Bandit Learning-based Distributed Computation in Fog Computing Networks: A Survey

Fog computing is a decentralized infrastructure that extends the capabilities of cloud closer to edge network. In fog network (FCN), resources, such as processing power, storage, and networking, are distributed at various points in network, including devices, nodes (FNs) access points, gateways, local servers. This architecture allows resource-limited end devices offload part their computational tasks nearby FNs achieve reduced response delay services energy efficiency. However, high dynamics complicated heterogeneity environment many application scenarios result uncertainty information raised critical challenge design efficient computation offloading strategies. Meanwhile, existing solutions centralized optimization, matching game theory-based inadequate be adopted this context because they require perfect knowledge system parameters. Considering promising approach deal with issues, bandit learning has used recently develop (DCO) algorithms for FCNs. paper, we aim reviewing these state-of-the-art DCO elaborate advantages limitations. Additionally, identify open research challenges provide future directions area. Our survey shows computing, expect will continue explore its potential improving performance efficiency computing-enabled systems.