— To scale for large ad hoc networks containing hundreds or even thousands of subnets, the networks must be hierarchically organized by partitioning them into clusters or domains. Many researchers have proposed partitioning of large networks into clusters, but this requires specifying cluster size bound. Graph Partitioning (GP) algorithms search for min-cut that balances the number of nodes in each partition. The limitations of GP algorithms are that they cannot be optimized for end-to-end performance requirements or take into account the characteristics of network topology. We propose a clustering algorithm based on the concept of Multiobjective Particle Swarm Optimization (MOPSO). Our algorithm minimizes the sum of the total hop count in each cluster and minimizes the edge-cut weight between clusters. Further, our algorithm provides quality of service (QoS) by taking into account end-to-end performance requirements. We find that minimizing the sum of the cluster hop count results in a balanced and bounded cluster size. For a given number of clusters, our clustering algorithm creates clusters with sizes closer to average size and avoids isolated nodes. In our study, the terms domain, cluster, and partition are used interchangeably.
— Ad-hoc network clustering, QoS, bounded cluster size, hop count, multiobjective optimization.
The authors are with Computer Science Department of Western Michigan University, USA (e-mail: firstname.lastname@example.org, email@example.com).
Cite: Nasser M. Alsaedi and Dionysios Kountanis, " Large Wireless Ad Hoc Network Clustering with End-to-End QoS Constraints Using Multiobjective Particle Swarm Optimization," International Journal of Information and Electronics Engineering vol. 4, no. 4, pp. 288-292, 2014.