Optimal chunk-based resource allocation for OFDMA systems with multiple BER requirements

We investigate the chunk-based resource allocation for OFDMA downlink, where data streams contain packets with diverse bit-error-rate (BER) requirements. Supposing adaptive transmissions based on a number of discrete modulation and coding modes, we derive the optimal scheme that maximizes the weighted sum of average user rates under the multiple BER and total power constraints. With the relevant optimization problem cast as an integer linear program, we show that the optimal strategy can be obtained through Lagrange dual-based gradient iterations with fast convergence and low computational complexity per iteration. Furthermore, a novel on-line algorithm is developed to approach the optimal strategy without knowledge of intended wireless channels a priori.