Real Time Scheduling for Cognitive Radio Networks

Real-time scheduling protocols for cognitive radio (CR) networks that can be implemented in the spectrum manager of IEEE 802.22 are proposed. The protocols take into consideration the inherent uncertainties in CR networks that are caused by the unknown future primary user channel usage and the dynamic channel allocation demands of the CR users in each time slot. Scheduling is performed at the beginning of each time slot. The proposed mean throughput maximization (MTM) scheduling protocol schedules the CRs available in the current time slot considering past channel allocations and possible future demands without introducing the delay of waiting for future CR resource demands to be revealed. A variation of the MTM called throughput loss minimization (TLM) is introduced to further reduce the delay in MTM scheduling. The upper bound on the throughput loss and the relation of TLM to MTM in terms of computational complexity and approximation ratio are analyzed. In addition, Markov chance decision processes are used for CR real-time scheduling. Tradeoffs in the upper hound of the throughput loss allow the cognitive base station to enhance the sum throughput of the CR network to maintain both scheduling time and throughput loss below a desired level.