Joint scheduling and power allocation in Cognitive Radio systems

This paper investigates a joint spectrum scheduling and power allocation problem in Cognitive Radio (CR) network composed of several secondary (SU) and primary users (PU). The objective is to minimize the overall SU system power consumption (over several orthogonal frequency bands) under minimum Quality of Service (QoS) constraints, maximum peak and average interference to the PUs constraints. The main challenge lies in the non-convexity of the problem because of the discrete spectrum scheduling policies. We overcome this issue by considering a simpler and convex problem based on Lagrangian relaxation. It turns out that, this problem has discrete solutions which are the optimal solutions to our initial non-convex problem. We propose an iterative sub-gradient algorithm to compute the optimal solution, when it exists, assuming that the channels are drawn randomly. In the case of symmetric systems, this algorithm is not suitable. Therefore, we study such a case for which we give an analytical solution. Several numerical results that sustain and give insight into the resolution of this problem are also provided.