An Analytical Comparison of Models for Rolling Stock Scheduling
A major step in the planning process of passenger railway operators is that of assigning rolling stock, i.e., train units, to the trips in the timetable. A rolling stock assignment should preferably satisfy the passenger demand, but also be attractive from the perspective of the railway operator. To support railway companies in scheduling their rolling stock, the literature has proposed a wide variety of models. These vary as a result of operational differences, and hence different requirements, between the considered railway companies. In this talk, we categorize the existing models according to these requirements and show the relations between them. Moreover, we make an analytical comparison between two models that have been proposed for the setting of Netherlands Railways (NS) and DB Fernverkehr AG (DB), respectively. Our analysis shows that these formulations lead, under some assumptions, to the same linear programming relaxation bound when considering the rolling stock scheduling setting of NS. Moreover, a numerical comparison shows that the model proposed for NS is able to find optimal solutions in shorter running times in this setting.
Meeting ID: 996 9625 8271