Model
Digital Document
Publisher
Florida Atlantic University
Description
In wireless orthogonal frequency division multiple-access (OFDMA) standards,
subcarriers are grouped into chunks and a chunk of subcarriers is made as the minimum allocation unit for subcarrier allocation. We investigate the chunk-based resource allocation for OFDMA downlink, where data streams contain packets with diverse bit-errorrate (BER) requirements. Supposing that adaptive transmissions are based on a number of discrete modulation and coding modes, we derive the optimal resource allocation scheme that maximizes the weighted sum of average user rates under the multiple BER and total power constraints. With proper formulation, the relevant optimization problem is cast as an integer linear program (ILP). We can rigorously prove that the zero duality gap holds for the formulated ILP and its dual problem. Furthermore, it is shown that the optimal strategy for this problem can be obtained through Lagrange dual-based gradient iterations with fast convergence and low computational complexity per iteration. Relying on the stochastic optimization tools, we further develop a novel on-line algorithm capable of dynamically learning the underlying channel distribution and asymptotically approaching the optimal strategy without knowledge of intended wireless channels a priori. In addition, we extend the proposed approach to maximizing the a-fair utility functions of average user rates, and show that such a utility maximization can nicely balance the trade-off between the total throughput and fairness among users.
subcarriers are grouped into chunks and a chunk of subcarriers is made as the minimum allocation unit for subcarrier allocation. We investigate the chunk-based resource allocation for OFDMA downlink, where data streams contain packets with diverse bit-errorrate (BER) requirements. Supposing that adaptive transmissions are based on a number of discrete modulation and coding modes, we derive the optimal resource allocation scheme that maximizes the weighted sum of average user rates under the multiple BER and total power constraints. With proper formulation, the relevant optimization problem is cast as an integer linear program (ILP). We can rigorously prove that the zero duality gap holds for the formulated ILP and its dual problem. Furthermore, it is shown that the optimal strategy for this problem can be obtained through Lagrange dual-based gradient iterations with fast convergence and low computational complexity per iteration. Relying on the stochastic optimization tools, we further develop a novel on-line algorithm capable of dynamically learning the underlying channel distribution and asymptotically approaching the optimal strategy without knowledge of intended wireless channels a priori. In addition, we extend the proposed approach to maximizing the a-fair utility functions of average user rates, and show that such a utility maximization can nicely balance the trade-off between the total throughput and fairness among users.
Member of