Skip to Main content Skip to Navigation
Theses

Développement et implémentation de modèles apprenants pour l’exploitation des grandes gares

Abstract : This thesis deals with uncertainty and robustness in decision problems, with the case of the train platforming problem subject to delays. A two-part methodology is proposed to address this problem. First, delay records are used to build models predicting probability distributions conditionnaly to a set of explanatory variables. A methodology to validate and evaluate these predictions is proposed to ensure their reliability for decision-making. As the train platforming problem can be seen as a weighted clique problem, these predicted distributions are used in a second part to add weights on edges to penalize risk of conflict. Local search algorithms are used and experiments show a significant decrease in conflicts.
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-03267838
Contributor : Abes Star :  Contact
Submitted on : Tuesday, June 22, 2021 - 4:50:54 PM
Last modification on : Thursday, June 24, 2021 - 3:18:24 AM

File

86519_MILLIET_DE_FAVERGES_2020...
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-03267838, version 1

Collections

Citation

Marie Milliet de Faverges. Développement et implémentation de modèles apprenants pour l’exploitation des grandes gares. Mathématiques générales [math.GM]. Conservatoire national des arts et metiers - CNAM, 2020. Français. ⟨NNT : 2020CNAM1283⟩. ⟨tel-03267838⟩

Share

Metrics

Record views

52

Files downloads

30