Multi-attribute assignment of trains to departures in rolling stock management

Abstract

The article describes our findings on the EURO/ROADEF 2014 Challenge problem. Several heuristic solution techniques have been implemented in a prototypical system for rolling stock management. First, the assignment of trains to departures is supported by a multi-attribute priority rule, for which extensive experiments have been conducted. The subsequent scheduling problem is then solved by a heuristic routing and scheduling concept. The feasibility of solutions is ensured by adopting a transaction model known from database programming to the scheduling problem domain. Besides our contributions to the solution of the optimization problem, a decision support system has been build that visualizes the movements of convoys in the network. Moreover, we make the source code of our optimization approach available with this article: doi:10.17632/nc642wfw2k.1.