Competitive Performance Assessment of Dynamic Vehicle Routing Technologies Using Sequential Auctions

Abstract

Technologies for a dynamic truckload pickup-and-delivery problem in a competitive environment by use of sequential auctions are compared. In this environment, demands arrive randomly over time and are described by pickup-and-delivery locations and hard time windows. On demand arrival, carriers compete for the loads in a second-price auction. Four fleet assignment technologies with different degrees of sophistication are tested with simulations. The technologies differ in how they deal with the combinatorial and stochastic elements of the online problem. A one-step look-ahead dynamic vehicle routing technology that tries to estimate the impacts of current decisions on serving future loads (not yet arrived) is introduced. The performance of each technology is analyzed in relation to different demand characteristics.