Cross-Layer Scheduling for Cooperative Multi-Hop Cognitive Radio Networks
The paper aims to design cross-layer optimal scheduling algorithms for cooperative multi-hop Cognitive Radio Networks (CRNs), where secondary users (SUs) assist primary user (PU)'s multi-hop transmissions and in return gain authorization to access a share of the spectrum. We build two models for two different types of PUs, corresponding to elastic and inelastic service classes. For CRNs with elastic service, the PU maximizes its throughput while assigning a time-share of the channel to SUs proportional to SUs' assistance. For the inelastic case, the PU is guaranteed a minimum utility. The proposed algorithm for elastic PU model can achieve arbitrarily close to the optimal PU throughput, while the proposed algorithm for inelastic PU model can achieve arbitrarily close to the optimal SU utility. Both algorithms provide deterministic upper-bounds for PU queue backlogs. In addition, we show a tradeoff between throughput/utility and PU's average end-to-end delay upper-bounds for both algorithms. Furthermore, the algorithms work in both backlogged as well as arbitrary arrival rate systems.