Séminaire doctorants - Franco Quezada Valenzuela

Nous aurons l'occasion d'écouter un exposé de Franco Quezada Valenzuela (post-doctorant à l'équipe OC) intitulé Stochastic programming with decision-dependent uncertainty: a probing-enhanced approach. le 13/05 à 14h en salle 2320.

Résumé : 

 In this work, we address the Multi-item Capacitated Lot-Sizing Problem (MCLSP) under decision-dependent uncertainty by proposing a probing-enhanced stochastic programming approach. In our setting, the decision-maker can proactively acquire information about uncertain demand through probing actions, refining the probability distribution knowledge at a cost. We introduce a compact reformulation that eliminates traditional non-anticipativity constraints, resulting in a stronger linear relaxation and improved computational tractability. To efficiently solve the problem, we develop a relaxation-based approach that decomposes the model into single-item subproblems, which provides tight bounds on the optimal value and enables the construction of high-quality feasible solutions with significantly reduced computational effort. Additionally, we introduce a branch-and-cut framework strengthened by valid inequalities, incorporating both classical demand-covering cuts and a novel family of value-function-based inequalities that exploit the cost structure. Computational experiments demonstrate the effectiveness of the proposed methodology in enhancing both solution quality and reducing runtime while underscoring the value of incorporating probing decisions in lot-sizing problems under uncertainty.