QUEZADA Franco

PhD graduated
Team : RO
    Sorbonne Université - LIP6
    Boîte courrier 169
    Couloir 26-00, Étage 4, Bureau 440
    4 place Jussieu
    75252 PARIS CEDEX 05
    FRANCE

Tel: +33 1 44 27 88 37, Franco.Quezada (at) nulllip6.fr
https://lip6.fr/Franco.Quezada

Supervision : Safia KEDAD-SIDHOUM

Co-supervision : GICQUEL CĂ©line

Génération des coupes et approches basées sur la décomposition pour résoudre des problèmes lot-sizing stochastiques à plusieurs étapes

This thesis studies one of the many problems to be addressed by industrial companies when managing their remanufacturing operations: production planning decision under uncertainty. Within a remanufacturing context, production planning consists in deciding about the products to be disassembled, refurbished and reassembled, the timing and level of production as well as the resources to be used so as to meet the customers’ demand for the remanufactured products in the most efficient and economical possible way. We consider a multi-stage decision process corresponding to the case where the value of the uncertain parameters unfolds following a discrete-time stochastic process and the production decisions can be made progressively as more and more information on the demand and cost realizations are collected. The problem can be formulated as a mixed-integer linear program (MILP), however, its direct resolution by a mathematical programming solver poses somes computational difficulties in practice. Our main objective in this thesis is thus to develop mathematical models and algorithms which could ultimately form the basis of decision support tools enabling industrial managers to efficiently plan production activities for complex remanufacturing production systems. We first propose several valid inequalities that can be used to effectively solve the medium-sized problems through a branch-and-cut algorithm. Then, we investigate a decomposition approach for solving (very) large-sized problems based on the stochastic dual dynamic integer programming (SDDiP) approach. For this novel approach, a new dynamic programming formulation and cuts generation strategy are proposed.

Defence : 10/28/2021 - 10h - Conservatoire National des Arts et Métiers, Amphithéâtre Gaston Planté, Accès 35 – 1er étage

Jury members :

Kerem Akartunali, Professeur, University of Strathclyde [Rapporteur]
Stéphane Dauzère-Pérès, Professeur, Ecole des Mines de Saint-Étienne [Rapporteur]
François Clautiaux, Professeur, Université de Bordeaux
Céline Gicquel, MCF, Université Paris-Sud
Safia Kedad-Sidhoum, Professeure, Conservatoire National des Arts et Métiers
Vincent Leclere, MCF, Ecole des Ponts ParisTech
Hande Yaman, Professeure, Katholieke Universiteit Leuven

2018-2021 Publications