Algorithm selection approach and features for liner shipping fleet repositioning


Liner carriers construct shipping networks that transport hundreds of millions
of containers on thousands of container ships each year. In order to keep
networks profitable, carriers must reposition vessels between services in the
network. This creates new services or changes the sizes of existing services.
There are a number of methods to solve repositioning problems available, with
some working better than others on certain types of instances. Repositioning
coordinators want to solve repositioning problems as quickly as possible with
the highest achievable profit.

The task of this thesis project is to use algorithm selection techniques to
provide repositioning coordinators with the method most likely to give them a
high quality solution within a provided time-budget. Your thesis will allow
liner carriers to more quickly plan repositioning scenarios and reduce the
overall cost of repositioning vessels.