Two-stage heuristic algorithm for a new variant of the multi-compartment vehicle routing problem with stochastic demands

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Gonçalves-Dosantos, Juan Carlos, Davila-Pena, Laura, Casas-Méndez, Balbina
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912082425806848
author Gonçalves-Dosantos, Juan Carlos
Davila-Pena, Laura
Casas-Méndez, Balbina
author_facet Gonçalves-Dosantos, Juan Carlos
Davila-Pena, Laura
Casas-Méndez, Balbina
contents This paper presents a model for a vehicle routing problem in which customer demands are stochastic and vehicles are divided into compartments. The problem is motivated by the needs of certain agricultural cooperatives that produce various types of livestock food. The vehicles and their compartments have different capacities, and each compartment can only contain one type of feed. Additionally, certain farms can only be accessed by specific vehicles, and there may be urgency constraints. To solve the problem, a two-step heuristic algorithm is proposed. First, a constructive heuristic is applied, followed by an improvement phase based on iterated tabu search. The designed algorithm is tested on several instances, including an analysis of real-world datasets where the results are compared with those provided by the model. Furthermore, multiple benchmark instances are created for this problem and an extensive simulation study is conducted. Results are presented for different model parameters, and it is shown that, despite the problems' complexity, the algorithm performs efficiently. Finally, the proposed heuristic is compared to existing solution algorithms for similar problems using benchmark instances from the literature, achieving competitive results.
format Preprint
id arxiv_https___arxiv_org_abs_2410_17302
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Two-stage heuristic algorithm for a new variant of the multi-compartment vehicle routing problem with stochastic demands
Gonçalves-Dosantos, Juan Carlos
Davila-Pena, Laura
Casas-Méndez, Balbina
Optimization and Control
90B06 (Primary), 90C15 (Secondary) 90C59
This paper presents a model for a vehicle routing problem in which customer demands are stochastic and vehicles are divided into compartments. The problem is motivated by the needs of certain agricultural cooperatives that produce various types of livestock food. The vehicles and their compartments have different capacities, and each compartment can only contain one type of feed. Additionally, certain farms can only be accessed by specific vehicles, and there may be urgency constraints. To solve the problem, a two-step heuristic algorithm is proposed. First, a constructive heuristic is applied, followed by an improvement phase based on iterated tabu search. The designed algorithm is tested on several instances, including an analysis of real-world datasets where the results are compared with those provided by the model. Furthermore, multiple benchmark instances are created for this problem and an extensive simulation study is conducted. Results are presented for different model parameters, and it is shown that, despite the problems' complexity, the algorithm performs efficiently. Finally, the proposed heuristic is compared to existing solution algorithms for similar problems using benchmark instances from the literature, achieving competitive results.
title Two-stage heuristic algorithm for a new variant of the multi-compartment vehicle routing problem with stochastic demands
topic Optimization and Control
90B06 (Primary), 90C15 (Secondary) 90C59
url https://arxiv.org/abs/2410.17302