Prediction based resource allocation in OFDMA

We present a prediction-based resource allocation algorithm (RA) for OFDMA downlink over mobile wireless channels, where prediction errors are incorporated in the problem formulation. A solution based on the empirical characterization of the prediction error is proposed. We test the proposed scheme on a realistic scenario using a practical low complexity channel predictor, suitable for implementation at the mobile stations. Simulation results show that the proposed algorithm outperforms previous prediction-based RA strategies without prediction error characterization, and the system throughput is comparable to the case with perfect channel state information in the transmitter.