Optimization of Vehicle Routing Problem with Time Windows (VRPTW) with Hybrid Dragonfly Algorithm Approach on Delivery Routes

Abstract

Efficient product distribution is a critical component of supply chain management, especially for small-scale business that operate under limited vehicle capacity and strict delivery time constraints. This research focused on solving the Vehicle Routing Problem with Time Windows (VRPTW) by applying a hybrid optimization strategy that integrates the Nearest Neighbor (NN) method and Dragonfly Algorithm (DA) to reduce total travel distance while ensuring compliance with capacity and time windows requirements. In that proposed approach, the Nearest Neighbor method is utilized to construct an initial feasible route based on proximity considerations, whereas the Dragonfly Algorithm is employed to enhance the route configuration through balanced exploration and exploitation processes. The effectiveness of the hybrid method is evaluated using real contribution data obtained from HoneyBee Bakery & Cake, a small-cake bakery enterprise located in Palembang, Indonesia. The experimental results indicate that the Nearest Neighbor method generates an initial route with a total distance of 72.54 km. After applying the hybrid NN–DA optimization, the total travel distance is reduced to 62.65 km, achieving a reduction of 9.89 km or an efficiency improvement of 13.64%, without increasing the number of vehicles used. Furthermore, the parameter sensitivity analysis reveals that variations in the number of dragonflies and iterations have a considerable impact on solution quality and convergence behavior. Overall, the findings confirm that the proposed hybrid method offers an effective and practical solution for VRPTW in real-world distribution contexts. Additionally, a web-based application is developed to support route optimization and data processing, enabling easier adoption by non-technical users in small-scale distribution environments.