Application of clustering methods in solving optimal set partitioning problems

The objective of this work is to study and implement an approach for initializing centersbefore executing the r-algorithm, based on clustering results, for solving optimal set parti-tioning (OSP) problems. The study focuses on single-product unconstrained OSP tasks andconsiders clustering algorithms such as EM, K-Means, and hierarchical clustering withvarious linkage types. The research results show a reduction in the number of r-algorithmiterations and the ability to achieve solutions with lower computational effort.