A Hybrid Metaheuristic for the Distance-constrained Capacitated Vehicle Routing Problem

• Published in: Procedia, social and behavioral sciences Vol. 109; pp. 779 - 783
• Main Authors: Tlili, Takwa, Faiz, Sami, Krichen, Saoussen
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2014
• Subjects: integer linear programming; Particle swarm optimization; Variable neighborhood search; Vehicle routing problem; integer linear programming; Variable neighborhood search; Particle swarm optimization; Vehicle routing problem
• ISSN: 1877-0428; 1877-0428
• DOI: 10.1016/j.sbspro.2013.12.543

The delivery of goods is a crucial operational process, which is at the heart of the supply chain and logistic fields. Numerous companies have known a significant drop of their running costs due to the efficient coping with distribution networks. Therefore, several studies in operational research area are interested in modeling and solving transportation problems, especially vehicle routing problems (VRPs). Due to the NP-Hardness of VRPs, various state-of-the-art metaheuristics were developed to generate near-optimal solutions in a reasonable computational time. Against this background, we study in this paper the capacitated vehicle routing problem with distance constraints (DCVRP) consisting in deriving the most favorable vehicle pathways that minimize the vehicles’ traveled distances subject to system requirements. The set of vehicles, based on a central depot, has a maximum weight capacity and can travel up to an allowed maximum distance. To tackle the DCVRP, we formulate it as an integer-programming problem and propose a hybrid swarm-based metaheuristic, named PSO-VNS, which integrates a variable neighborhood search within the particle swarm optimization. Results conducted on benchmark instances show that the proposed PSO-VNS approach is highly competitive, compared to existing solution approaches, and converges to promising solutions.