Modeling approaches for empty container management

Abstract

Although empty container repositioning is a costly and non-revenue generating activity, it is an integral part of an overall efficient global transportation system. Due to the natural imbalance of trade, certain areas develop a surplus of containers, while others have a deficit. As a consequence, there is a need for carriers to reposition their empty containers in order to be able to fulfill future demand of customers for empty containers. Furthermore, at a regional level, empty container movements around ports contribute considerably to regional traffic congestion, which is a source of external costs such as air pollution, wasted energy and driver inefficiency. Currently, empty container depots are usually located nearby ports. After being unloaded at a customer's site, most empty containers are directly moved back to the port of origin for global repositioning or in expectation of future requests for empty containers. When a demand for empty containers arises in the hinterland, empty containers are subsequently transported back and forth again. An opportunity for optimizing empty container ows exists in partly preventing empty containers to be transported back to the port immediately. The objective of empty container management at a regional level is to reposition empty containers efficiently in order to minimize costs, while fulfilling empty container demands. Several ways of achieving this objective are proposed in literature, such as inland depots for empty containers, street turns (moving containers directly between importers and exporters), container substitution (fulfilling a request for a container of one type by a container of another type) and container leasing. Planning models for these proposed solutions are described in scientific literature. The modeling of empty container repositioning is rather complex due to several reasons. Uncertainties concerning demand and supply of empty containers play an important role. Secondly, container substitution is only possible for some, not all, empty container requests. Finally, rationalizing empty container repositioning requires decisions to be taken at all three decision-making levels: strategic, tactic and operational. Therefore, designing a single model for the empty container repositioning problems is not feasible. An overview of the proposed modeling techniques is given and opportunities for future research are identified.