Enhancing Green Vehicle Routing Efficiency Through Machine Learning-Optimized Genetic Algorithms

Abstract

The study addresses the Multiple Depot Vehicle Routing Problem (MDVRPTW) challenge, focusing on optimizing pathways for a fleet of vehicles to distribute goods to customers while adhering to delivery window constraints. The study introduces a new category, Green MDVRPTW, which aims to reduce environmental impact through innovative approaches. By combining machine learning and genetic algorithms, the research optimizes routing decisions to minimize journey costs, fleet size and aggregate delays. The use of clustering techniques to group clients based on depot proximity enhances routing efficiency. The paper employs a Tuned Genetic Algorithm and metaheuristics to tackle the Green MDVRPTW problem effectively. Furthermore, the paper provides a comprehensive overview of MDVRPTW solutions, presents a mathematical model for Green MDVRPTW, and details the implementation of the Tuned Genetic Algorithm and clustering concepts. Computational outcomes are analyzed against benchmark standards. The research contributes valuable insights for sustainable transportation planning and optimization.