基于强化学习的天地融合网络算力路由发展与挑战

王明乾; 吴晓梅; 杨桄; 雷广旺; 张婷婷; 刘子深

doi:10.11959/j.issn.1000-0801.2025034

您当前的位置：

首页 >

文章列表页 >

基于强化学习的天地融合网络算力路由发展与挑战

综述 | 更新时间：2026-01-09

- 基于强化学习的天地融合网络算力路由发展与挑战
- Development and Challenges of Computing Power Routing in Space-Ground Integrated Networks Based on Reinforcement Learning
- 天地一体化信息网络 2025年6卷第4期页码：1-8
- 作者机构：
  
  1.北京理工大学，北京 100081
  2.中国铁塔股份有限公司，北京 100195
- 作者简介：
  
  [ "王明乾（1999- ），男，北京理工大学博士生，主要研究方向为强化学习、卫星通信、路由优化。" ]
  [ "吴晓梅（1968- ），女，中国铁塔股份有限公司通信技术研究院副总经理、教授级高级工程师，主要研究方向为移动通信、卫星通信、新能源、绿色低碳、算力网络等。" ]
  [ "杨桄（1989- ），男，中国铁塔股份有限公司通信技术研究院工程师，主要研究方向为卫星通信、物联网。" ]
  [ "雷广旺（2002- ），男，北京理工大学硕士生，主要研究方向为卫星通信、路由优化。" ]
  [ "张婷婷（1989− ），女，北京理工大学预聘助理研究员，主要研究方向为天地融合网络、软件定义网络、网络功能虚拟化、智能组网。" ]
  [ "刘子深（2000− ），男，北京理工大学硕士生。主要研究方向为AI 赋能无线通信、卫星通信、贝叶斯推断。" ]
- 基金信息：
  
  国家自然科学基金(62401047)
- DOI：10.11959/j.issn.1000-0801.2025034
  中图分类号： TP393
- 收稿：2025-06-19，
  
  修回：2025-08-14，
  
  纸质出版：2025-12-20
- 稿件说明：
移动端阅览
王明乾,吴晓梅,杨桄等.基于强化学习的天地融合网络算力路由发展与挑战[J].天地一体化信息网络,2025,06(04):1-8.

WANG Mingqian,WU Xiaomei,YANG Guang,et al.Development and Challenges of Computing Power Routing in Space-Ground Integrated Networks Based on Reinforcement Learning[J].Space-Integrated-Ground Information Networks,2025,06(04):1-8.
王明乾,吴晓梅,杨桄等.基于强化学习的天地融合网络算力路由发展与挑战[J].天地一体化信息网络,2025,06(04):1-8. DOI： 10.11959/j.issn.1000-0801.2025034.

WANG Mingqian,WU Xiaomei,YANG Guang,et al.Development and Challenges of Computing Power Routing in Space-Ground Integrated Networks Based on Reinforcement Learning[J].Space-Integrated-Ground Information Networks,2025,06(04):1-8. DOI： 10.11959/j.issn.1000-0801.2025034.

摘要

天地融合网络通过协同卫星、空中平台和地面设施构建立体化通信架构，其动态拓扑和异构算力资源对路由机制提出了全新挑战。面对这一复杂场景，强化学习因其强大的环境适应能力和智能决策优势成为解决算力路由问题的研究热点。首先，系统梳理并深入探究强化学习在算力路由中的应用实践；然后，从算法范式与优化目标两个维度解析其核心设计原理；最后，深入剖析并前瞻性地指出其在天地融合网络环境下所面临的关键挑战。

Abstract

Space-air-ground integrated networks (SAGIN) establish a multi-dimensional communication architecture by synergistically coordinating satellites

aerial platforms

and terrestrial facilities. The inherent characteristics of dynamic topologies and heterogeneous computing resources within this architecture present novel challenges for routing mechanisms. In response to this complex scenario

reinforcement learning (RL) has emerged as a prominent research focus for addressing computing-aware routing problems

owing to its robust environmental adaptability and intelligent decision-making capabilities. This paper systematically reviews and investigates the application of RL in computing-aware routing. It analyzes the core design principles from the dual perspectives of algorithmic paradigms and optimization objectives. Finally

this paper deeply analyzes and prospectively identifies the key challenges that RL faces in the context of SAGIN environments.

关键词

Keywords

references

吴巍 . 天地一体化信息网络发展综述 [J ] . 天地一体化信息网络 , 2020 , 1 ( 1 ) : 1 - 16 .

WU W . Survey on the development of space-integrated-ground information network [J ] . Space-Integrated-Ground Information Networks , 2020 , 1 ( 1 ) : 1 - 16 .

曹素芝 , 孙雪 , 王厚鹏 , 等 . 星地融合网络智能路由技术综述 [J ] . 天地一体化信息网络 , 2021 , 2 ( 2 ): 11 - 19 .

CAO S Z , SUN X , WANG H P , et al . Review of intelligent routing technology in integrated satellite-ground network [J ] . Space-Integrated-Ground Information Networks , 2021 , 2 ( 2 ): 11 - 19 .

田厉锋 , 张佳宇 , 李昕 , 等 . 空天地海一体化算力资源管理技术研究及展望 [J ] . 天地一体化信息网络 , 2025 , 6 ( 2 ): 1 - 10 .

TIAN L F , ZHANG J N , LI X , et al . Research and prospect of integrated computing resource management technologies for space-air-ground-sea networks [J ] . Space-Integrated-Ground Information Networks , 2025 , 6 ( 2 ): 1 - 10 .

SERRANO P , GRAMAGLIA M , MANCINI F , et al . Balloons in the sky: unveiling the characteristics and trade-offs of the google loon service [J ] . IEEE Transactions on Mobile Computing , 2023 , 22 ( 6 ): 3165 - 3178 .

LU D , XIONG C K , LYU B Z , et al . Multi-mode hybrid aerial underwater vehicle with extended endurance [C ] // Proceedings of the 2018 OCEANS - MTS/IEEE Kobe Techno-Oceans (OTO) . Piscataway : IEEE Press , 2018 : 1 - 7 .

许柳飞 , 罗志勇 . 面向低轨卫星网络的算力路由策略 [J ] . 移动通信 , 2025 , 49 ( 6 ): 35 - 42 .

XU L F , LUO Z Y . Computing power routing strategy for low earth orbit satellite networks [J ] . Mobile Communications , 2025 , 49 ( 6 ): 35 - 42 .

孔梦燕 , 张亚生 , 董飞虎 . 基于深度强化学习的低轨卫星网络算力路由研究 [J ] . 计算机测量与控制 , 2025 , 33 ( 2 ): 286 - 292 .

KONG M Y , ZHANG Y S , DONG F H . Research on computing power routing for LEO satellite networks based on deep reinforcement learning [J ] . Computer Measurement & Control , 2025 , 33 ( 2 ): 286 - 292 .

ALBRECHT S V , CHRISTIANOS F , SCHAFER L . Title of the monograph [M ] . Cambridge : MIT Press , 2024 .

CAO J , ZHANG S , CHEN Q , et al . Computing-aware routing for LEO satellite networks: a transmission and computation integration approach [J ] . IEEE Transactions on Vehicular Technology , 2023 , 72 ( 12 ): 16607 - 16623 .

NGUYEN T H , PARK L . A survey on deep reinforcement learning-driven task offloading in aerial access networks [C ] // Proceedings of the 2022 13th International Conference on Information and Communication Technology Convergence (ICTC) . Piscataway : IEEE Press , 2022 : 822 - 827 .

FU J , LIANG L , LI Y , et al . Deep deterministic policy gradient algorithm for space/aerial-assisted computation offloading [C ] // Proceedings of the International Conference on Communications and Networking in China . Cham : Springer International Publishing , 2021 : 523 - 537 .

谢明熹 . 算力网络中基于深度强化学习的路由算法研究 [D ] . 北京 : 北京邮电大学 , 2024 .

XIE M X . Research on routing algorithm based on deep reinforcement learning in computing power network [D ] . Beijing : Beijing University of Posts and Telecommunications , 2024 .

CHEN Z Y , ZHANG S C , TANG Y , et al . Optimization of computing power network routing strategies based on multi-agent soft actor-critic [C ] // Proceedings of the 2024 9th International Conference on Intelligent Computing and Signal Processing (ICSP) . Piscataway : IEEE Press , 2024 : 426 - 430 .

LI Y , LIANG L , FU J , et al . Multiagent reinforcement learning for task offloading of space/aerial‐assisted edge computing [J ] . Security and Communication Networks , 2022 ( 1 ): 4193365 .

HUANG L , SUN R , CHENG N , et al . Delay-oriented knowledge-driven resource allocation in SAGIN-based vehicular networks [C ] // Proceedings of the 2023 IEEE Wireless Communications and Networking Conference (WCNC) . Piscataway : IEEE Press , 2023 : 1 - 6 .

杨桂松 , 黄业超 , 何杏宇 . 低轨卫星网络中面向时延优化的跨域边缘计算卸载策略 [J ] . 天地一体化信息网络 , 2025 , 6 ( 2 ): 65 - 75 .

YANG G S , HUANG Y C , HE X Y . Cross-domain edge computing offloading strategy for delay-optimized in low earth orbit satellite network [J ] . Space-Integrated-Ground Information Networks , 2025 , 6 ( 2 ): 65 - 75 .

HAN C , XIONG W , YU R H . Deep reinforcement learning-based multipath routing for LEO Mega constellation networks [J ] . Electronics (Switzerland) , 2024 , 13 ( 15 ): 3054 .

DONG F , SONG J , ZHANG Y , et al . DRL-based load-balancing routing scheme for 6G space-air-ground integrated networks [J ] . Remote Sensing , 2023 , 15 ( 11 ): 2801 .

QIU Y , LUO X , NIU J , et al . Category attribute-oriented heterogeneous resource allocation and task offloading for SAGIN edge computing [J ] . Journal of Sensor and Actuator Networks , 2025 , 14 ( 4 ): 81 .

LIU X , CHAI Z Y , LI Y L , et al . Multi-objective deep reinforcement learning for computation offloading in UAV-assisted multi-access edge computing [J ] . Information Sciences , 2023 , 64 ( 2 ): 119154 .

浏览量

133

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

天基算力网络关键技术综述

面向新型电力系统的天空地一体化网络架构设计