Performance Analysis of Multiuser Selection Diversity

In this paper, the performance of scheduling algorithms exploiting the multiuser selection diversity is studied. The authors consider schedulers with affordable-rate transmission and adaptive transmission based on the absolute signal-to-noise ratio (SNR) and the normalized SNR. In contrast to previous studies on multiuser diversity systems, channel dynamics is taken into consideration in this paper by a novel formulation based on the level-crossing analysis of stochastic processes. Then, a connection is made between the Doppler frequency shift, which indicates the channel temporal correlation, and the average (channel) access time, the average waiting time between accesses, and the average access rate of active users. These properties are important for the scheduler design, especially for applications where delay is a concern. In addition, analytical expressions for the system throughput and the degree of fairness when users have nonidentical average channel conditions are presented. These expressions quantify the effect of disparateness in users' average channel conditions on the system performance.