Fast train: A computationally efficient train routing and scheduling engine for general rail networks

A number of tactical and operational applications, such as train timetabling and dispatching, require sophisticated and computationally efficient software to optimize train schedules. An open-source train scheduling package, namely FastTrain has been designed and implemented to generate feasible schedules and optimality gap estimates for trains on a generic rail network with both single and double tracks. This paper describes its two major modelling components: (1) a network cumulative flow model to capture the complex safety rules of infrastructure resources, and (2) a Lagrangian relaxation solution framework which efficiently decomposes the original multi-trains complex model to a set of single train oriented subproblems. Two experimental cases based on the adapted datasets released by INFORMS RAS are conducted to demonstrate the effectiveness and efficiency of the developed algorithm under different network and data availability conditions.