Antenna Selection for AF MIMO Relay Networks and the Capacity Scaling Tendency

We propose a novel antenna selection algorithm for amplify-and-forward (AF) multiple-input multiple-output (MIMO) relay networks. Assuming equal-power allocation across the activated antennas, we iteratively select the antenna pairs to greedily maximize the network capacity in closed-form expressions. Simulation results show that our algorithm greatly outperforms the existing schemes, with the gain being more pronounced when the noise at relay nodes is relatively high. Moreover, from the results, we propose a conjecture that the network capacity with antenna selection shall scale logographically with respect to the number of relays, which achieves the capacity upper bound in the sense of asymptotic capacity scaling tendency.