ENHANCING SCHOOL BUS ROUTING EFFICIENCY: A STUDY OF CLUSTERING METHODS

School Bus Routing Problem aims to create an efficient routes and allocate serviceable areas for student school buses. Buses pick up students from various locations based on their capacities and closeness or distances. It is a typical clustering problem for different locations. However, the choice of algorithm may vary depending on student locations and area shapes. This study focuses on to compare and evaluate the performance of four well-known clustering methods like K-Means, DBSCAN, Hierarchical Clustering and Gaussian Mixture Model on 500 randomly generated geographical points within the boundaries of Izmir (T¨urkiye). The evaluation is conducted based on their density and distribution characteristics. Silhouette Score, Davies-Bouldin Score metrics, and running time are used to assess clustering quality. This analysis highlights the advantages and disadvantages of the algorithms to provide insights into their applicability in different scenarios. Additionally, a visual representation of clustering outcomes offers a deeper understanding of the spatial distribution of the data