Variable Neighborhood Search for the Electric Vehicle Routing Problem
David Woller, Viktor Kozák, Miroslav Kulich, Libor Přeučil
TL;DR
This work tackles the Capacitated Green Vehicle Routing Problem (CGVRP), a minimal yet representative EVRP variant, by proposing a Variable Neighborhood Search (VNS) framework that blends VRP-inspired components with EVRP-specific mechanisms. The method employs a three-phase initial solution (Density-Based Clustering + Modified Clarke-Wright Savings + Relaxed ZGA repair), a Randomized RVND-based local search with problem-specific AFS relocation operators (AFS-realloc-1 and AFS-realloc-more), and a generalized Double-Bridge perturbation with Repair. Key contributions include the Robust Relaxed ZGA repair, the CVRP-to-EVRP constructive pipeline, and extensive ablation studies demonstrating the value of each component, culminating in state-of-the-art results on the IEEE WCCI 2020 CGVRP dataset and competitive performance against BACO. The approach yields practical impact by delivering a robust, adaptable solver for EVRP variants, with open data and code to facilitate replication and extension.
Abstract
The Electric Vehicle Routing Problem (EVRP) extends the classical Vehicle Routing Problem (VRP) to reflect the growing use of electric and hybrid vehicles in logistics. Due to the variety of constraints considered in the literature, comparing approaches across different problem variants remains challenging. A minimalistic variant of the EVRP, known as the Capacitated Green Vehicle Routing Problem (CGVRP), was the focus of the CEC-12 competition held during the 2020 IEEE World Congress on Computational Intelligence. This paper presents the competition-winning approach, based on the Variable Neighborhood Search (VNS) metaheuristic. The method achieves the best results on the full competition dataset and also outperforms a more recent algorithm published afterward.