A Comparative Analysis of Heuristic and Dynamic Algorithms for Route Optimization in Johor’s Delivery Hubs

Abstract

Delivery route optimization is crucial for enhancing logistics efficiency and reducing operational costs in the e-commerce industry. During the COVID-19 pandemic, Movement Control Order (MCO) in Malaysia led to a surge in online shopping as physical stores were closed. This study focuses on optimizing delivery routes between J&T hubs in Johor using three algorithms: Dynamic Programming (DP), Genetic Algorithm (GA) and Simulated Annealing (SA). The objectives include employing these algorithms to determine optimal routes, considering both distance and time and comparing GA and SA against DP as a benchmark. Data from 18 delivery hubs were analyzed using Python, with distance and travel times from Google Maps. All three optimazation methods were successfully applied to determine the optimal delivery route. The results demonstrated that DP consistently provides optimal solutions and emerged as the most effective method. The ideal departure time for both weekdays and weekends was identified as 10 p.m., with 667 minutes for weekdays and 641 minutes for weekends, respectively.In the comparison between GA and SA, GA outperformed SA in 8 out of 9 cases. However, at 6 p.m. on a weekend, SA achieved a shorter duration of 720 minutes compared to GA’s 742 minutes. These findings suggest that GA could be effectively adopted by logistics companies to optimize operations, reduce delivery times and meet the growing demands of e-commerce. Future applications could involve integrating real-time traffic data to further refine route optimization in dynamic environments. Additionally, hybrid approaches combining the strengths of DP, GA and SA could be explored to address complex logistics challenges in various regions, contributing to optimized delivery systems for congested urban areas, faster deliveries, and reduced the environmental impact.