4 présentations

• 
  15h30 - 15h55

  Neighborhood and Hybrid Genetic Search for Inventory Routing Problems

  • Jingyi Zhao, prés., Polytechnique Montréal

  In this study, we address the Multi-Vehicle Inventory Routing Problem (MIRP), focusing on optimizing both inventory management and routing decisions for distributing products from a supplier to retailers. This involves integrated decisions on (i) routing, and (ii) delivery days and quantities subject to inventory constraints.
  Despite the importance of Large Neighborhood Search (LNS) in advancing solutions for vehicle routing problems, its exploration in MIRP has been limited. Traditional neighborhood strategies often become inefficient or overly complex when applied to MIRP, indicating a significant gap in optimization techniques and the need for innovative solutions.
  Therefore, we introduced a novel LNS operator, the PI operator, to tackle this complex issue.
  This PI operator is specifically designed for the MIRP and enables the reselection of optimal distribution plan for each retailer in the current solution, including the insertion of vehicle routes and delivery quantities.
  It iteratively re-optimizes each retailer in a random order until no further improvements can be made. This operator relies on efficient preprocessing, DP techniques, and acceleration tricks, and is integrated into a hybrid genetic search.
  Our method significantly enhances the heuristic's ability to navigate and assess large neighborhoods efficiently, leading to effective solutions. Even in newer large-scale instances, it can produce high-quality solutions that were previously unattainable.
  Furthermore, it addresses MIRP with out-of-stock penalties successfully, solving this problem with excellent results for the first time. Extensive experiments on MIRP confirm the superiority of our approach, demonstrating a remarkable number of new benchmark solutions across diverse datasets, marking a notable advancement in the field.

• 
  15h55 - 16h20

  Integrated poultry production-distribution optimization

  • Samuel Gbeya, prés.,
  • Maryam Darvish, Université Laval
  • Jacques Renaud, Université Laval, CIRRELT

  This study proposes a mixed integer linear programming model to develop production planning and delivery schedules for the poultry supply chain. The production planning comprises the number and breed of chicks to be raised on each farm, when to start raising and the duration of the raising period. The delivery schedule focuses on when and on the number of chickens to be delivered to each slaughterhouse at each period regarding the transportation distance, which is crucial for poultry welfare. This study aims to develop proper production and delivery schedules that reduce the total costs over a planning period. To achieve this objective, we have developed a two-stage algorithm that minimizes the transportation costs and penalty costs for not reaching the target weight in the first stage and the penalty costs for not fulfilling the slaughterhouse’s demands in the second phase. A case study of a Quebec poultry cooperative is used to illustrate the application of the proposed model.

• 
  16h20 - 16h45

  Investigating different service levels in the context of the stochastic production routing problem with adaptive routing

  • Ali Kermani, prés., HEC Montréal
  • Jean-François Cordeau, HEC Montréal, GERAD, CIRRELT
  • Jans Raf, HEC Montréal

  In this research we study the Production Routing Problem (PRP) with demand uncertainty and adaptive routing, integrating service level constraints. Considering service levels can help companies manage stockouts and improve customer satisfaction, yet this issue has not been studied in the context of PRP. To address this gap, we propose a two-stage stochastic model with adaptive routing. This model incorporates chance-constrained formulations for four service level types, allowing for flexible routing based on realized demand scenarios. Given the complexity of the problem, especially with routing decisions handled in the second stage, we develop an efficient iterative matheuristic algorithm. This algorithm decomposes the problem into subproblems for setup decisions, service level considerations, and feasible routing with production and inventory flexibility. We then iterate over a modified version of these subproblems to refine the solution. Finally, we test the proposed algorithm on benchmark instances and leverage the findings to provide valuable managerial insights. This research contributes to the literature by presenting a novel model for the PRP with adaptive routing and service levels, and offers an efficient solution algorithm to navigate the complexities of this integrated problem. We also compare these four service levels in terms of their costs and impacts on the problem.

• 
  16h45 - 17h10

  Multi-Depots Inventory Routing Problem with Heterogeneous Vehicles: Delivery and Backhauling

  • Farzad Avishan, prés., HEC Montreal
  • Amira Dems, Institut de recherche d’Hydro-Québec
  • Yossiri Adulyasak, HEC Montréal
  • Okan Arslan, GERAD, HEC Montréal
  • Jean-François Cordeau, HEC Montréal, GERAD, CIRRELT

  We investigate an inventory routing problem, encompassing the efficient distribution of commodities to sites and the backhauling of hazardous materials from these sites back to the depots. We aim to optimize the distribution of multiple products from different depots to the customers, utilizing capacitated heterogeneous vehicles within discrete and finite-time horizons. Notably, a vehicle cannot transport regular delivery commodities and hazardous materials simultaneously. Moreover, we explore the concept of split deliveries, wherein a single customer may receive commodities by multiple vehicles. We present a mathematical model and incorporate valid inequalities and cuts through a branch-and-cut algorithm. Our approach effectively addresses instances with approximately 40-60 satellite points, 30-40 products, and 4-5 periods within an hour. In cases where the problem exceeds manageable sizes, we have developed a heuristic algorithm as an alternative approach.