and can withstand impacts from parameter changes in a large degree. However, the input data can contain some sources of error, connected to how the costs and revenues are developing through the planning horizon. The results from the expected value of perfect information indicate that the testing is done with too few scenarios. The scenarios could be improved by introducing a better method for scenario generating, in addition to a probability distribution. The scenarios developed in this thesis can be seen as a representative example, which give the possibility of doing tests and evaluations of the model.Strategic fleet renewal of ships is a crucial and difficult problem in maritime transportation, and the proposed model may serve as a decision support tool for fleet renewal for offshore shipping. The key findings from the computational study have not been the results themselves, but on the different ways in which the model can be handled as a strategic decision support tool for a fleet of OSVs. For the presented problem, there are limitations connected to the lack of earlier studies about this topic. In addition, the computational study is performed based on input data provided by second-hand distributors. The model performs sufficient regarding fleet renewal decisions, and the underlying operational decisions are also satisfactorily performed. The presented model identifies the strategic decisions regarding fleet renewal, in order to maximize profit for future deployment of the fleet.