Maritime Logistics Fleet Size and Mix - development of methods to improve SOT, based on SOR and SOR/SOT in other modes of transport

Maritime fleet size and mix, MARFLIX, a fundamental and complex decision The renewal of a vessel fleet is a fundamental decision for all ship-owners. The number of vessels (fleet size) and the groups of different vessels (fleet mix) can decide the business potential for a ship-owning company. Optimizing the commercial operations and the capacity utilisation of the fleet can only be made within the confines of the present and available fleet. History has shown that being positioned with a good fleet is more important than effective operations, and that the most successful ship-owners are those that manage both. Underlying data for a fleet size and mix problem, are defined by the vessel alternatives, and the future demand for sea-transport or vessel operations. But, when investing in a new vessel with an operative life-time of at least 25 years, uncertainty about future demand, fuel prices, freight rates or contract values, regulations and ship technology, could have an important impact for the vessel investment decision. Hence, uncertainty should be taken directly into account and modelled, so that the decision process is robust, and that the chosen fleet provide adequate flexibility to handle different future scenarios and operational events impacting the vessel segment. The MARFLIX project set out to develop and test methods for decision support regarding vessel fleet size and mix. Strategic decision support in shipping Decisions regarding fleet renewal will in general follow a procedure like this: First, gather market forecasts for the coming years, then use the collective experience and future strategies of the ship-owning company, to make decisions with respect to which vessels to invest in, lay up or scrap. Few ship-owners have analytical tools to support this, in stead relying on experienced planners with spread-sheets and gut-feeling. This is a time and resource demanding process, and a feasible hypothesis is that there is seldom time or resources available to explicitly model uncertainty. To develop an alternative approach, state-of-art theory and methods were reviewed, models were developed and tested, leading to an alternative decision support process for maritime fleet size and mix. The objectives were to introduce the benefit of using analytical tools in a strategic process, then to test a model based approach which explicitly take uncertainty impacting fleet size and mix into account. The researchers approach Researchers from NTNU and MARINTEK used time with stakeholders in a ship owning and operating company, to better understand the fleet renewal process. By developing an optimization model to conduct the fleet renewal calculations, an optimized fleet size and mix selection based upon given optimization criteria was the result per given input scenario. The required time and resource demand was considerably less, which enabled the analysts and stakeholders to spend more time at analysing alternative future scenarios and the results of these. The fleet size and mix decision support process was based upon data and knowledge of future market developments, regulatory changes, and alternative vessel technology. The given approach can better take future uncertainty into account, and most probably reach decisions that are able in a better way to handle the deviation between the unfolding future compared to the estimated future(s) upon which the decision where made. Evaluation The decision support process was tested interactively against decision analysts and other decision stakeholders, focusing on the viability of the approach for supporting strategic decision processes. The feedback has been positive, primarily with respect to the value of quickly assessing many fleet size and mix alternatives, finding new alternatives in the solution space, based on a model that more people could use, supporting organizational robustness. The results have been presented at conferences and in scientific journals, attracting high interest, and the feedback on both the process between researchers and industry, the feasibility of the approach and results have been good. Conclusion The MARFLIX project have based upon close collaboration between researchers and industrial stakeholders, contributed to exchange industrial know-how to the researcher community, and vice versa exchange knowledge of analytical, method based approaches to the maritime industry. Based on a common knowledge platform, an optimization based decision support process for maritime fleet size and mix, have through common testing and revision shown that appropriate use of model-based decision support can contribute to increased insight into fleet renewal (investment) alternatives, better use of organisational knowledge and competence, and providing better organisational robustness in conducting such analyses. Insight into which type of uncertainty and how it impacts such decision processes has also been assessed and documented.

Literature reviews show that the uptake of state-of-art results in applied decision making, is lower for maritime fleet planning than in other modes of transport as road and air. The project objective is to build methodologies for performing strategic fle et analyses. This will be done by combining analytic techniques and methods from different disciplines that are relevant for these kinds of analyses. Given long time-spans and high uncertainty in such decisions, it is valuable to make thorough forecasts a nd possible scenarios for the future demand-supply situation and also for the future technology. When it comes to evaluating the performance of different fleet alternatives, a range of quantitative disciplines are relevant. The FSMP has been considered to some extent within such disciplines earlier. However, there have been few attempts to combine the disciplines in a broader methodology that would be more applicable for the industry. The methodologies developed should be able to solve specific real-lif e problems, and at the same time, the research should strive to keep the methodologies generic and thus applicable for a range of similar cases. Further, the ability to produce reliable quantitative output that can be re-examined and verified is an essent ial feature of the methodology. The decision-maker needs to be confident that the numbers are correct when making major investment decisions. Proving the reliability of results is one of the challenges that have limited the use of quantitative methods fro m operations research, simulation and optimisation in the industry. Another area of focus for this project will be structuring and presentation of the results for high-quality decision support. The output could consist of a large set of variables, and the ir values may point in different directions. It will be aimed for methodologies that do not only provide data, but have the capability to present the results in a way that gives intelligible information.