Fog Computing Tasks Management Based on Federated Reinforcement Learning.
Saved in:
| Title: | Fog Computing Tasks Management Based on Federated Reinforcement Learning. |
|---|---|
| Authors: | Azarkasb, Seyed Omid, Khasteh, Seyed Hossein |
| Source: | Journal of Grid Computing; Mar2025, Vol. 23 Issue 1, p1-46, 46p |
| Abstract: | The exponential growth of IoT networks and the emergence of fog computing have created a critical need for innovative solutions to manage tasks efficiently at the network edge, where traditional cloud-based task allocation methods often fall short. These conventional approaches struggle to meet the increasing demands for bandwidth, low-latency processing, and robust privacy in distributed IoT ecosystems. To address these challenges, this study introduces a novel Federated Reinforcement Learning (FRL)-based framework for adaptive, decentralized task scheduling in fog computing environments. Our approach combines the collective learning power of federated learning with the adaptability of reinforcement learning, enabling fog nodes to make real-time, autonomous decisions that minimize reliance on centralized cloud services, conserve bandwidth, and safeguard data privacy. By learning from local interactions while sharing aggregated insights in a privacy-preserving manner, each fog node independently optimizes its task allocation policies. The proposed framework was implemented in a three-layer fog architecture and rigorously tested in real-world conditions. The results demonstrated significant improvements across multiple key performance metrics. Specifically, FRL achieved 51.69% improvement in data privacy protection, 21.84% faster convergence rate, 9.75% reduction in communication overhead, and 4.98% higher decision accuracy. Moreover, the system showed 37.65% better scalability and 48.68% stronger data security compared to traditional methods. These gains culminated in an overall average improvement of 28.28%, underscoring the efficacy of the FRL architecture in enhancing adaptability, optimizing resource utilization, and significantly reducing task completion times and delays.This research offers a robust, scalable solution for intelligent task management in fog computing environments, making substantial contributions to next-generation IoT applications that demand high privacy, low latency, and adaptability. By achieving superior performance in unpredictable, dynamic conditions, our FRL framework paves the way for more efficient and secure fog-based IoT systems. [ABSTRACT FROM AUTHOR] |
| Copyright of Journal of Grid Computing is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Complementary Index |
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstract: | The exponential growth of IoT networks and the emergence of fog computing have created a critical need for innovative solutions to manage tasks efficiently at the network edge, where traditional cloud-based task allocation methods often fall short. These conventional approaches struggle to meet the increasing demands for bandwidth, low-latency processing, and robust privacy in distributed IoT ecosystems. To address these challenges, this study introduces a novel Federated Reinforcement Learning (FRL)-based framework for adaptive, decentralized task scheduling in fog computing environments. Our approach combines the collective learning power of federated learning with the adaptability of reinforcement learning, enabling fog nodes to make real-time, autonomous decisions that minimize reliance on centralized cloud services, conserve bandwidth, and safeguard data privacy. By learning from local interactions while sharing aggregated insights in a privacy-preserving manner, each fog node independently optimizes its task allocation policies. The proposed framework was implemented in a three-layer fog architecture and rigorously tested in real-world conditions. The results demonstrated significant improvements across multiple key performance metrics. Specifically, FRL achieved 51.69% improvement in data privacy protection, 21.84% faster convergence rate, 9.75% reduction in communication overhead, and 4.98% higher decision accuracy. Moreover, the system showed 37.65% better scalability and 48.68% stronger data security compared to traditional methods. These gains culminated in an overall average improvement of 28.28%, underscoring the efficacy of the FRL architecture in enhancing adaptability, optimizing resource utilization, and significantly reducing task completion times and delays.This research offers a robust, scalable solution for intelligent task management in fog computing environments, making substantial contributions to next-generation IoT applications that demand high privacy, low latency, and adaptability. By achieving superior performance in unpredictable, dynamic conditions, our FRL framework paves the way for more efficient and secure fog-based IoT systems. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 15707873 |
| DOI: | 10.1007/s10723-025-09796-4 |