Mahrus Sholeh, Gracyella Exaudi Girsang, R. Billiyan Mulkan Ghifari, Harris Imam Fathoni, Herman Tolle, Aryo Pinandito
—Conventional Ant Colony Optimization (ACO) algorithms, when applied to complex logistics routing problems, often suffer from premature convergence, leading to solution stagnation and suboptimal routing results. This paper introduces an improved ACO algorithm that overcomes this limitation through two novel and synergistic mechanisms: a dynamic ant population and a boosted pheromone update strategy. The dynamic population mechanism improves global exploration by adaptively scaling the number of search agents to the complexity of the problem, while the boosted pheromone mechanism refines exploitation by allowing multiple high-quality solutions to guide the search, thus preventing premature commitment to local optima. Empirical results demonstrate that this enhanced approach produces high-quality routes while achieving a transformative improvement in computational efficiency, reducing execution times by over 98% (from minutes to seconds) compared to conventional ACO without a significant increase in memory consumption. This leap in performance renders the algorithm a practically viable tool for real-time route optimization on mobile-based delivery platforms, enabling its deployment in dynamic, real-world logistics systems. © 2025 IEEE.
Department of Informatics Engineering, Universitas Brawijaya, Malang, Indonesia; Media, Game, and Mobile Laboratory, Department of Information Systems, Faculty of Computer Science, Universitas Brawijaya, Malang, Indonesia
