Reliability of Weight Metric based Clustering Models in Large Scale Mobile Ad Hoc Networks

T.M. Rasheed, L. Reynaud, Y. Gourhant, and K. Al Agha (France)


Wireless, Ad Hoc, Clusters, Routing, Algorithms and Performance Analysis.


Clustering is an efficient method for partitioning an ad hoc network by dynamically grouping the geographically adjacent nodes into clusters. Clustering provides for efficient radio resources allocation, location and energy management, and routing and backbone formation and is an effective method for tackling scalability issues pertinent to the Mobile Ad Hoc Networks (MANET) networking domain. In this paper, we consider the weight based clustering approach and analyse the reliability of the weight metric estimation schemes. We then propose an adaptive and reliable weight metric estimation approach at mobile nodes which considers the cost incurred by different device properties in clustering. On the basis of the estimated weight metric, we present the Stable Clustering Algorithm, a robust, adaptive and constraint aware heuristic clustering algorithm for heterogeneous wireless networks applicable to any level of network scalability and density. The proposed algorithm is evaluated using simulations, and the result indicates that our protocol is more reliable and form stable clusters in dynamic large scale mobile ad hoc networking environments.

Important Links:

Go Back