Flexible load management algorithm in future 6G networks based on federated deep reinforcement learning

The article discusses the use of federated deep reinforcement learning (FDRL) to optimize network load using edge computing in 6G mobile networks. As you know, sixth-generation networks pose new challenges related to data processing (in terms of processing speed and data volume), data transfer speed and confidentiality. The proposed approach integrates the advantages of federated and deep learning with reinforcement for efficient distribution of computational load. The use of FDRL helps to reduce the load on central servers by ensuring data processing at peripheral nodes, which potentially not only increases efficiency, but also improves system confidentiality, minimizing data transmission over the network. The study shows the possibility of adaptive response of nodes to changes in the network and a reduction in computing load of at least up to 70%, foreshadowing broad prospects for resource management in 6G networks.