Abstract
Empty container logistics is a significant component in the shipping industry’s operational cost management. The logistics process involves decisions regarding holding, leasing and repositioning empty containers. These decisions get increasingly more complicated as the number of ports in the network increases, adding different cost parameters. In this study, we consider the problem of optimizing the total cost related to empty container logistics. The objective of the optimization is to find a structure to the cost and provide a decision support tool for logistics managers to perform strategic planning and find cost saving opportunities. The first part of the thesis focuses on modelling a two-port setting. Extensive numerical experiments are carried out. From the numerical results, we observed that the value function is convex. We take advantage of this result to improve the efficiency of the optimal algorithm by pruning the decision space in the dynamic programming (DP) model. In addition, we find that a carrier would only have to send the net difference of container demand between two ports. This observation allows us to halve the number of decision variables. To further deal with large-scale problems, we propose a heuristic that overcomes the curse of dimensionality which the DP algorithm suffers from. We also provide an integer programming (IP) model for a network of ports which gives us optimal solutions. This formulation is implemented for an actual shipping company’s network of 48 ports and the resulting heuristic and IP’s performances are then compared to the company’s existing empty repositioning policies. The results are encouraging and show that there are further cost saving opportunities of 25% and a reduction of 5% in fleet size for the company. Numerical results, under the optimal policy, shows that we tend to hold containers for export opportunities, instead of aggressively repositioning empty containers, to cut down on high empty repositioning costs. If empty repositioning is necessary, there are cost savings opportunities in pre-emptively sending containers and holding them at the destination port where the holding costs are lower. The results also show that empty container related costs are proportional to fleet sizes and uncertainty in demand can lead to cost increase. The shipping company is advised to consider the uncertainty that they face and optimize the different scenarios with our model to understand the cost fluctuations. The cost difference should be managed properly to ensure profitability.