一种低复杂度的多用户分集调度算法

Low-complexity Scheduling Algorithm in Multiuser Diversity System

利用多用户分集增益可以显著提高系统容量,但最大化吞吐量的调度算法会导致用户不公平地占用资源,比例公平算法在损失较大容量的条件下可以让用户公平分享资源。重点考虑多用户分集系统中的系统容量和公平性的折中,结合最大化吞吐量和比例公平调度算法的思想,提出一种低复杂度的加权调度算法,并推导出门限反馈和全反馈时的系统容量和公平性的理论表达式。仿真结果表明,所提算法在损失较少容量的同时,可以保持较高的公平性。并且在用户数较少时,性能优于已有的折中算法,在用户数较多时,可以通过设置反馈门限来降低用户的反馈开销,同时不会带来容量和公平性的损失。

The system capacity can be increased by exploiting the multiuser diversity gain, but the scheduling algorithm based on maximizing throughput （MT） leads to unfairness of the resource allocation. The proportional fairness （PF） algorithm can increase the fairness at the cost of system capacity. In the paper, the tradeoff on the system capacity and fairness in the multiuser diversity system are focused on. Considering the advantages and disadvantages of MT and PF algorithms, a low-complexity weighted scheduling algorithm is proposed. The theoretical expression of system capacity and fairness is de- rived with and without the threshold feedback. The simulation results show the proposed algorithm can achieve high fairness at the low cost of system capacity, and the performance is better than the previous algorithms when the number of the users is small. In addition, the appropriate feedback threshold is chosen to reduce the feedback overhead without the cost of capacity and fairness.