Diseño de un modelo híbrido para el mejoramiento del proceso de ruteo de vehículos en el centro de distribución de Coca Cola FEMSA en Barranquilla

This project proposes a hybrid model that enhances the routing efficiency of Coca Cola FEMSA's distribution center in Barranquilla. The design combines methods such as the sweep algorithm and machine learning-based clustering, along with the application of heuristic algorithms such as the nearest neighbor and local search. The addressed problem is a version of the Vehicle Routing Problem (VRP), where the aim is to define and assign a set of customers distributed in a geographic space to a number of available vehicles, ensuring to serve all customers and visit them exactly once. The main objective is to minimize the total transportation cost while satisfying various logistical constraints. For the problem outlined in this project, results were obtained that demonstrate significant improvements of 19.51% in the number of routes dispatched per day compared to the current plan, providing not only a reduction in operating costs but also a noticeable improvement in responsiveness and customer satisfaction. Additionally, this approach allows for greater flexibility and adaptation to changes in demand and environmental conditions, which is crucial for maintaining efficiency and competitiveness in the current market.

• Autores: Posada Rodríguez, Ángel Alberto; Fernández Rodríguez, Sharon Angelith
• Fecha de publicación: 2024
• Institución: Universidad del Norte
• Repositorio: Repositorio Uninorte
• Idioma: spa
• OAI Identifier: oai:manglar.uninorte.edu.co:10584/12021
• Acceso en línea: http://hdl.handle.net/10584/12021
• Palabra clave: Problema de ruteo de vehículos (VRP), Algoritmo hibrido, Algoritmo de barrido, Machine learning, Vecino más cercano, Mejoras locales.; Vehicle routing problem (VRP), Hybrid algorithm, Sweep algorithm, Machine learning, Nearest neighbor, Local improvements.
• License: Universidad del Norte