Optimizing Distributed MIMO Wi-Fi Networks with Deep Reinforcement Learning
share
This paper explores the feasibility of leveraging concepts from deep reinforcement learning (DRL) for dynamic resource management in wireless networks. Specifically, this work considers the case of distributed multi-user MIMO (D-MIMO) based Wi-Fi networks and studies how DRL can help improve the performance of these networks, particularly in dynamic scenarios. D-MIMO is a technique by which a set of wireless access points are synchronized and grouped together to jointly serve multiple users simultaneously. This paper addresses two dynamic resource management problems pertaining to D-MIMO Wi-Fi networks in detail -- channel assignment of D-MIMO groups, and deciding how to cluster access points to form D-MIMO groups. A DRL framework is constructed through which an agent interacts with a D-MIMO network, learns about the network environment, and is successful in converging to policies which augment the performance of the D-MIMO network in different scenarios. Through extensive simulations and on-line training based on dense Wi-Fi deployments, this paper demonstrates the merits of using DRL by achieving up to a 20 the performance of D-MIMO networks with DRL compared to when DRL is not used.
READ FULL TEXT
