Analysis of performance of proportional fair algorithm with reduced scheduling overhead

Abstract

The throughput (total bit per second per hertz) and delay of a modified proportional fair algorithm (PFA) are computed under different fading environments. The modified PFA uses a dynamic threshold to filter the feedbacks of channel quality (CQ) based on which the users' priorities are calculated for scheduling. Only those in good CQs (relative to their own average CQ) can be the candidates for scheduling. The results show that the modified PFA with low CQ thresholds cannot only reduce scheduling overhead for transmitting and processing CQ feedback but also maintain throughput and delay of the traditional PFA. Interestingly, for a certain environment studied, throughput gain over the traditional PFA can be achieved by the modified PFA. It should be noted that the basic strategy of reducing scheduling overhead behind the modified PFA might also be applied to other resources management scenarios.