In empty loading unit repositioning, there is a set of empty and reusable loading units that should be shipped back from the location
they are emptied to the locations they are needed. These loading units do not directly incorporate in the profit of transportation,
but guarantee requested service level. In empty repositioning problem, future customer demand is unknown and the objective
function is to minimize empty transportation plus storage costs, stockout costs, and in some cases, substitution costs. These problems
are usually modeled as NFP problems with continuous variables to represent the movement of the loading units. In the
literature, Erera, Morales, and Savelsbergh (2005) compare a base repositioning strategy (a current state-of-the-practice) with three
alternative strategies which integrate the repositioning and routing of the containers simultaneously. These three strategies are
weekly, bounded daily, and unbounded daily repositioning. Erera et al. (2005) compare these strategies on a network with 10 ports
and 900 orders and up to 1000 containers. They show that the proper timing of repositioning is more important than deciding
on the number of containers to be repositioned, and unbounded daily repositioning in overall is the best strategy.