A novel harmony search based spectrum allocation technique for cognitive radio networks

This paper outlines the application of the heuristic Harmony Search (HS) algorithm for efficient spectrum allocation in cognitive radio networks under a minimum Bit Error Rate (BER) criterion. Our proposed algorithm provides a higher degree of diversity in the search process by virtue of its particular improvisation procedure, as opposed to evolutionary computation techniques used so far for this optimization problem. In our work both centralized and distributed implementations of our approach are proposed and detailed. The first set of simulation results made for one single HS instance running over a fixed network show, on one hand, that our approach achieves nearoptimum spectral channel assignments at a very low computational complexity. On the other hand, satisfactory results obtained for a distributed implementation of our algorithm pave the way for future research aimed at comparing our approach with avantgarde genetically-inspired spectrum allocation techniques.