Please use this identifier to cite or link to this item: https://hdl.handle.net/11499/46871
Title: FogAI: An AI-supported fog controller for Next Generation IoT
Authors: Kok, Ibrahim
Okay, Feyza Yildirim
Ozdemir, Suat
Keywords: FogAI
Fog computing
Artificial intelligence
DeepQ-learning
Task offloading
IoT
NGIoT
Resource-Allocation
Cloud Control
Edge
Security
Internet
Sdn
Communication
Optimization
Things
Reliability
Publisher: Elsevier
Abstract: In this paper, we present a novel artificial intelligence-based fog controller, called FogAI that provides a versatile control mechanism to the fog layer. FogAI not only abstracts the control mechanism from the fog environment but also offers potential solutions for the problems of fog-based Next Generation Internet of Things (NGIoT) systems. To this end, we first present a comprehensive examination of challenging issues in Fog Computing (FC). Then, we outline possible FogAI based solutions to these challenges from different perspectives. To illustrate the feasibility of our FogAI concept, we design a use case scenario for task offloading problem in FC. Then, we propose a Deep Q-Learning (DQL) algorithm that autonomously performs task offloading in delay-sensitive and computationally-intensive IoT applications and test it on FogAI. The results show that the proposed FogAI-assisted DQL algorithm is superior to existing offloading policies.
URI: https://doi.org/10.1016/j.iot.2022.100572
https://hdl.handle.net/11499/46871
ISSN: 2543-1536
2542-6605
Appears in Collections:Mühendislik Fakültesi Koleksiyonu
Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
WoS İndeksli Yayınlar Koleksiyonu / WoS Indexed Publications Collection

Show full item record



CORE Recommender

SCOPUSTM   
Citations

10
checked on Oct 13, 2024

WEB OF SCIENCETM
Citations

10
checked on Oct 31, 2024

Page view(s)

62
checked on Aug 24, 2024

Google ScholarTM

Check




Altmetric


Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.