Advances in Industrial Engineering

Advances in Industrial Engineering

A Hybrid VNS–Hopfield Approach for Capacitated Green Vehicle Routing under a CO₂ Emission Cap

Authors
1 PhD candidate, Department of Industrial Engineering, Faculty of Engineering, Bu-Ali Sina University, Hamedan, Iran
2 Professor, Department of Industrial Engineering, Faculty of Engineering, Bu-Ali Sina University, Hamadan, Iran
10.22059/aie.2026.408744.1965
Abstract
Green vehicle routing seeks to minimize distribution costs while satisfying vehicle-capacity and greenhouse-gas emission constraints. This study develops a hybrid method combining Variable Neighborhood Search (VNS) with a Hopfield neural network for a capacitated vehicle routing problem with an explicit CO₂ emission cap. VNS explores customer assignments to vehicles, while a Hopfield neural network uses an energy-minimization formulation to determine visiting sequences for the resulting vehicle-level traveling-salesperson problems. The method is evaluated on 15 randomly generated instances and compared with standalone VNS and GAMS/CPLEX, with a 1,500-second time limit for the latter. Heuristic results are averaged over 10 independent runs. The hybrid method matches the objective values reported by GAMS/CPLEX on the three smallest instances and matches standalone VNS on the two smallest instances, while obtaining lower objective values than standalone VNS on the remaining 13 instances. It also generates solutions for larger instances where the exact solver reaches its time limit. Although the hybrid method requires longer runtimes than standalone VNS, the results illustrate the potential of combining energy-based neural optimization with metaheuristic search for cost-efficient green vehicle routing under emission constraints.
Keywords
Subjects


Articles in Press, Accepted Manuscript
Available Online from 22 November 2026