Optimization for Container Logistics: Freight Containerization and Berth Allocation with Transshipment Handling

Container logistics is at the core of global trade, with around 90% of non-bulk goods being transported in these standardized steel boxes between production and consumption sites. Ensuring the smooth, timely flow of these containers
along shipping routes is therefore critical for the efficiency and reliability of supply chains. However, the complexity of container logistics operations gives rise to a multitude of optimization challenges throughout the container’s journey, which often require the use of advanced Operations Research methods to be addressed. This thesis aims to tackle two such problems faced by the major container logistics company Maersk.

The first problem revolves around the containerization of freight at a warehouse. Here, the logistics provider must determine the minimum-cost set of containers needed to transport a set of cargo items, while adhering to mixing constraints imposed by the cargo owner. The problem can be formulated as a Multi-Class Constrained Variable Size Bin Packing Problem, and we propose several heuristics to find solutions as part of the first chapter of the dissertation. Those are benchmarked against lower bounds calculated using a Column Generation approach, and are shown to improve the quality of loading plans.

The second problem deals with the optimization of berth allocation at container terminals, where a particular focus is placed on the handling of transshipments, which are operations where containers are transferred from one vessel to another. Transshipments require careful coordination of vessel berthing times, and missed connections require costly rerouting of affected containers. In the second chapter of the dissertation, we consider the problem from the perspective of terminal planners that seek to minimize inconvenience to liner shipping companies, and introduce a mathematical formulation. A heuristic approach is benchmarked against an exact method, highlighting the need for further research. The third chapter deals with the same problem, this time from the perspective of a liner shipping company which seeks to recommend berth plans to the terminal that satisfy its own performance indicators without affecting those of third parties. A weighted-sum optimization model is implemented and tested on various real-life instances for different prioritizations of performance indicators. Berth allocation and transshipment handling are solved in an integrated way, and this approach is compared to sequential ones.

Together, these contributions demonstrate how tailored optimization models can support more efficient decision-making in container logistics, improving the quality of operational plans in both freight containerization and berth allocation.