User Grouping and Resource Allocation for Uplink of MU-MIMO-OFDMA-Enabled WLAN Using Multi-Agent Reinforcement Learning

Wireless local area network (WLAN) standards aim to provide reliable wireless connectivity in dense environments. With the advent of multi-user (MU) communication technologies, such as orthogonal frequency division multiple access (OFDMA) and multiple input multiple output (MIMO), spectrum resources can be shared over time, frequency, and space. Therefore, WLAN capacity can be significantly improved by scheduling MU-MIMO transmissions for groups of spatially compatible users/stations (STAs) over several OFDMA resource units. In this paper, we study user grouping and resource allocation for the uplink scheduled access (UL-SA) of a joint MU-MIMO-OFDMA-enabled WLAN, which is considered a combinatorial optimization problem. Performing the optimization task is computationally exhaustive and time-consuming for a WLAN access point (AP). Moreover, for the UL-SA, the AP cannot jointly optimize the multi-channel UL access and the transmit power of the STAs, which makes it challenging to evaluate the effectiveness of a resource allocation strategy in advance. Thus, we propose user grouping and resource allocation using multi-agent reinforcement learning (GRAMARL). Leveraging the multi-agent framework and advances in deep Q-learning, GRAMARL can learn an effective resource management strategy to optimize the UL-SA. Our numerical results validate GRAMARL’s superior performance over benchmark techniques. Furthermore, we demonstrate GRAMARL’s ability to operate within the spatial degrees of freedom provided by different antenna settings on the AP and adapt to dynamic WLAN environments.