A Multicommodity Integrated Freight Origin–destination Synthesis Model

Abstract

This paper introduces a multi-commodity, single (generic) vehicle formulation of freight ODS model that combines a commodity-based model to estimate loaded truck trips and a complementary model of empty trips. This integration is important because explicit modeling of empty trips—that account for 30% to 40% of total truck trips—is required to avoid significant errors in the estimation of the directional traffic. The formulation is then applied to a case study. Two cases of the proposed model are studied. The first one uses total traffic in the estimation; while the second one is based on loaded and empty traffic. The results conclusively show that the models that consider an empty trip submodel significantly outperform the model that does not in their ability to replicate the observed traffic counts. The comparison between the results from the multi-commodity ODS and the single commodity ODS previously developed by the authors indicates that the multi-commodity formulation brings about substantial reductions in the error associated with the estimation of observed traffic counts. These reductions, in the order of 20% for empty traffic and 40% for loaded and total traffic, seem larger than the spurious improvement to be expected from the increased number of parameters, suggesting that the multi-commodity ODS formulation performs better. The results also show some minor improvements in the ability of the multi-commodity ODS formulation to estimate the OD matrices. In terms of the model's ability to correctly estimate the “true” value of the parameters of the models used, i.e., the parameter values estimated by calibrating the model directly from the OD data, it was found that the multicommodity ODS procedure is able to provide fairly good estimates Noortman and van Es's model parameters, though the parameters of the gravity models that came out to be quite different than the “true” values. The overall assessment of the formulation introduced here is that it represents a solid improvement with respect to comparable techniques.