Skip to Main content Skip to Navigation

Modélisation par contraintes et heuristiques pour l'évaluation de la capacité d'infrastructures ferroviaires

Abstract : In comparison with others transportation system, the railway transport has many advantages in terms of cost and sustainable development. Nonetheless, transportation is a highly competitive field and the benefits of rail are not sufficient to enable it to prevail. Thus, the SNCF (French National Railway Society) must improve the efficiency of rail transport and the quality of services to clients. Punctuality, train frequency and prices are a central concern. A important work resource on planning is necessary. The problem of assessing the capacity of rail infrastructure sets limits on the network and the impact of modification in the infrastructure. In order to provide solutions to the SNCF, this problem has been modelled in a Constraint Satisfaction Problem. The exact resolution of this problem, with the ILOG SOLVER libraries, shows limits, so improvements are proposed for both the model and the resolution. The improvements focus on reducing the number of constraints, the rewriting of the temporal constraints, suppression of symmetries, cuts and an upper bound. An approached resolution algorithm is proposed, it uses the principles of the meta-heuristics Large and Variables Neighbourhood Search, and the Depth-bounded Discrepancy Search algorithms. The different algorithms and each improvement are validated on a set of random instance and on real data from the railway junction of Pierrefite-Gonesse. The results are the best known
Document type :
Complete list of metadatas
Contributor : Marie Zoia <>
Submitted on : Tuesday, November 10, 2020 - 11:45:15 AM
Last modification on : Saturday, November 28, 2020 - 3:21:12 AM


Files produced by the author(s)


  • HAL Id : tel-02997843, version 1



Fabien Degoutin. Modélisation par contraintes et heuristiques pour l'évaluation de la capacité d'infrastructures ferroviaires. Informatique [cs]. Université de Valenciennes et du Hainaut-Cambrésis, UVHC, (France), 2007. Français. ⟨NNT : 2007VALE0042⟩. ⟨tel-02997843⟩



Record views


Files downloads