A coupled linear programming model with geospatial dynamic trip assignment for global-scale intermodal transportation

We present an ad hoc geospatial origin–destination model that utilizes the rudiments of operations research optimization techniques to estimate feasible multimodal routes that may or may not be geospatially connected. The model is designed to determine and optimize large-scale container flows from US trade partners to inland markets mainly located within the lower 48 states. The dynamic traffic assignment for the freight flow considers several factors such as congestion, volume/capacity (V/C) ratio, throughput and distance-based impedance and the Geographic Information Systems platform provides a visualization environment for all feasible routes. Visual analytics with the resultant geographical distribution maps indicate that maritime containers are concentrated on the West Coast whereas rail shipments are densely concentrated along rail transshipment points such as Chicago, Memphis and Dallas. The highway transportation densities are mainly visible with origins from marine ports to associated market destinations starting from hinterland and radiating further away from these primal areas to other inland intermodal terminals that serve major landlocked metropolitan cities that are not served by railway networks. Our critical finding indicates that some routes, which are not generally considered as prime routes, may actually offer the best alternatives especially when considering global supply chains.