一种针对用户数量变化而改进的比例公平算法

An Improved Proportional Fairness Algorithm for Changes in Number of Users

针对复杂环境下用户数量改变的情况,对用户数量和不同资源分配算法吞吐量之间的关系进行了分析,找出其变化趋势,在现有比例公平算法的基础上,提出了一种调节因子,在公平性指标可以接受的范围内,使得改进后的比例公平算法可以良好适应100以内的用户数量变化,避免过于逼近吞吐量的理论下限值,在吞吐量较低时最高可以获得接近9%的性能提升。提出的算法将能够更好地适应时变环境,有效改善用户数激增带来的系统性能下降问题。

In view of the changes in number of users in complex environment, this paper analyzes the relationship between the number of users and the throughput of different resource allocation algorithms, finds out the trends of changes. Based on existing proportional fairness algorithm, this paper puts forward a regulation factor. In acceptable range of fairness specification, a modified proportional fairness algorithm can adapt well to the changes in number of user less than 100, and avoid too close to the theoretical lower bounds on the throughput values,at lower throughput rates,9% performance promotion can be obtained.The algorithm proposed in this paper will be able to better adapt to the time-varying environment, effectively improve the performance degradation of system caused by proliferation in number of users.