Joint Power Allocation and Relay Selection Optimization for Cognitive Radio Networks

This paper investigates the joint power allocation (PA) and relay selection scheme (RS) in two-way relaying cognitive radio networks consisting of multiple user-pairs and multiple relays. In order to reduce the computational complexity for practical scenario, we propose a branch and bound based (BnB-based) power allocation and relay selection scheme and a greedy power allocation and relay selection scheme to maximize the system throughput. The system is assumed under the constraint that the interference power from the secondary nodes in two way relay systems to primary user (PU) shall be less than a predefined interference threshold which can guarantee the normal communication of PU. Numerical simulation results show that the optimal PA and RS scheme has the highest system capacity, however, the greedy PA and RS scheme has the lowest complexity. The proposed BnB-based PA and RS scheme has the better tradeoff of system throughput and complexity than the above two schemes.