Dynamic control channel management in presence of spectrum heterogeneity

Numerous architectures, applications, and algorithms and software-defined radios have been proposed for cognitive radio networks. In order to work or to achieve maximum performance, many of the developed systems and control algorithms require the exchange of control and configuration messages among the network nodes, typically on a dedicated control channel. In a practical deployment, however, the presence of such a commonly known, dedicated control channel might not be feasible, and for some scenarios should especially be avoided. Since such a signaling channel will be the critical link for the network's operation, link failure or interference will cause the entire network to loose its ability to adapt. In this paper, we evaluate our previously biologically inspired distributed channel selection algorithm and compare it to an "optimal" centralized algorithm formulated as a standard multi-commodity flow linear program.