Centralized and Distributed Optimization of Ad-Hoc Cognitive Radio Network

In this paper, we study coexistence and optimization of an ad hoc cognitive radio network (CRN) coexisting with multicell primary radio networks (PRNs). We assume the PRN cells operate in multiple frequency subbands, and the ad hoc CRN can utilize several subchannels in each PRN subband using spectrum underlay, which maintains that the pre-specified PRN signal-to-interference-plus-noise ratio (SINR) outage probability is not violated. To jointly optimize the throughput of the ad hoc secondary user (SU) links, we utilize the Lagrange duality optimization tool and design fast-convergent weighted sum rate (WSR) maximization schemes under important system and quality of service constraints, including the power spectral mask (PSM), the available transmit power of SUs, the maximum-subchannel-rate, and the minimum-rate per SU link. Both continuous rate (C-rate) and discrete rate (D-rate) modulations are considered. Additionally, we design a distributed access duality scheme which uses the mini-slot competition approach and involves only CRN local information exchange for the dual update, and achieves fast and stable convergence. Effects of many system operating parameters are presented via simulation results, which show that the optimal duality scheme can perform substantially better than the suboptimal duality scheme, and that the performance loss of the distributed scheme is small compared to the centralized scheduling.