Random generation of periodic event schedules with an application to railway timetabling

In this paper, the authors discuss the problem of randomly generating fixed-interval railway timetables in order to use them as scenarios when drawing up and/or analyzing specifications of future railway infrastructure. The timetables are sampled from a so-called timetable structure, which is a mixed integer system of linear inequalities, the solutions of which represent timetables and the constraints of which model physical or operational requirements. The periodicity of the timetables is expressed by an integral variable for each constraint in the timetable structure. This paper presents a heuristic approach to randomly generate train orders, given the timetable structure, according to two specific distributions.