Decarbonizing Freight Transportation: Joint Optimization of Intermodal Cargo Routing, Scheduling, and Service Networks

Abstract

Intermodal transportation, which exploits the strengths of alternative modes of transportation, such as rail and water, offers a promising solution for the growing freight transportation sector to reduce carbon emissions and operating costs. Optimization of the intermodal transportation system has been widely studied in the literature, but conventional approaches typically focus on the broad averages of emission and cost statistics or rely on existing inflexible timetables that lead to suboptimal results. In this study, we optimize the U.S. intermodal system consisting of highway, railway, and waterway freight transportation. We construct a mixed integer programming model on a muti-layered transportation network to jointly optimize freight routing, scheduling, and service network design decisions. The model minimizes intermodal operating costs, carbon emission costs as carbon taxes, and late penalties for delayed freight arrival. The complexity of the intermodal system imposes great computational challenges on state-of-the-art solution methods. To address this, we investigate mathematical properties of the model to explore structural benefits. We further propose a vehicle scheduling decomposition approach that expands the Lagrangian cutting plane methods towards nonlinear linking constraints. Structural properties are integrated into the decomposition to derive novel algorithms. Real-world intermodal infrastructure, cost, and freight projection data are collected to validate the model and the algorithms through comprehensive experiments. Carbon emission data from combinations of vehicle engines and fuel types are utilized in large-scale case studies to provide insights on intermodal decarbonization, targeting a carbon-neutral U.S. freight transportation sector by 2050.