A DRL-based RAQ-GERT dynamic resource allocation algorithm considering utility for multibeam satellite system

With the evolution and popularity of smart devices, the demand and requirement (e.g., communication, file transfer) of satellite users have increased rapidly. Moreover, users have different preferences for services and the quality of service (QoS), like delay and throughput, which leads to user hete...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Computer networks (Amsterdam, Netherlands : 1999) Ročník 257; s. 110940
Hlavní autoři: Wu, Shuang, Fang, Zhigeng, Hua, Chenchen, Tao, Liangyan, Zhang, Jingru
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 01.02.2025
Témata:
Deep reinforcement learning (DRL)
Dynamic resource management (DRM)
Multibeam satellite system (MSS)
Quality of service (QoS)
Queuing graphical evaluation and review technique (Q-GERT)
System utility
Multibeam satellite system (MSS)
Quality of service (QoS)
Queuing graphical evaluation and review technique (Q-GERT)
Deep reinforcement learning (DRL)
Dynamic resource management (DRM)
System utility
ISSN:1389-1286
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:With the evolution and popularity of smart devices, the demand and requirement (e.g., communication, file transfer) of satellite users have increased rapidly. Moreover, users have different preferences for services and the quality of service (QoS), like delay and throughput, which leads to user heterogeneity. Facing numerous, time-varying, and heterogeneous users, how to dynamically allocate limited spectrum and on-board power while satisfying user requirements is the major challenge for the multibeam satellite system (MSS). Aiming to seek a solution, firstly, the resource allocation queue graphical evaluation and review technique (RAQ-GERT) network is constructed to describe the service process of the MSS, as well as to compute the channel condition parameters during the whole process. Next, appropriate QoS indicators are selected based on user requirements. Then, QoS indicators are calculated from the results of the RAQ-GERT network, which are combined to form the optimization objective of the MSS by drawing on the Cobb–Douglas utility function. After that, guided by the utility of the MSS, the proximal policy optimization (PPO) algorithm is applied to explore the optimal resource allocation scheme in this heterogeneous user scenario. Finally, the simulation comparisons show that the proposed scheme has enhancements in several performances, up to 42.19 % in service rate, 53.58 % in system capacity, and 3.42 % in throughput with minimal increase in latency.
ISSN:1389-1286
DOI:10.1016/j.comnet.2024.110940