MIMO Broadcast Scheduling with Quantized Channel State Information

We develop and analyze a simple, low-complexity system architecture for scheduling over a Gaussian multiple-input multiple-output (MIMO) broadcast channel with infinite message backlogs. In the system of interest, there is a transmitter with m antennas, and n receiving users, where n Gt m. We show that the proposed architecture is strongly asymptotically optimal with respect to average throughput. We further characterize the feedback requirements of the architecture, and highlight various tradeoffs available to the system designer