Distributed link adaptation for multicast traffic in MIMO-OFDM systems

Multicast traffic exploits the broadcast nature of the wireless medium to deliver the same data to multiple users improving the bandwidth efficiency. Link adaptation can be used in multicast transmission to further increase data rates exploiting feedback from the users. However, it is not easy to have the quality of service (QoS) of every intended receiver met while pushing the data rate to the link capacity. Due to this difficulty, the conventional approach is to transmit isotropically with a fixed basic rate giving up the opportunity of increased throughput. For point-to-point unicast traffic, machine learning algorithms have recently found successful application in link adaptation due to their flexibility and ability to capture more environmental effects implicitly than classical adaptation algorithms. In this paper, we propose a machine learning based distributed algorithm for link adaptation for multicast traffic in multiple-input multiple-output (MIMO) orthogonal frequency-division multiplexing (OFDM) systems. Our computations show that the data driven approach for link adaptation provides good prediction of the optimal modulation and coding scheme (MCS) outperforming the fixed MCS policies collectively. The distributed algorithm using dual decomposition reduces the required feedback amount significantly while maintaining the near-optimal throughput performance.