Quantized scheduling for radio networks with heterogeneous fading characteristics

As the 3G cdma2000^® networks with high data rates get deployed, it is interesting to study the impact of the scheduling algorithms used in these networks, and to improve over their performance. In this paper, we introduce a quantized scheduling that takes into account the fact that mobile nodes and access points can communication only through a finite set of rates or power vectors. This approach allows also to improve over the traditional scheduling algorithm used in the cdma2000 networks, proportional fairness, in terms of fairness in heterogeneous channels (that is channels were the fading is governed by different distributions for different nodes) by mapping all channels to a vector of i.i.d. random variables from which to base the scheduling decision. There is no performance degradation for using this algorithm in Rayleigh fading channels, but improved fairness in the heterogeneous fading case.