Towards a fair non-convex resource allocation in Wireless Networks

This paper presents a non-convex optimization framework for the Network Utility Maximization problem in Wireless Networks, which incorporates the interference among links and introduces a power penalty term in the objective function to assure both convergence and energy efficiency of the method. Moreover, a distributed gradient based algorithm is proposed that converges to the optimal solution for problems with zero duality gap and a fair-allocation heuristic is presented to resolve user oscillations when they occur. Finally, numerical results regarding the performance of the heuristic and the distributed approach are presented.