Integrated Origin–Destination Synthesis Model for Freight with Commodity-Based and Empty Trip Models

Origin–destination (O-D) synthesis techniques offer the potential of producing estimates of freight O-D matrices by using secondary sources such as traffic counts, bypassing the need for expensive data collection efforts. In the case described in this paper, this promise is particularly important because of the significant cost associated with conducting surveys to obtain O-D patterns of freight movements and because of the reluctance of freight providers to provide information they consider commercially sensitive. The paper proposes an integrated O-D synthesis model that combines a commodity-based model to estimate loaded truck trips with a complementary model of empty trips. This integration is important because explicit modeling of empty trips—which account for 30% to 40% of total truck trips—is required to avoid significant errors in the estimation of directional traffic. The proposed model is applied to a case study for which both the actual O-D matrix and traffic counts are known. Two objective functions are considered for the proposed model based on two scenarios: (a) only total link traffic is known and (b) only the split of loaded and empty link traffic is known. With data from a case study, it was found that the proposed model performs significantly better than an alternative formulation that does not consider an empty trip model. In addition, it is observed that the added information in terms of observed empty trips improves the O-D estimates. The model was found to produce reasonable estimates of the true parameters of the underlying models.