Matching Theoretic Beam Selection in Millimeter-Wave Multi-User MIMO Systems

Beamspace multiple-input multiple-output (MIMO) with beam selection offers an attractive solution to reduce the number of radio-frequency chains in a high-dimensional millimeter-wave MIMO system. Considering a multi-user MIMO system in which an access point, having an N element antenna array, communicates with K users (K < N), each having a single antenna, we address the problem of selecting K beams with sum rate as the performance metric. We model beam selection as a two-sided matching between the two sets of players (users and beams) and consider two different ways of modeling players preferences. When the interdependences between the players' preferences are considered, it becomes matching with externalities. The proposed algorithm, which finds a stable matching in such a case, outperforms the existing beam selection algorithms.