| dc.description.abstract | An important challenge within the offshore industry and sea traffic preparedness is how to plan and optimise the positioning of a fleet of vessels in a dynamic, everchanging environment such that operating costs and risks related to health, the environment, and safety are reduced, whilst services and profit are improved. In Norway, the Norwegian Coastal Adminstration runs a vessel traffic service centre in the town of Vardø that is responsible for a fleet of tugs patrolling the northern coastline of Norway. Tug fleet optimisation algorithms can be designed to solve the problem of dynamically positioning such a fleet of tugs in order to mitigate the risk of oil tanker drifting accidents. In this project, we modelled the tug fleet optimisation problem in both 1D and 2D and developed new methods for solving it, including 14 different variants of a receding horizon genetic algorithm as well as several variants of a receding horizon mixed integer programming algorithm. We also developed two evaluation heuristics for measuring the performance, or merit, of such algorithms. The algorithms were tested for a large number of simulation scenarios. Extending the original project description, we also re-implemented the receding horizon genetic algorithm in a more suitable programming language and studied parallelisation of genetic algorithms and pseudo-random number generators. The work resulted in several conference presentations and seven publications, far exceeding the target of two publications. Research on fleet optimisation has continued after the project ended through the work of a PhD candidate expected to complete his work late 2016. In addition, the knowledge, skills, and methods gained and developed during the project have led to further research funding in other research areas as well as integration in research-based education both at the bachelor and master level at Aalesund University College. | nb_NO |
