Efficient Vehicle Routing Problem: A Machine Learning and Evolutionary Computation Approach

Abstract

The Vehicle Routing Problem with Time Windows (VRPTW) is an extension of VRP that introduces time window constraints to the routing optimization process. Scaling Evolutionary Computation algorithms for VRPTW to handle large-scale problems poses significant challenges. Machine Learning assisted Evolutionary Computation strategy have been proposed to enhance optimization algorithms' efficiency and effectiveness. This study proposes a machine-learning model that exploits the graphical nature of VRP to design and improve evolutionary computational methods. The aim is to improve the resilience and efficiency of VRPTW optimization and provide better-quality solutions for practical applications.