Aryo Pinandito, Agi Putra Kharisma, Herman Tolle, Chandrawati Putri Wulandari
With the exponential growth in package delivery demands, optimizing delivery routes has become critical for maintaining customer satisfaction and operational efficiency. Ant Colony Optimization (ACO) is an algorithm that uses a probabilistic approach to optimizing and solving delivery route problems. ACO optimizes a path by mimicking the behavior of real ants that follow the pheromone trail left by other ants. How ACO solves the problem greatly depends on how the pheromone level builds up on the graph during optimization. This study investigates how heuristic pheromone initialization contributes to Ant Colony Optimization (ACO) algorithm performance for solving parcel delivery problems, which resembles the Traveling Salesperson Problem (TSP). Instead of initializing the pheromone level to a small uniform value, the study suggests a reverse distance value that allows shorter paths to have a greater probability of being chosen as the next destination in the early optimization phase. To see how the algorithm behaves when the pheromone levels differ in the graph, an experiment with a real-world case study was carried out. The findings suggested empirical evidence that heuristic pheromone initialization in ACO could yield a better optimized solution but marginal impact to user experience. However, significant improvement in route optimization could translate to significant operational efficiency gains, contributing valuable insights for mobile application developers implementing ACO algorithms in the domains of logistics and transportation. © 2025 IEEE.
Media, Game, and Mobile Laboratory, Faculty of Computer Science, Universitas Brawijaya, Malang, Indonesia; Department of Industrial Engineering, Faculty of Advanced Technology and Multidiscipline, Airlangga University, Surabaya, Indonesia
