A genetic algorithm for joint resource allocation in cooperative cognitive radio networks

Existing literature in Cooperative Cognitive Radio Networks (CCRNs) always assumed a scenario where only one Primary User (PU) and several Secondary Users (SUs) coexist. However, in practice, multi-PUs and multi-SUs always coexist and the number of SUs is usually greater than that of PUs. Under such complex yet real scenarios, we assume that each PU not only allows a set of SUs to access its pre-allocated channel, but can leverage some of these SUs to improve its transmission rate via cooperative technologies. We consider a joint channel allocation and cooperation set partition problem in CCRNs, in which we aim to allocate a channel and assign a cooperation set that consists of several SUs for each PU, such that for a given period of time, the average transmission rates gained by all the users achieve maximum proportional fairness. We formulate the problem as a 0-1 non-linear programming model. Due to its NP-hardness, we propose a suboptimal Centralized Genetic Algorithm (CGA) for the problem. Extensive simulations demonstrate that CGA not only converges rapidly, but is shown to perform as well as 92% of the optimal solution delivered by brutal search, in terms of the fitness that reflects the fairness degree of the transmission performance gained by all the users.