Channel Allocation & Power Control for Dynamic Spectrum Cognitive Networks using a Localized Island Genetic Algorithm

This paper formulates a distributed cross layer channel allocation-power control problem for an ad-hoc cognitive radio network in dynamic spectrum access and applies a localized variation of the island Genetic Algorithm (iGA) to solve this problem. For comparison purposes, we also implement the standard iGA, which works with global information, and compare it to the localized iGA to analyze the effect of working with partial knowledge on network performance. Simulation results demonstrate that the localized iGA results in efficient spectrum allocation.