Multi-objective optimisation of ship routes

Abstract

In this master thesis two different approaches of solving a three-criteria multi-objective ship route optimisation are developed and compared. The first is a grid search approach, while the second one is a modification of the distance based Pareto genetic algorithm, which has previously been proven useful in other multi-objective optimisation problems. It is found that the modified distance based Pareto genetic algorithm can give an equivalently good result using approximately 1% of the computing time, but also that a penalty needs to be introduced in the second method in order for it to be used in an actual product. Keywords: Ship route optimisation, Multi-objective optimisation, Weather routing, Voyage planning, Genetic algorithms