The Anchor-Robust Project Scheduling Problem - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Operations Research Année : 2019

The Anchor-Robust Project Scheduling Problem

Résumé

A project scheduling framework is considered where processing times of jobs lie in some uncertainty set. The decision maker needs to compute a baseline schedule in advance, while guaranteeing that some jobs may not be rescheduled later. A subset of jobs is said to be anchored with respect to a baseline schedule if for any realization of processing times in the uncertainty set, the baseline schedule can be repaired in a second stage without changing the starting times of anchored jobs. Each job has an anchoring weight. The Anchor-Robust Project Scheduling Problem (AnchRobPSP) is to find in a first stage a baseline schedule satisfying a given deadline, and a subset of anchored jobs with maximum total weight. AnchRobPSP is considered for several uncertainty sets, such as box or budgeted uncertainty set, and dedicated graph models are presented. AnchRobPSP is shown to be NP-hard even for budgeted uncertainty. Polynomial or pseudo-polynomial algorithms are devised for box uncertainty and special cases of budgeted uncertainty. Numerical experiments for AnchRobPSP under budgeted uncertainty are presented. AnchRobPSP solutions are compared to those of state-of-the-art robust techniques. Finally it is shown how to achieve a trade-off between the number of anchored jobs and the makespan of the baseline schedule.
Fichier principal
Vignette du fichier
20190531_AnchRobPSP_HAL.pdf (358.11 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02144834 , version 1 (31-05-2019)

Identifiants

Citer

Pascale Bendotti, Philippe Chrétienne, Pierre Fouilhoux, Adèle Pass-Lanneau. The Anchor-Robust Project Scheduling Problem. Operations Research, 2019, ⟨10.1287/opre.2022.2315⟩. ⟨hal-02144834⟩
351 Consultations
378 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More