4 Presentations

• 
  10:30 AM - 10:55 AM

  Planification tactique des projets : Intégration du nivellement des ressources au RCCP.

  • Anis Noureddine, presenter, Polytechnique Montréal
  • Francois Soumis, GERAD et Polytechnique
  • Robert Pellerin, École Polytechnique de Montréal

  Ce travail présente un nouveau modèle linéaire mixte en temps continu pour le problème RCCP, adapté à la planification tactique des projets. Utilisé en début de projet, le RCCP fixe les intensités d’exécution des lots de travaux par période. Nous intégrons le nivellement des ressources tout en minimisant la durée totale (makespan), les intensités variables aggravant les fluctuations coûteuses des ressources. Notre approche exacte accélère le meilleur modèle existant, optimisé pour le makespan, afin de générer efficacement le front de Pareto, qui dévoile les compromis entre nivellement et durée lorsqu’on intègre le nivellement des ressources. Nous reformulons le RCCP avec des contraintes de capacité et des ajustements temporels pour lisser l’utilisation des ressources. Nous testons notre méthode sur des instances que nous avons générées et d’autres issues de la littérature. Ce travail propose au gestionnaire un compromis optimal entre gestion des ressources et respect des délais.

• 
  10:55 AM - 11:20 AM

  Strengthening the RCPSP lower bounds.

  • Oussama Siwane, presenter, Polytechnique Montreal, Gerad
  • El Hallaoui Issmail, GERAD & Polymtl
  • Robert Pellerin, École Polytechnique de Montréal

  The Resource-Constrained Project Scheduling Problem (RCPSP) is one of the most challenging optimization problems of the past decades. Its strong symmetry and the exponential number of possible solutions make it particularly difficult to solve using exact methods. When addressing this class of problems, closing the optimality gap requires either improving the lower bound through an exact procedure or finding a better solution using an exact or heuristic method. Consequently, many researchers have shifted their focus toward approximate methods to obtain high-quality heuristic solutions. In this work, we focus on improving the lower bound for the RCPSP. We propose an exact approach based on a decomposition method—Dantzig-Wolfe decomposition—within a column generation framework. Experimental results on the PSPLIB dataset demonstrate that our method achieves competitive convergence times and delivers better lower bound quality compared to existing approaches in the literature.

• 
  11:20 AM - 11:45 AM

  Explaining Unsatisfiable Constraint Problems with Parameter Correction

  • Antoine Laviolette, presenter,
  • Claude-Guy Quimper, Université Laval
  • Raphaël Boudreault, Thales

  Constraint problems can be use as decision support system in real-world application. However, end-users of those systems needs to be able to deal with a erroneous or faulty input data that can make the constraint problem unsatisfiable. Existing research for explaining such unsatisfiability mainly focus on finding a subset of problematic constraints. We propose an approach that can find a minimal correction at a level the end-user can understand. Instead of focusing only on constraints, we focus on finding a subset of parameter correction. This is a subset of data parameters and their correction to make the instance satisfiable. Our proposed approach is constraint agnostic. We empirically evaluate our approach on existing CP benchmark problems that we have modified for unsatisfiability and demonstrate that our approach can correct the unsatisfiability.

• 
  11:45 AM - 12:10 PM

  Optimizing Project Delivery System Selection: Assessing External Environment Impact and Mitigation Strategies in Quebec’s Public Infrastructure Sector

  • Basma Ben Mahmoud, presenter, Département de génie mécanique, Université Laval
  • Lehoux Nadia, Full Professor at Laval university
  • Gabriel Jobidon, ÉTS
  • Pierre Blanchet, Faculté de la Foresterie,géographie et géomatique

  The Project Delivery System (PDS) selection process significantly influences construction project success and team performance. While extensive research has focused on improving PDS selection, the impact of external environment criteria such as economic constraints, political pressure, and market conditions remains largely unexplored. These dynamic factors disrupt project execution and hinder overall sector productivity. Addressing this gap, this study evaluates the influence of external environment criteria on PDS selection and proposes mitigation strategies using a multi-objective optimization model. The proposed model helps in allocating the suitable PDS for each project and provides a strategic scheduling considering project characteristics, owner’s preferences, human resources constraints and external environment criteria including political and economic constraints, contractor availability, and innovation needs. A case study on Quebec’s public infrastructure projects was applied to evaluate how these criteria shape the PDS selection results. Findings demonstrate the critical role of external criteria in promoting alternative and collaborative PDS. Moreover, mitigation measures such as prioritizing politically constrained projects and dividing larger projects into smaller ones to anticipate contractor availability issues prove effective in managing associated risks. This research provides a valuable insight for improving PDS selection and enhancing project outcomes in a dynamic environment.