Distributed Spectrum Leasing via Vertical Cooperation in Cognitive Radio Networks

In hierarchical cognitive radio networks, unlicensed secondary users can increase their achievable rates by assisting licensed primary user transmissions via cooperation. In this paper, a novel approach to maximizing the transmission rates in the secondary network by optimizing the relay selection, the secondary transmit powers, and the cooperative relaying power splitting parameters is proposed. The resulting optimization problem is mixed integer and nonconvex, which makes it NP hard to find the optimal solutions. Therefore, centralized and distributed solution methods to find near-to-optimal solutions of this challenging problem are proposed. The methods are based on iteratively solving the secondary relay selection by a greedy approach, and the optimal power allocation problem by a fixed-point approach together with alternating direction method of multipliers. It is established that both centralized and distributed solution methods always converge. The numerical results illustrate how the performance of the proposed solution methods depend on the primary performance margins, and show that they give a near-to-optimal solution in few iterations.