An algorithm with performance guarantee for the Online Container Relocation Problem

This paper introduces and investigates the Online Container Relocation Problem, where containers have to be retrieved from a bay in a container terminal so as to minimize the number of relocations. Unlike the offline version of the problem, the order of container retrievals is revealed one at a time in an online fashion. We analyze the so-called leveling heuristic using the perspective of worst-case competitive analysis of online algorithms and derive its competitive ratio. We then provide some computational experiments which give insights on the actual average performance of the heuristic.