Cluster Modelling for Cognitive Radio Ad-hoc Networks Using Graph Theory

With the swift expansion of wireless technologies, demand for radio spectrum is continuously mounting. Along with the spectrum scarcity problem, radio spectrums are also underutilized. Cognitive radio practices an open spectrum allocation technique, which can ensure efficient handling of the frequency bands. Suitable network model is a must for the implementation of cognitive radio networks. In this paper, an efficient cluster model for cognitive radio ad-hoc networks is presented using graph theory. The proposed clustering model is defined as a maximum edge biclique problem, where the spatial variations of spectrum availability are considered. This clustering scheme aims to maintain set of free common channels in every cluster, which allows smooth shifting among control channels. A parameter called Cluster Head Determination Factor (CHDF) is also introduced to select cluster-heads where clusters' operations are coordinated by cluster-heads. Each cluster comprises of a secondary cluster-head to combat the re-clustering issue for mobile nodes. Finally, simulation and comparative studies are conducted to evaluate the performance of the proposed method.