Three-stage service network design in rail-road networks with demand and capacity uncertainty

Abstract

This paper addresses the service network design (SND) problem in a railroad network with demand and capacity uncertainty from the perspective of a logistics service provider (LSP) in a synchromodal transport setting. Based on consultations with LSPs, we propose a three-stage model that includes an initial stage to book capacity, an intermediate stage to update capacity decisions with partial information in the form of improved demand forecasts, and a third stage with complete information for capacity updates, recourse actions and routing decisions. Stochasticity is included by means of a scenario generation approach which results in a scenario tree. The impact of replanning rail capacity once when complete information is available, and twice with both partial and complete information, is compared against the case without replanning on a set of theoretical instances with a full factorial design. Results indicate that replanning once when complete information is available yields large cost savings and increases the share of rail transport. Replanning is more advantageous with cheaper train services, networks with more transhipment options, more rail capacity, and higher levels of demand uncertainty. Introducing an additional replanning stage, with partial information, yields diminishing returns, though savings increase with the amount of uncertainty that is resolved at the time of replanning.