Optimasi Rute Pengambilan Sampah Menggunakan Algoritma K-Means Dan Ant Colony Optimization
Abstract
Waste management is a major challenge in urban logistics, particularly in optimizing transportation routes to ensure efficient waste collection without exceeding vehicle capacity. This issue is classified as a Capacitated Vehicle Routing Problem (CVRP), which can be addressed using metaheuristic algorithms such as Ant Colony Optimization (ACO). However, ACO performance is highly dependent on parameter selection, making parameter tuning essential to achieve optimal solutions. Furthermore, dividing transportation zones effectively is crucial for structured and efficient waste collection. This study proposes a Modified Ant Colony Optimization method, which integrates the K-Means algorithm for automatic clustering of waste collection areas. A total of 75 datasets comprising latitude, longitude, and waste volume data were used. The system successfully divided the area into clusters and generated optimal routes within each cluster without exceeding truck capacity. The fitness results showed variations in some clusters, while others remained stable. Parameter tuning experiments using different numbers of ants (10, 20, 30, and 50) demonstrated that the algorithm consistently achieved optimal fitness values with convergence stability. Among these, the 50-ant configuration produced the most stable convergence curve, although with the longest execution time. These results confirm that the proposed Modified ACO method is effective in determining optimal waste collection routes, and that parameter tuning plays a crucial role in enhancing algorithm performance.
