DQN and Heuristic Precoding-aware Scheduling in Cell-free mMIMO Networks

We propose a deep Q-Learning (DQN) and heuristic-based radio resource scheduling (RRS) algorithms, which determine serving access points (APs), for maximum ratio transmission (MRT) precoder in the cell-free massive multiple-input multiple-output (CF mMIMO) networks. We show that, DQN-based algorithm yields up to 6% higher sum spectral efficiency (SE) than the heuristic and the reinforcement learning based algorithms for MRT precoder, but the heuristic algorithm yields 1000 times shorter computation time than the DQN-based algorithm. Thus, the low complexity MRT heuristic algorithms are the trade off between the performance and the cost.