Self-Organizing Dynamic Spectrum Management for Cognitive Radio Networks

Abstract

Dynamic spectrum management (DSM) is one of the key problems in the design of cognitive radio (CR) networks. It is a time-varying Dynamic spectrum management (DSM) is one of the key problems in the design of cognitive radio(CR) networks. It is a time-varying and location-dependent optimization problem, equivalent to the well-known graph-colouring problem in graph theory. This problem is known to be NP-hard and computationally challenging to solve. Accordingly, finding the exact solution for the DSM optimization problem is typically not practical. In this paper, we introduce a novel self-organizing DSM scheme, which solves the DSM problem in a decentralized manner. The use of self-organization to address the DSM problem offers several benefits: decentralization and scalability ofthe network behaviour, computational simplicity, cost-effectiveness and bandwidth conservation. In the paper, we address the underlying principles involved in the design and implementation of the self-organizing DSM as well as a software testbed for demonstrating this novel approach. Experimental results are presented to justify this new approach.and location-dependent optimization problem, equivalent to the well-known graph-colouring problem in graph theory. This problem is known to be NP-hard and computationally challenging to solve. Accordingly, finding the exact solution for the DSM optimization problem is typically not practical. In this paper, we introduce a novel self-organizing DSM scheme, which solves the DSM problem in a decentralized manner. The use of self-organization to address the DSM problem offers several benefits: decentralization and scalability ofthe network behaviour, computational simplicity, cost-effectiveness and bandwidth conservation. In the paper, we address the underlying principles involved in the design and implementation of the self-organizing DSM as well as a software testbed for demonstrating this novel approach. Experimental results are presented to justify this new approach.