NetAutoLLM: Network Autonomous Health Management Based on LLM

TANG Fengxiao; WANG Xiaonan; ZHANG Junjian; ZHAO Ming

doi:10.12263/DZXB.20250440

您当前的位置：

首页 >

文章列表页 >

NetAutoLLM: Network Autonomous Health Management Based on LLM

PAPERS | 更新时间：2026-06-16

- NetAutoLLM: Network Autonomous Health Management Based on LLM
- ACTA ELECTRONICA SINICA Vol. 54, Issue 3, Pages: 1147-1160(2026)
- 作者机构：
  
  中南大学计算机学院，湖南长沙 410083
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(62302527);Hunan Provincial Natural Science Foundation(2025JJ90177;2024JJ9173);Hunan Provincial Natural Science Foundation(2023jj40774)
- DOI：10.12263/DZXB.20250440
  CLC： TP277;
- Received：29 May 2025，
  
  Accepted：26 February 2026，
  
  Published：25 March 2026
- 稿件说明：
移动端阅览
唐枫枭, 王啸楠, 张君健, 等. NetAutoLLM：基于LLM的自主化网络健康管理[J]. 电子学报, 2026, 54(03): 1147-1160.

TANG Fengxiao, WANG Xiaonan, ZHANG Junjian, et al. NetAutoLLM: Network Autonomous Health Management Based on LLM[J]. Acta Electronica Sinica, 2026, 54(03): 1147-1160.
唐枫枭, 王啸楠, 张君健, 等. NetAutoLLM：基于LLM的自主化网络健康管理[J]. 电子学报, 2026, 54(03): 1147-1160. DOI：10.12263/DZXB.20250440

TANG Fengxiao, WANG Xiaonan, ZHANG Junjian, et al. NetAutoLLM: Network Autonomous Health Management Based on LLM[J]. Acta Electronica Sinica, 2026, 54(03): 1147-1160. DOI：10.12263/DZXB.20250440

摘要

在现代网络运营和维护中，可靠的网络服务与系统健康管理至关重要。传统的网络健康管理方案依赖于规则匹配的简单算法或机器学习模型，难以应对通信状况复杂、信息难以拟合的网络通信环境。大型语言模型（Large Language Model，LLM）因其具备自主的思考与推理能力可以解决这些问题。目前，最先进的基于LLM的算法在通信网络领域取得了显著进展。然而，现有方案通常将LLM作为特征梳理与学习的工具，LLM本身被动接收数据，其自主性和思考能力受到了极大的限制。这导致目前基于LLM的方案无法直接应用于网络运营与维护任务中。为了解决这个问题，本文通过外接行为树（Behavior Trees，BTs）赋予LLM感知与修改网络环境的能力，并构建自主化的网络健康管理框架。此外，该框架通过双池设计保证各专家信息隔离与案例经验更新。同时，通过多专家讨论的方式加固操作安全性，以保证网络运维任务的完成。实验证明：相较于目前流行的模型，基于LLM的自主化网络健康管理（Network Autonomous health management based on LLM，NetAutoLLM）模型在异常检测任务中精度提高了8.44个百分点；在故障溯源任务中，精度提高了21.7个百分点，同时可以自动缓解故障。

Abstract

In modern network operation and maintenance

reliable network services and system health management are paramount. Traditional network health management solutions relying on machine learning models or rule-based algorithms struggle to address complex network communication environments with heterogeneous devices. Large language models (LLMs)

with their reasoning and generalization capabilities

offer potential solutions to these challenges. While state-of-the-art LLM-based algorithms have achieved significant progress in network domains

existing approaches typically utilize LLMs merely as tools for feature extraction and learning

where models passively receive data with severely constrained autonomy and cognitive capabilities. This limitation hinders direct application of current LLM-based solutions to network operation and maintenance tasks. To address this issue

this paper proposes an autonomous network health management framework by integrating external behavior trees (BTs) that empower LLMs with network environment perception and modification capabilities. The framework ensures information authenticity through a dual-pool design distinguishing public and private information

while employing multi-expert discussions to verify action effectiveness and guarantee successful network maintenance. Experimental results demonstrate that compared to conventional models

network autonomous health management based on LLM (NetAutoLLM) achieves an 8.44 percentage points improvement in anomaly detection accuracy over mainstream models

and enhances fault localization precision by 21.7 percentage points

while enabling automated fault mitigation.

关键词

Keywords

references

Yuan X , Tang F X , Zhao M , et al . Joint rate and coverage optimization for the THz/RF multi-band communications of space-air-ground integrated network in 6G [J ] . IEEE Transactions on Wireless Communications , 2024 , 23 ( 6 ): 6669 - 6682 . DOI: 10.1109/twc.2023.3336016 http://dx.doi.org/10.1109/twc.2023.3336016

Arun V , Balakrishnan H . Copa: Practical delay-based congestion control for the Internet [C ] // Proceedings of the 2018 Applied Networking Research Workshop . New York : ACM , 2018 : 19 . DOI: 10.1145/3232755.3232783 http://dx.doi.org/10.1145/3232755.3232783

Meng Z L , Wang M H , Bai J S , et al . Interpreting deep learning-based networking systems [C ] // Proceedings of the Annual Conference of the ACM Special Interest Group on Data Communication on the Applications, Technologies, Architectures, and Protocols for Computer Communication . New York : ACM , 2020 : 154 - 171 . DOI: 10.1145/3387514.3405859 http://dx.doi.org/10.1145/3387514.3405859

Yuan X , Wang X N , Tang F X , et al . MPITE: Multidimensional performance evaluator for interpretable and traceable network performance evaluation [J ] . IEEE Transactions on Networking , 2025 , 33 ( 5 ): 2458 - 2473 . DOI: 10.1109/ton.2025.3562348 http://dx.doi.org/10.1109/ton.2025.3562348

Li Z , Zhu X Q , Gahm J , et al . Probe and adapt: Rate adaptation for HTTP video streaming at scale [J ] . IEEE Journal on Selected Areas in Communications , 2014 , 32 ( 4 ): 719 - 733 . DOI: 10.1109/jsac.2014.140405 http://dx.doi.org/10.1109/jsac.2014.140405

Lin X J , Xiong G , Gou G P , et al . ET-BERT: A contextualized datagram representation with pre-training transformers for encrypted traffic classification [C ] // Proceedings of the ACM Web Conference 2022 . New York : ACM , 2022 : 633 - 642 . DOI: 10.1145/3485447.3512217 http://dx.doi.org/10.1145/3485447.3512217

Guo Z Q , Tang F X , Luo L F , et al . A survey on applications of large language model-driven digital twins for intelligent network optimization [J ] . IEEE Communications Surveys & Tutorials , 2026 , 28 : 3388 - 3411 . DOI: 10.1109/comst.2025.3568637 http://dx.doi.org/10.1109/comst.2025.3568637

Tang J , Tang F X , Long S F , et al . Utilizing large language models for advanced optimization and intelligent management in space-air-ground integrated networks [J ] . IEEE Network , 2025 , 39 ( 5 ): 173 - 181 . DOI: 10.1109/mnet.2024.3519664 http://dx.doi.org/10.1109/mnet.2024.3519664

Xia Z X , Zhou Y J , Yan F Y , et al . Automatic curriculum generation for learning adaptation in networking [PP/OL ] . V2.arXiv ( 2022-09-08 )[ 2025-05-29 ] . https://doi.org/10.48550/arXiv.2202.05940 https://doi.org/10.48550/arXiv.2202.05940 .

Dhamdhere A , Teixeira R , Dovrolis C , et al . NetDiagnoser: Troubleshooting network unreachabilities using end-to-end probes and routing data [C ] // Proceedings of the 2007 ACM CoNEXT conference . New York : ACM , 2007 : 1 - 12 . DOI: 10.1145/1364654.1364677 http://dx.doi.org/10.1145/1364654.1364677

Szilagyi P , Novaczki S . An automatic detection and diagnosis framework for mobile communication systems [J ] . IEEE Transactions on Network and Service Management , 2012 , 9 ( 2 ): 184 - 197 . DOI: 10.1109/tnsm.2012.031912.110155 http://dx.doi.org/10.1109/tnsm.2012.031912.110155

Peng Y H , Bao Y X , Chen Y R , et al . DL2: A deep learning-driven scheduler for deep learning clusters [J ] . IEEE Transactions on Parallel and Distributed Systems , 2021 , 32 ( 8 ): 1947 - 1960 . DOI: 10.1109/tpds.2021.3052895 http://dx.doi.org/10.1109/tpds.2021.3052895

Mao H Z , Netravali R , Alizadeh M . Neural adaptive video streaming with pensieve [C ] // Proceedings of the Conference of the ACM Special Interest Group on Data Communication . New York : ACM , 2017 : 197 - 210 . DOI: 10.1145/3098822.3098843 http://dx.doi.org/10.1145/3098822.3098843

Bentaleb A , Timmerer C , Begen A C , et al . Bandwidth prediction in low-latency chunked streaming [C ] // Proceedings of the 29th ACM Workshop on Network and Operating Systems Support for Digital Audio and Video . New York : ACM , 2019 : 7 - 13 . DOI: 10.1145/3304112.3325611 http://dx.doi.org/10.1145/3304112.3325611

Mei L F , Hu R C , Cao H W , et al . Realtime mobile bandwidth prediction using LSTM neural network [M ] // Passive and Active Measurement . ChamSpringer International Publishing , 2019 : 34 - 47 . DOI: 10.1007/978-3-030-15986-3_3 http://dx.doi.org/10.1007/978-3-030-15986-3_3

朱晓荣 , 张佩佩 . 基于GAN的异构无线网络故障检测与诊断算法 [J ] . 通信学报 , 2020 , 41 ( 8 ): 110 - 119 .

Zhu Xiaorong , Zhang Peipei . Fault detection and diagnosis method for heterogeneous wireless network based on GAN [J ] . Journal on Communications , 2020 , 41 ( 8 ): 110 - 119 . (in Chinese)

Kan N W , Jiang Y K , Li C L , et al . Improving generalization for neural adaptive video streaming via meta reinforcement learning [C ] // Proceedings of the 30th ACM International Conference on Multimedia . New York : ACM , 2022 : 3006 - 3016 . DOI: 10.1145/3503161.3548331 http://dx.doi.org/10.1145/3503161.3548331

Jeong Y , Yang E , Ryu J H , et al . AnomalyBERT: Self-supervised transformer for time series anomaly detection using data degradation scheme [PP/OL ] . V1.arXiv ( 2023-05-08 )[ 2025-05-29 ] . https://doi.org/10.48550/arXiv.2305.04468 https://doi.org/10.48550/arXiv.2305.04468 .

Mao H Z , Schwarzkopf M , Venkatakrishnan S B , et al . Learning scheduling algorithms for data processing clusters [C ] // Proceedings of the ACM Special Interest Group on Data Communication . New York : ACM , 2019 : 270 - 288 . DOI: 10.1145/3341302.3342080 http://dx.doi.org/10.1145/3341302.3342080

Yen C Y , Abbasloo S , Chao H J . Computers can learn from the heuristic designs and master Internet congestion control [C ] // Proceedings of the ACM SIGCOMM 2023 Conference . New York : ACM , 2023 : 255 - 274 . DOI: 10.1145/3603269.3604838 http://dx.doi.org/10.1145/3603269.3604838

Wang X W , Lin X H , Dang X C . Supervised learning in spiking neural networks: A review of algorithms and evaluations [J ] . Neural Networks , 2020 , 125 : 258 - 280 . DOI: 10.1016/j.neunet.2020.02.011 http://dx.doi.org/10.1016/j.neunet.2020.02.011

Wu D , Wang X D , Qiao Y Q , et al . NetLLM: Adapting large language models for networking [C ] // Proceedings of the ACM SIGCOMM 2024 Conference . New York : ACM , 2024 : 661 - 678 . DOI: 10.1145/3651890.3672268 http://dx.doi.org/10.1145/3651890.3672268

Liu B X , Liu X Y , Gao S J , et al . LLM4CP: Adapting large language models for channel prediction [J ] . Journal of Communications and Information Networks , 2024 , 9 ( 2 ): 113 - 125 . DOI: 10.23919/jcin.2024.10582829 http://dx.doi.org/10.23919/jcin.2024.10582829

Tang F X , Wang X N , Yuan X , et al . MSADM: Large language model (LLM) assisted end-to-end network health management based on multi-scale semanticization [PP/OL ] . V4.arXiv ( 2026-03-23 )[ 2025-05-29 ] . https://doi.org/10.48550/arXiv.2406.08305 https://doi.org/10.48550/arXiv.2406.08305 .

Yao S , Zhao J , Yu D , et al . ReAct: Synergizing reasoning and acting in language models [C ] // The Eleventh International Conference on Learning Representations . 2023 .

Qin Y G , Tang J , Tang F X , et al . Multi-agent reinforcement learning in adversarial game environments: Personalized anti-interference strategies for heterogeneous UAV communication [J ] . IEEE Transactions on Mobile Computing , 2025 , 24 ( 9 ): 8886 - 8898 . DOI: 10.1109/tmc.2025.3559123 http://dx.doi.org/10.1109/tmc.2025.3559123

Zhang L , Wang B L , Zhao Y Q , et al . Collaborative multimodal fusion network for multiagent perception [J ] . IEEE Transactions on Cybernetics , 2025 , 55 ( 1 ): 486 - 498 . DOI: 10.1109/tcyb.2024.3491756 http://dx.doi.org/10.1109/tcyb.2024.3491756

Peigné P , Kniejski M , Sondej F , et al . Multi-agent security tax: Trading off security and collaboration capabilities in multi-agent systems [J ] . Proceedings of the AAAI Conference on Artificial Intelligence , 2025 , 39 ( 26 ): 27573 - 27581 . DOI: 10.1609/aaai.v39i26.34970 http://dx.doi.org/10.1609/aaai.v39i26.34970

Iodice F , De Momi E , Ajoudani A . Intelligent framework for human-robot collaboration: Dynamic ergonomics and adaptive decision-making [J ] . Journal of Intelligent & Robotic Systems , 2026 , 112 : 5 . DOI: 10.1007/s10846-025-02341-1 http://dx.doi.org/10.1007/s10846-025-02341-1

Xu M R , Peng J L , Gupta B B , et al . Multiagent federated reinforcement learning for secure incentive mechanism in intelligent cyber-physical systems [J ] . IEEE Internet of Things Journal , 2022 , 9 ( 22 ): 22095 - 22108 . DOI: 10.1109/jiot.2021.3081626 http://dx.doi.org/10.1109/jiot.2021.3081626

Pacheco F , Exposito E , Gineste M , et al . Towards the deployment of machine learning solutions in network traffic classification: A systematic survey [J ] . IEEE Communications Surveys & Tutorials , 2019 , 21 ( 2 ): 1988 - 2014 . DOI: 10.1109/comst.2018.2883147 http://dx.doi.org/10.1109/comst.2018.2883147

Ren H S , Xu B X , Wang Y J , et al . Time-series anomaly detection service at microsoft [C ] // Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . New York : ACM , 2019 : 3009 - 3017 . DOI: 10.1145/3292500.3330680 http://dx.doi.org/10.1145/3292500.3330680

Tariq S , Lee S , Shin Y , et al . Detecting anomalies in space using multivariate convolutional LSTM with mixtures of probabilistic PCA [C ] // Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . New York : ACM , 2019 : 2123 - 2133 . DOI: 10.1145/3292500.3330776 http://dx.doi.org/10.1145/3292500.3330776

Yang Y Y , Zhang C L , Zhou T , et al . DCdetector: Dual attention contrastive representation learning for time series anomaly detection [C ] // Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . New York : ACM , 2023 : 3033 - 3045 . DOI: 10.1145/3580305.3599295 http://dx.doi.org/10.1145/3580305.3599295

Tuli S , Casale G , Jennings N R . TranAD: Deep transformer networks for anomaly detection in multivariate time series data [PP/OL ] . V6.arXiv ( 2022-05-14 )[ 2025-05-29 ] . https://doi.org/10.48550/arXiv.2201.07284 https://doi.org/10.48550/arXiv.2201.07284 .

Chen Y H , Zhang C Y , Ma M H , et al . ImDiffusion: Imputed diffusion models for multivariate time series anomaly detection [PP/OL ] . V2.arXiv ( 2023-11-14 )[ 2025-05-29 ] . https://doi.org/10.48550/arXiv.2307.00754 https://doi.org/10.48550/arXiv.2307.00754 .

Ma M , Xu J M , Wang Y , et al . AutoMAP: Diagnose your microservice-based web applications automatically [C ] // Proceedings of The Web Conference 2020 . New York : ACM , 2020 : 246 - 258 . DOI: 10.1145/3366423.3380111 http://dx.doi.org/10.1145/3366423.3380111

Li M J , Li Z Y , Yin K L , et al . Causal inference-based root cause analysis for online service systems with intervention recognition [C ] // Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . New York : ACM , 2022 : 3230 - 3240 . DOI: 10.1145/3534678.3539041 http://dx.doi.org/10.1145/3534678.3539041

Wu Y K , Zhang J T , Hu N , et al . MLDT: Multi-level decomposition for complex long-horizon robotic task planning with open-source large language model [C ] // Database Systems for Advanced Applications . Singapore : Springer , 2024 : 251 - 267 . DOI: 10.1007/978-981-97-5569-1_16 http://dx.doi.org/10.1007/978-981-97-5569-1_16

Qiao S F , Fang R N , Qiu Z S , et al . Benchmarking agentic workflow generation [PP/OL ] . V3.arXiv ( 2025-02-23 )[ 2025-05-29 ] . https://doi.org/10.48550/arXiv.2410.07869 https://doi.org/10.48550/arXiv.2410.07869 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research on SDN Network Automated Configuration Based on Large Language Models

Lightweight Intrusion Detection in the Edge Internet of Things with Dual-Teacher Distillation

Stacked Rotating Alamouti Code for Quad-Antenna Backscatter Tags

Recent Advances in Modulation Mechanisms of Polarity and Chirality of Magnetic Vortices and Magnetization Dynamic Properties

Test-Time Adaptation of Graph Neural Networks under Distribution Shifts

Related Author

YANG Xing-yuan

TIAN Le

YAO Ying

PAN Fan

HU Yu-xiang

TIAN Le

HU Yu-xiang

PAN Fan

Related Institution

University of Information Engineering

⁰