Revenue-based admission control for cognitive radio cellular systems

In a cognitive radio cellular system, secondary users (SUs) are allowed to share the spectrum with primary users (PUs), and inevitably cause interference to PUs. To guarantee the interference to PUs under a predefined threshold, admission control has to be investigated. In addition, different SUs may require different levels of quality of service (QoS), and pay differently based on their services. The network operator cannot provide complete fairness between different SUs. The problem of our interest is to maximize the total revenue output subjected to the interference constraint on the PU, and QoS (in terms of packet error rate (PER)) requirements on both SUs and the PU. In this paper, we transform the admission control problem into a 0-1 knapsack problem, and propose two schemes: AC-Greedy and AC-Dynamic. Simulation results are presented to demonstrate the performance of the proposed schemes.