A high-performance QoS-enhanced proportional fair scheduling algorithm over downlink OFDMA-based networks

摘要

OFDMA-based networks have been considered to be a promising solution to wireless networks especially for broadband multimedia mobile services. The major challenges pertaining to such OFDMA-based networks are to design a medium access control scheme that is expected to exploit multi-channel multi-rate diversity and with QoS provision. The previous existing scheduling algorithms either target at maximizing system throughput while starving the low-rate users, or focus on throughput fairness but seriously decreasing the system throughput. While proportional fair (PF) has emerged as a prominent candidate in wireless networks, as it achieves a good tradeoff between system throughput and fairness, PF still leads to low throughput of bad channel users, resulting in limiting the attainable QoS throughput for the entire system. In this paper, we propose a QoS-enhanced proportional fair (QEPF) scheduling algorithm, which aims at maximizing QoS throughput while still retaining fairness as well as maximizing system throughput in downlink OFDMA networks. QEPF proposes a QoS-enhanced PF-based utility function, which supports QoS and is a low-complexity algorithm for multi-channel OFDMA networks. We demonstrate via extensive simulation results, that QEPF outperforms PF in terms of system throughput and mean delay. The results illustrate that QEPF surprisingly attains great performance for poor channel users, while not sacrifices the performance of good channel users. Furthermore, the results show that QEPF well accommodates nearly twice the amount of high priority data than PF does.