Subchannel-Sharing Based Distributed Optimization of Ad-Hoc Cognitive Radio Network

In this paper, we study the optimization of an ad-hoc cognitive radio network (CRN) coexisting with a multi-cell primary radio network (PRN) utilizing spectrum underlay. To maximize the weighted sum rate (WSR) of the CRN, in contrast to the exclusive subchannel assignment (ESA) method considered in former results, we design shared subchannel assignment (SSA) method to approach the performance limit assuming discrete-rate modulation. The considered SSA method involves interference-channel-sharing based optimization, and in cognitive radio it becomes more complicated when the CRN-to-PRN sum-interference constraint has to be strictly satisfied. We design fast-convergent SSA duality schemes and use the interior point search to satisfy the various system constraints. Additionally, we design distributed duality schemes for both SSA and ESA schemes which involve only CRN local information exchange, using multichannel parallel dual update and a novel mini-slot competition. Effects of many system parameters are presented via simulation results, which show that our SSA duality scheme can perform significantly better than the near-optimal ESA duality scheme, and that the distributed schemes entail only small overhead and convergence losses.