4 Presentations

• 
  03:30 PM - 03:55 PM

  A stochastic lead times and stochastic demand in a production-distribution problem

  • Dorian Dumez, presenter, HEC Montréal
  • Jean-François Cordeau, HEC Montréal, GERAD, CIRRELT
  • Raf Jans, HEC Montréal

  Plants and distribution centers often deliver their products to numerous customers spread over a vast territory and can thus rely on a combination of air, road, rail, and maritime transportation.
  These means of transportation have different costs but also different lead times, and there is thus a fundamental trade-off to be considered: a shorter lead time typically comes at a higher cost but offers more flexibility to react quickly to changes in demand.
  Hence, the plant faces the complex problem of making simultaneous production and transportation decisions, which include the selection of the transportation modes to use for shipping the goods to different customers.
  The objective is to minimize the expected cost of production, transportation, inventory, and lost sales.
  We extend the production-distribution problem setting with uncertainty on the demand and on the lead times.
  We consider this problem in a setting with a discrete finite time horizon.
  Thus, at the beginning of each time period, the time to ship goods to each customer with each carrier is revealed.
  Then, after the production and shipping decision have been made, the demand at each customer is also revealed.
  Concurrently taking into account two types of uncertainty in a complex system constitute a very difficult stochastic optimization problem.
  Consequently, we develop a heuristic to solve the proposed problem in a rolling horizon framework.
• 
  03:55 PM - 04:20 PM

  Minsum capacity provisioning for evacuation on path networks

  • Oluwaseun Lijoka, presenter, University of Lethbridge
  • Robert Benkoczi, University of Lethbridge

  Dynamic networks and flow together with sink location problems are at the forefront of research for effective evacuation planning, in the advent of large-scale disasters. Many models developed in these areas are motivated by considering densely populated areas, with known road capacities and the possibility of evacuating by foot. However, when considering sparse or remote communities, the evacuation models require some adjustments as effective evacuation may not be achievable with the classical assumption, and emergency vehicles may have to be supplied for the rescue mission. In this work, we present an extension of capacity provisioning for evacuation in dynamic networks with the objective of minimizing the sum of evacuation times on the network. We exploit the connection between a fixed resources and the network capacities and design a new objective function suitable for the minsum problem. We consider a simple path network with a fixed sink and our approach leads to convex objective functions which could change as the optimization process progresses. We exploit the power of sequential quadratic programming and matrix decomposition to handle the changing objective function, to accommodate the associated constraints, and make the optimization process efficient.

  Keywords: Facility location, sink location, capacity allocation, evacuation problem
• 
  04:20 PM - 04:45 PM

  A Two-Stage Production Routing Problem with Uncertain Availability of Vehicles (CANCELED)

  • Alline Zanette, presenter, Polytechnique Montréal
  • Walter Rei, Université du Québec à Montréal
  • Michel Gendreau, Polytechnique Montréal
  • Jorge Mendoza, HEC Montréal

  The Production Routing Problem (PRP) is a complex integrated problem that allows for the achievement of competitive advantage. Most of the literature on PRP considers only deterministic data, and the ones that take stochastic parameters into account focus mainly on uncertain demand. In this study, we consider a PRP with a single capacitated production facility, a single product, and a homogeneous fleet of capacitated vehicles. The availability of these vehicles is unknown, so it is assumed as a stochastic parameter. We propose a two-stage mathematical formulation with two possible recourse policies for cases in which there are not enough vehicles to perform all the planned routes. In the first policy, non-served customers are inserted in existing routes, if the remaining capacities of the latter allow for it; in the second one, they are served by third parties, which are outsourced from the spot market. We present a Benders decomposition approach to solve the proposed model.
• 
  04:45 PM - 05:10 PM

  A Progressive Hedging-based Matheuristic for the Stochastic Production Routing Problem with Adaptive Routing

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

  In order to enhance the efficiency of a supply chain, a number of processes related to production, distribution, and inventory planning must be optimized. Conventionally, these processes are addressed individually without considering their interdependence. However, research suggests that an integrated approach is more cost-effective within the network. Consequently, there has been a growing interest in addressing these subproblems in an integrated manner, commonly referred to as the Production Routing Problem (PRP). In the PRP, a company is required to make simultaneous decisions regarding production (setups and quantities) and delivery routes to multiple customers to fulfill their demand over a multi-period horizon. Uncertainty, particularly demand uncertainty, is a major factor that can have adverse effects on the supply chain, leading to unfavorable outcomes and increased costs. In existing PRP models, customers are visited in a fixed sequence irrespective of demand realization. This reduces flexibility and may significantly increase transportation costs as some customers may be visited unnecessarily. To address this issue, a two-stage stochastic formulation is proposed, where routing plans are adaptive and changed based on demand realization. A three-phase matheuristic algorithm, coupled with the progressive hedging approach, is developed to solve this intricate problem and obtain high-quality solutions.