DQN-Based Adaptive MCS and SDM for 5G Massive MIMO-OFDM Downlink

In this letter, we propose a deep Q-network (DQN)-based joint adaptive scheduling algorithm of modulation and coding scheme (MCS) and space division multiplexing (SDM) in 5G massive MIMO-OFDM, that enables the next-generation NodeB (gNB) to choose optimal MCS and SDM for downlink scheduling that maximizes the spectral efficiency (SE) and throughput while in a low block error rate (BLER). The simulation results show that in the typical high-speed mobile and uplink interference scenarios, the proposed DQN-based algorithm has a 10% to 40% improvement in SE and a 10% to 20% improvement in throughput compared to representative algorithms, and can effectively compensate for the channel state information (CSI) feedback delay.