Hybrid Hughes-Hartogs power allocation algorithms for OFDMA systems

This work analyzes the discrete solution of Hughes-Hartogs (HH) for the transmission rate maximization problem with power constraint in the OFDMA systems and explores mechanisms to reduce the computational complexity of greedy algorithms. In addition to the solution characterization, a computational complexity analysis is developed, considering the number of executed operations for running time purpose. Moreover, we have compared the system capacity via the throughput obtained with the HH solution, and its variants combined with three complexity reduction mechanisms. These tools consist of an initial allocation bit vector calculated by rounding the results of the water-filling (WF) solution, the multiple subchannels per iteration updating, and the adoption of a subchannel grouping procedure. Our findings indicate that the update of multiple subchannels and the subcarriers grouping techniques reduce the number of iterations required for convergence of the original HH, with some throughput degradation. Also, the bit-allocation mechanism based on the WF is deployed as an alternative to overcome the HH solution, increasing the computational complexity.