全局公平的自适应比例公平调度

Adaptive proportional fair scheduling with global-fairness

传统的比例公平调度通过牺牲系统的速率性能获得公平性,但该公平性具有"长期"的特点,无法保证进入系统时间较短或在系统中短暂停留的用户的公平性,具有实时业务的用户的时延需求也难以满足.针对以上问题,提出一种全局公平的自适应比例公平调度算法.基站根据全体用户的调度优先级的离散程度,动态调整比例公平算法中的遗忘因子,进而影响用户调度权重的更新.仿真结果表明,与传统的比例公平调度算法相比,自适应比例公平调度算法能够兼顾长期和短期公平性以及系统的和速率,并且能为用户业务保证良好的时延性能.

Conventional proportional fair （PF） scheduling achieves fairness at the cost of the system＇s rate performance. Such fairness is characterized by long-term, and hence cannot guarantee the fairness of subscribers who enter the system temporarily or stay for just a short period of time. In addition, the delay requirement of real-time service users can hardly be met. In order to remedy the above problems, we propose an adaptive proportional fair （APF） scheduling algorithm with global-fairness. The base station dynamically adjusts the forgetting factor in the PF algorithm based on the degree of dispersion of all the users＇ scheduling priorities so as to influence the update of users＇ scheduling weights. Simulation results show that compared to conventional PF scheduling, the APF can achieve both the long-term and short-term fairness and high system sum-rate, and additionally guarantee good delay performance for users＇ service.