Power optimization for dynamic spectrum access with convex optimization and intelligent algorithm

In dynamic spectrum access networks, relay transmission has been considered as a potential method to help cognitive users communicate by reducing their interferences to primary networks especially in the underlay mode. In this paper, we propose a power optimization algorithm for both regenerative and non-regenerative relay systems in condition of Rayleigh fading channels. Based on analyzing the features of dynamic spectrum access networks, a relevant interference model is first built. Then, we propose a combined power allocation strategy in order to minimize the outage probabilities in the cooperative transmission. For regenerative system, we give a closed-form expression for the power allocation by taking into account the characteristics of the fading channels. For non-regenerative system, we utilize pattern search algorithm to solve the optimization problem since the objective function is complex and uneasy to be figured out directly. A competitive strategy for initial point selection is also designed to guarantee a global optimal outcome for the proposed pattern search algorithm. Numerical results show that the system performances with optimum power allocation outperform those with uniform power allocation whereas lower outage probabilities can be obtained.