CONICET | Buscador de Institutos y Recursos Humanos

This work deals with a general distribution logistics problem involving a set of facilities (production plants, suppliers, warehouses) from which multiple products are to be efficiently delivered to a high number of consumer nodes (retailers, stores, customers) in order to meet some specified product demands while satisfying capacity and timing constraints. In contrast to the classical pick-up and delivery problem (PDP) that assumes a set of transportation requests with predefined pickup and delivery locations, a more realistic problem definition that better fits many real-world supply chain networks has been adopted. The proposed supply chain-oriented PDP (SC-PDP) defines the choice of source (pickup) nodes for each product demand as a set of additional decisions to be made based on the geographical locations of product demands and product inventories, respectively. Due to its inherent high complexity and combinatorial nature, most of the existing solution approaches to this logistic problem are of heuristic type and, therefore, provide good but not necessarily optimal solutions. This work introduces an exact MILP model that can be solved using a branch-and-cut commercial package to find the best vehicle routes and schedules for moderate-size problems. The higher flexibility in the vehicle routing design process permits to discover more efficient operational strategies to manage complex multi-site distribution systems. The mathematical formulation has been generalized to consider time windows at customer sites, multiple vehicle visits to pickup and delivery nodes, a heterogeneous fleet of vehicles with given weight and volume capacities, multiple depots and the shipment of multicommodities by every vehicle. Optimal solutions for a variety of SC-PDP are reported and discussed.