Energy-Efficient 2-D Resource Allocation with Fairness Constraints for OFDMA Networks

Abstract

In this paper, a 2-Dimension (2-D) resource allocation problem for energy efficiency maximization in the uplink of an OFDMA network is studied. Firstly, with the metric of energy efficiency, measured as bits-per-joule, we convert the complex problem of optimizing the system energy efficiency, defined as sum of users' bits-per-joule, into an optimal achievable rate comparison problem, and devise a low-complexity sub-optimal resource allocation scheme for maximizing it while satisfying each user's resource requirement. The proposed scheme implements both power allocation and Resource Unit (RU) assignment devised by exploiting the optimal achievable rate maximizing user's energy efficiency. The numerical result shows that the proposed scheme outperforms widely-used schemes with Equal Power Allocation (EPA), such as maximum SNR and Proportional Fair, in terms of performance of bits-per-joule with different resource requirement and can achieve similar results as the exhaustive search scheme in much lower complexity.

Energy efficient network-flow-based algorithm for multiuser multicarrier systems

“…The algorithm stops when it reaches a local optimum, at which none of the neighbouring solutions leads to any improvement. Many other two-stage algorithms have been proposed recently, including sorting-and-assignment algorithms [11,12], subcarrier-and-power allocation algorithms [13] and userand-rate allocation algorithms [14]. The aforementioned heuristic algorithms, however, achieve loose gaps from the optimum, or their complexities are hard to predict as the number of iterations needed is a tuning parameter.…”