Rate-Maximization Scheduling Schemes for Uplink OFDMA

Abstract

In this paper, we propose and study several sum rate maximization algorithms for uplink orthogonal frequency division multiple access (OFDMA). For uplink scheduling without fairness consideration, we propose two Lagrangian duality optimization-based methods to maximize the weighted sum rate, which include a cyclic dual-update algorithm and a per-stage dual-update algorithm. For a low-complexity alternative, we design and analyze the transmit power and signal-to-noise ratio (SNR) product (PSP) based selective multiuser diversity (SMuD) schemes. Next, for fair scheduling, we propose rate maximization schemes under access proportional fairness (APF) and rate proportional fairness (RPF) constraints, respectively. The APF is achieved using normalized channel SNR (n-SNR) ranking-based SMuD for user selection per carrier, and the RPF is realized using dynamical carrier assignment based on the target rate ratios. Analytical throughput and fairness metrics are derived and verified via simulations. Numerical results illustrate the sum rate loss caused by rate fairness and access fairness constraints compared to the duality approach. Also, we show that unlike the downlink case, for uplink OFDMA the correlated frequency channels (carriers) cause significant ergodic sum rate degradation compared to the independent channels. These results provide new insight into the achievable uplink OFDMA performance with and without fairness constraints.