Heuristic Approaches to Solve the Capacitated Facility Location Problem in Terms of Clustering Internal Evaluation

The challenging problem is always happened when dealing with the capacitated facility location problem (CFLP), mostly because of the difference between the capacity of facilities and the number of customers that needs to be served. A facility with small capacity may result in uncovered customers. These customers need to be re-allocated to another facility that still has available capacity. An existing heuristic approach to solve CFLP by Liao and Guo use k -means clustering algorithm as a base idea to decide the facilities location and customers allocation. This existing approach is yet to be evaluated against connectedness, compactness and separations factors for clustering evaluation. In this research, a new approach is proposed to solve CFLP considering those three factors. This approach use k- Nearest Neighbors as a base idea. This new approach is compared to the existing approach using three clustering evaluation methods with connectedness, compactness, and separations factors.