Solving Maritime Fuel Distribution Routing Problem Using Cluster-First, Route-Second

Supply Chain Management plays a vital role in sustaining company operations, particularly in distribution. The Inventory Routing Problem is widely used to ensure product availability while minimizing distribution costs. A key variant, the maritime inventory routing problem that involves time windows requirement, is especially relevant to inter-island distribution, such as fuel distribution practice in Indonesia. This study implements a Cluster-First, Route-Second (CFRS) approach with two different clustering methods: K-means and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) methods. The case study examines a fuel distribution problem in the eastern region of Indonesia involving 13 seaports. The model was modified to calculate distribution costs in accordance with the Ministry of Energy and Mineral Resources' regulations, supporting Indonesia's "One Price Policy"for fuel. This study enhances reliability and scalability of maritime distribution systems through a CFRS approach, ensuring consistent delivery despite logistical complexity. Results show that DBSCAN yields better cluster quality, while the K-means method achieves faster computation time and higher vessel utilization. Overall, the CFRS approach, particularly when implemented with the K-means method, provides an effective and efficient solution to the fuel distribution case study.