Within the ACEM-Rail project of the European Seventh Framework Programme new measurement and inspection techniques for monitoring the track condition are developed. By means of these new techniques the prediction of future track condition will be highly improved. To our knowledge mid-term maintenance planning is done for projects and preventive tasks, but predictions of the track condition are not incorporated into the planning process up to now. To efficiently utilize this new kind of information one task within the ACEM-Rail project is the development of methods for planning predictive maintenance tasks along with preventive and corrective ones in a mid-term planning horizon. The scope of the mid-term or tactical maintenance planning is the selection and combination of tasks and the allocation of tasks to time intervals where they will be executed. Thereby a coarse maintenance plan is determined that defines which tasks are combined together to form greater tasks as well as the time intervals for executing the selected tasks. This tactical plan serves as the base for booking future track possessions and for scheduling the maintenance tasks in detail.

In this paper an algorithmic approach is presented which is able to react on dynamic and uncertain changes due to any track prediction updating. To this end optimisation algorithms are implemented within a rolling planning process, so it is possible to respond to updated information on track condition by adapting the tactical plan. A novel optimisation method is developed to generate cost-effective and robust solutions by looking ahead in the future and evaluating different solutions in several scenarios.

railway maintenance, tactical planning, optimisation under uncertainties