Net Throughput Maximization of Per-Chunk User Scheduling for MIMO-OFDM Downlink

Per-chunk user scheduling for multiple-input-multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) downlink is considered. By grouping adjacent subcarriers into chunks, the amount of required channel state information feedback is reduced. Based on the net throughput criterion, which accounts for the reduction in sum rate due to the feedback overhead, it is shown that there exists an optimal chunk size that maximizes the net throughput. To reduce the feedback requirement even further, an opportunistic feedback scheme is proposed, and a close approximation for its net throughput is derived. The net throughput of per-chunk user scheduling with optimized chunk size is compared to various other limited-feedback MIMO-OFDM downlink strategies. The results show that increasing the total number of users in the system results in the net throughput of most existing MIMO-OFDM downlink schemes decreasing to zero for moderate-size user pools, whereas the net throughput of per-chunk user scheduling with opportunistic feedback increases with the total number of users, even when that number is very large (> 1000).