Fair data scheduling in OFDM wireless networks based on maximizing utility

This paper proposes a joint layer scheme for fair downlink data scheduling in multiuser OFDM wireless networks. Based on the optimization model formulated as the maximization of total utility function with respect to the mean waiting time of user queue,we present an algorithm with low complexity for dynamic subcarrier allocation (DSA). The decision for subcarrier allocation was made according to delay utility function obtained by the algorithm that instantaneously estimated both channel condition and queue length using an exponentially weighted low-pass time window and pilot signals respectively. The complexity of algorithm was reduced by varying the length of the time window to make use of time diversity,which provided higher throughput ratio. Simulation results demonstrate that compared with the conventional approach,the proposed scheme achieves better performance and can significantly improve fairness among users,with very limited delay performance degradation by using a decreasing concave utility function when the traffic load increases.

This paper proposes a joint layer scheme for fair downlink data scheduling in muhiuser OFDM wireless networks. Based on the optimization model formulated as the maximization of total utility function with respect to the mean waiting time of user queue, we present an algorithm with low complexity for dynamic subcarrier allocation （DSA）. The decision for subcarrier allocation was made according to delay utility function obtained by the algorithm that instantaneously estimated both channel condition and queue length using an exponentially weighted low-pass time window and pilot signals resPectively. The complexity of algorithm was reduced by varying the length of the time window to make use of time diversity, which provided higher throughput ratio. Simulation results demonstrate that compared with the conventional approach, the proposed scheme achieves better performance and can significantly improve fairness among users, with very limited delay performance degradation by using a decreasing concave utility function when the traffic load increases.