Modeling Uncertainty in Rail Freight Operations: Implications for Service Reliability

Abstract

This thesis presents an operational simulation tool to evaluate different rail operational policies aimed at increasing service reliability in large-scale multi-carrier rail networks. Operational policies that improve shipment connection reliability at shunting yards, such as priority-based classification, train holding and train cancellation policies can be evaluated using the tool. To support operational decisions needed to implement priority-based classification, an optimization based framework is proposed. Operational policies to improve train schedule reliability, such as including slack time in timetables to handle minor delays, and rescheduling strategies to manage large delays can also be evaluated using the tool. For minor delays, analytical method for deterministic analysis of propagation of delays in train traffic networks is proposed and demonstrated on the Washington DC Metrorail Network. Rescheduling strategies required to manage large delays in multi-carrier rail networks are also discussed herein. A dynamic slot request mechanism is proposed, wherein each carrier requests slots for N blocks ahead, to model rescheduling requests of multiple carriers competing for the slots. The proposed simulation tool is applied on a European rail freight network, the REORIENT network, to evaluate the effect of variability in border crossing times, slack time in timetable design, different rescheduling policies and slot request size (N) on service reliability and average delay to the trains in the system.