Dynamic Multiobjective Approach for Power and Spectrum Allocation in Cognitive Radio Networks

Cognitive radio (CR) is a promising technology for addressing resource scarcity in wireless networks. However, in the current CR-based framework, when resource availability changes, existing resource allocation algorithms restart the optimization process without using historical information. In addition, most previous studies have focused on one or two objectives and some have only addressed pure power control or spectrum allocation, limiting the potential of CR. Thus, this article proposes a dynamic multiobjective approach for power and spectrum allocation in a CR-based environment in which the available spectrum channels vary over time and multiple objectives are involved. This article presents a dynamic multiobjective optimization problem (MOP) with the objectives of energy efficiency, fairness, and spectrum utilization and an approach of Pareto optimality. A dynamic resource allocation algorithm comprising a hybrid initialization method and feasible point generation mechanisms is proposed to solve the dynamic MOP. To dynamically adjust resource allocation, historical approximate Pareto optimal solutions, represented by a center and a manifold, are used to predict the new optima. The proposed approach can yield a better convergence level and convergence rate than those attained by comparable multiobjective resource allocation algorithms. Compared with conventional single-objective approaches, it achieves an excellent balance between the objectives.