A Survey of Clustering Algorithms for Determining Optimal Locations of Distributed Centers

The use of efficient machines and algorithms in planning, distribution, and optimization methods is of paramount importance, especially when it comes to supporting the rapid development of technology. Cluster analysis is an unsupervised machine learning function for clustering objects based on some similarity measure. In this paper, we review different types of clustering algorithms for clustering data of different sizes and their applications. This survey reviews five primary clustering approaches—Partitioning, Hierarchical, Density-Based, Model-Based, and Grid-Based clustering—highlighting their strengths, limitations, and suitability for location-based optimization. Each algorithm is evaluated on key performance criteria, including noise handling, computational efficiency, scalability, and the ability to manage spatial constraints. Key evaluations demonstrate that DBSCAN achieved an average silhouette score of 0.76, indicating strong cluster cohesion and separation, while K-Means showed the fastest computational time for datasets under 10,000 points. The Grid-Based method excelled in scalability, handling datasets exceeding 1 million points with minimal computational overhead. Case studies and real-world applications demonstrate the practical utility of these algorithms in optimizing center placement across diverse industries. The results provide valuable insights for practitioners and researchers seeking to improve distributed network design, resource efficiency, and location optimization using advanced clustering methodologies.