Throughput Enhancement of High-Bandwidth Wireless Network Using Deep Double Q-Network Algorithm

In high-bandwidth network environments, multiple users and devices frequently share similar high-frequency bands, but signal propagation limitations lead to unstable network links. These can affect the throughput of high-bandwidth wireless networks, reducing the continuity and efficiency of data tra...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Mobile networks and applications Jg. 30; H. 1-2; S. 177 - 190
Hauptverfasser: Wang, Chengman, Aljohani, Abeer, Khan, Fazlullah
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.04.2025
Springer Nature B.V
Schlagworte:
Algorithms
Bandwidths
Communications Engineering
Computer Communication Networks
Data transmission
Deep learning
Electrical Engineering
Engineering
Frequencies
Internet of Things
IT in Business
Machine learning
Mathematical models
Networks
Neural networks
Performance enhancement
Propagation
Scheduling
User needs
Wireless communications
Wireless networks
Double Deep Q-Network algorithm
High bandwidth
Throughput performance enhancement
Wireless network
Mathematical modelling
Joint optimisation
ISSN:1383-469X, 1572-8153
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In high-bandwidth network environments, multiple users and devices frequently share similar high-frequency bands, but signal propagation limitations lead to unstable network links. These can affect the throughput of high-bandwidth wireless networks, reducing the continuity and efficiency of data transmission. Therefore, this paper constructs a high-bandwidth wireless network throughput performance enhancement model based on the Double Deep Q-Network algorithm (Double DQN). This method designs a mathematical model for high-bandwidth wireless networks, visually presents node connections using undirected graphs, and constructs a throughput performance enhancement model based on channel allocation and link scheduling by combining interference model analysis. In addition, this method introduces the Double DQN algorithm, integrates the reinforcement learning framework, and realizes dynamic adjustment of optimal channel allocation and link scheduling strategies through continuous interaction and learning between the agent and the environment. Experimental results confirm that this model significantly improves the throughput of high-bandwidth wireless networks to above 2500 kb/s and reduces the call drop rate of voice and video network services to 0.01–0.02%, meeting the demand for high data continuous transmission.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1383-469X
1572-8153
DOI:10.1007/s11036-025-02448-7