A parallel branch-and-bound method for cluster analysis

• Published in: Annals of operations research Vol. 90; p. 65
• Main Authors: Iyer, Lakshmi S, Aronson, Jay E
• Format: Journal Article
• Language: English
• Published: New York Springer Nature B.V 01-01-1999
• Subjects: Algorithms; Cluster analysis; Clustering; Communication; Flexible manufacturing systems; Heuristic; Hypothesis testing; Information systems; Integer programming; Linear programming; Mathematical models; Mathematical programming; Methods; Operations research; Optimization; Optimization techniques; Studies; Systems design
• DOI: 10.1023/A:1018925018009
• EISSN: 1572-9338
• ISSN: 0254-5330

Cluster analysis is a generic term coined for procedures that are used objectively to group entities based on their similarities and differences. The primary objective of these procedures is to group n items into K mutually exclusive clusters so that items within each cluster are relatively homogeneous in nature while the clusters themselves are distinct. In this research, we have developed, implemented and tested an asynchronous, dynamic parallel branchandbound algorithm to solve the clustering problem. In the developmental environment, several processes (tasks) work independently on various subproblems generated by the branch-and-bound procedure. This parallel algorithm can solve very large-scale, optimal clustering problems in a reasonable amount of wall-clock time. Linear and superlinear speedups are obtained. Thus, solutions to real-world, complex clustering problems, which could not be solved due to the lack of efficient parallel algorithms, can now be attempted. [PUBLICATION ABSTRACT]