Application of Expert Knowledge for Train Dispatching Decision Support

The challenge of increasing train velocities, highly interdependent schedules and higher traveller demands force train traffic providers to improve the punctuality and reliability of their train services. The dispatchers, which try to reduce the effects of unavoidable disturbances and to keep the trains close to the schedule, are overcharged by their tasks and need to be supported by advanced support systems, which should be able to develop dispatching actions and to propose them to the dispatcher, together with reasons, explanations and alternatives. In this paper, a new decision support system, which is especially designed for future high speed maglev train networks, is described. The system visualises condensed traffic situation information, for which the traffic development of the next hour is estimated by simulation. The available data is analysed for existing or arising conflicts. Applying the stored dispatching expert knowledge, promising dispatching actions are developed. After having checked the consistency of these actions with various hard constraints, the probable effects of the selected dispatching measures are evaluated by simulation. The action which resulted in the greatest improvement in traffic quality is proposed to the dispatcher. Thus, the system contributes to an improvement in traffic performance, reliability and customer satisfaction. Transactions on Information and Communications Technologies vol 16, © 1996 WIT Press, www.witpress.com, ISSN 1743-3517