Graph Neural Network With Soft Actor-Critic and Attention-Based Large Model for Intelligent Edge Routing in Consumer Internet of Things

With the rapid advancement of 5G and AI, numerous Consumer Internet of Things (CIoT) applications have emerged, enhancing everyday life. However, the increasing connection of these applications to edge networks has led to surging traffic, challenging network stability and Quality of Service (QoS). T...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE transactions on consumer electronics Ročník 71; číslo 3; s. 9061 - 9074
Hlavní autoři: Wang, Zhi, Gong, Tao, Hui Huang, Shu
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York IEEE 01.08.2025
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Témata:
Adaptive algorithms
Artificial intelligence
attention mechanism
Consumer electronics
Consumer Internet of Things
deep reinforcement learning
Feature extraction
Graph neural networks
Heuristic algorithms
intelligent edge routing
Internet of Things
Neural networks
Quality of service
Quality of service architectures
Real-time systems
Routing
Routing (telecommunications)
Shortest-path problems
Vehicle dynamics
ISSN:0098-3063, 1558-4127
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:With the rapid advancement of 5G and AI, numerous Consumer Internet of Things (CIoT) applications have emerged, enhancing everyday life. However, the increasing connection of these applications to edge networks has led to surging traffic, challenging network stability and Quality of Service (QoS). The heterogeneous and dynamic nature of CIoT Edge Network Environments (CEEs) exacerbates these challenges. To address these challenges, this study introduces SAG, an adaptive intelligent routing algorithm that integrates Soft Actor-Critic (SAC), Attention Mechanism (AM), and Graph Neural Networks (GNN) to construct a large model tailored for edge routing in consumer-centric environments. In this integrated framework, SAG employs GNN to dynamically capture network topological features and node states, enabling seamless adaptation to evolving consumer demands. AM refines these GNN-extracted features by prioritizing critical QoS parameters, ensuring tailored QoS guarantees for diverse CIoT applications. Finally, SAC with Multi-Constraint Shortest Path (MCSP) determines to make the final routing decisions, providing a customized QoS routing strategy for CIoT applications. Simulation results demonstrate that SAG outperforms state-of-the-art algorithms by significantly enhancing delay, packet loss, and throughput. Additionally, SAG effectively meets the diverse QoS requirements of various CIoT applications while showcasing robust generalization capabilities in dynamic environments.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0098-3063
1558-4127
DOI:10.1109/TCE.2025.3544813