Stochastic Service Network Design for Intermodal Freight Transportation

In view of the accelerating climate change, greenhouse gas emissions from freight transportation must be significantly reduced over the next decades. Intermodal transportation can make a significant contribution here. During the transportation process, different modes of transportation are combined, enabling a modal shift to environmentally friendly alternatives such as rail and inland waterway transportation. However, at the same time, the organization of several modes is more complex compared to the unimodal case (where, for example, only trucks are employed). In particular, an efficient management of uncertainties, such as fluctuating transportation demand volumes or delays, is required to realize low costs and transportation times, thereby ensuring the attractiveness of intermodal transportation for a further modal shift. Stochastic service network design can explicitly consider such uncertainities in the planning in order to increase the performance of intermodal transportation. Decisions for the network design as well as for the mode choice are defined by mathematical optimization models, which originate from operations research and include relevant uncertainities by stochastic parameters. As central research gap, this dissertation addresses important operational constraints and decision variables of real-life intermodal networks, which have not been considered in these models so far and, in consequence, strongly limit their application in everyday operations. The resulting research contribution are two new variants of stochastic service network design models: The "stochastic service network design with integrated vehicle routing problem" integrates corresponding routing problems for road vehicles into the planning of intermodal networks. This new variant ensures a cost- and delay-minimal mode choice in the case of uncertain transportation times. The "stochastic service network design with short-term schedule modifications" deals with modifications of intermodal transportation schedules in order to adapt them to fluctuating demand as best as possible. For both new model variants, heuristic solution methods are presented which can efficiently solve even large network instances. Extensive case studies with real-world data demonstrate significant savings potentials compared to deterministic models as well as (simplified) stochastic models that already exist in literature.