Stochastic scheduled service network design in the presence of a spot market for excess capacity

Abstract

We study the impacts of a priori knowledge of spot markets for excess capacity on stochastic service network design problems. We do this by investigating whether it is beneficial to take spot markets into account at the design stage when building a transportation plan facing uncertain demands. We show that in most situations, taking spot markets into account at design level reduces the total costs, not just from the benefit of selling excess capacity, but also from changes in the design of the service network itself. We also show that when deterministic models are used to produce solutions for stochastic service network design problems, introducing spot markets at the design stage has undetermined effects. This is important since in most real cases, only the deterministic models can actually be solved. We only study cases where it would not be profitable for our vehicles to operate solely in the spot market as the income per capacity unit is much lower than our costs per capacity unit; the spot market is a means to obtain some income from excess capacity, but it is not our primary market.