Integration of Regression-Based Guidance Ant for Enhanced Exploration and Convergence in Ant Colony Optimization (ACO)

Traditional Ant Colony Optimization (T-ACO) algorithms often face challenges in dynamic environments, particularly a tendency to become trapped in local minima, resulting in suboptimal path planning. To address these limitations, this research incorporates a linear regression line as a directional guide for the ants, helping them navigate toward the optimal path more efficiently. This study presents an improved Ant Colony Optimization (I-ACO) algorithm by integrating regression-based guidance to enhance both exploration and convergence. The proposed I-ACO method was tested using real-world geographic data from the Universitas Sriwijaya campus, and its performance was compared with that of T-ACO in both obstacle-free and obstacle-present scenarios. The results indicate that I-ACO consistently produces more accurate and efficient paths, with lower distance errors and faster computation times. Specifically, I-ACO achieved an average computation time of 0.8 seconds, compared to T-ACO’s 1.2 seconds. Moreover, I-ACO demonstrated superior adaptability when encountering obstacles, providing shorter and more efficient routes. These findings highlight the effectiveness of regression-based guidance in improving the performance of the ACO algorithm, making it more suitable for real-time autonomous vehicle navigation in complex environments. This approach not only enhances path accuracy but also reduces computational costs and improves obstacle avoidance capabilities.