Solving the Twin Hub Bundling Problem

Abstract

Optimisation of transport networks can yield substantial reductions in transport costs and improve the quality of service for the users of these networks. When optimising intermodal rail-road networks, organising large transport volumes on transport links is key for bringing down the costs per container-kilometre. By bundling container flows in Twin Hub Bundling (THB) networks, thereby integrating the flows of the Dutch and Belgian seaports, the Twin Hub project aims to stimulate intermodal rail-road transport in North-West Europe. The goal of this thesis is to create a bundling tool that finds the most promising THB networks based on projected container flows. In this thesis, a modelling is presented of the THB problem, introducing three innovative features: the organisation of intermodal transport services in batches, the introduction of indirect trains, and the integration of train and truck services in one service network. We then prove this problem to be an NP-hard optimisation problem, and present an exact algorithm based on Integer Linear Programming (ILP) as well as a relaxation of that formulation. An experimental evaluation shows that both the exact algorithm and the combination of the relaxation and the ILP algorithm (cascade algorithm) are capable of producing good-quality solutions to real-world sized instances within 24 hours. Finally, we present the bundling tool itself, which provides an easy interface for running the algorithms, evaluating the results, and selecting the most promising THB network to be implemented in a pilot operation.