Railway planning: optimization approaches for the rolling stock rotation problem

11/22/2023

Rolling stock is the set of train units and is one of the most expensive assets that a railway company can own. The rolling stock rotation planning aims to schedule the rotation of the rolling stocks at minimal costs and is a very complex and time-consuming problem when properly integrated. The rotation must ensure that it covers all trips already defined in the timetable, and respects the requirements imposed by the railway network. This work integrates a significant number of industrial railway requirements such as vehicle compositions, station capacity, regularity bonification, deadhead trip penalization, shunting, and turn operations. We solved the problem through the hypergraph model methodology, and our contribution is the addition of a new objective function, implying the model to be considered as bi-objective. The new objective function aims to minimize the damaging emissions produced within the operations concerning the rolling stock rotation. To solve the model, we used the augmented ε-constraint method. We showed that the complexity of our model is mainly due to shunting operations hyperarcs and that by varying the regularity bonification, we would change the solution routes while the variation of the deadhead trip penalty would not change as much. Moreover, when considering the minimization of the emissions, the solutions tend to have fewer deadhead trips than the solutions provided by the cost minimization. Lastly, to validate and test the model, we used the Chicago railway subway’s open data.

Authors in the community:

• 

  M

  Mickaël Anthony da Silva

  ist193671

Supervisors of this institution:

• 

  D

  Daniel Rebelo dos Santos

  ist167904

other-engineering-and-technologies - Other engineering and technologies

eng - English

Instituto Superior Técnico