An examination of railroad capacity and its implications for rail-highway intermodal transportation.

摘要

After many years of decline in market share, railroads are now experiencing an increasing demand for their services. Service intensive intermodal transportation seems to be an especially promising market area. Since the historic decline in traffic has been accompanied by a reduction in network infrastructure, however, the railroads' ability to handle sizable traffic increases, at least in the short term, is in question. Since rail transportation is critical to the domestic economy of the nation, and is increasingly important in international logistics channels, shortfalls in railroad capacity are not desirable.; The published literature on railroad capacity is relatively sparse, especially in comparison to the highway mode. Much of what is available pertains to individual network components such as lines or terminals. Evaluation of system capacity, considering the interactive effects of traffic flowing through a network of lines and terminals, has received less attention. A tool specifically designed for evaluating freight railroad system capacity issues could be a useful addition to the rail analyst's toolbox.; The research conducted in this study resulted in the formulation and application of RAILNET, a multicommodity, multicarrier network model for predicting equilibrium flows within a railroad network. Designed for strategic planning with a short term horizon, the model assumes fixed external demand. The predicted flows meet the conditions for Wardropian system equilibrium. At completion, the solution algorithm predicts the expected delay per train on each link, allowing the analyst to identify areas of congestion.; Following completion of the model, it was applied to a case study examining the railroad network in the southeastern U.S. The public use version of the Interstate Commerce Commission's Commodity Waybill Sample (CWS) provided flow data. The dissertation describes the procedure used to develop the case study and presents some results. The case points to major deficiencies in the CWS data which resulted in substantially less traffic in the network than is actually present. In general, given this limitation, the model behaved well and results appear reasonable, although not necessarily reflective of actual network conditions.