图书情报工作

图书情报工作 >

2026 , Vol. 70 >Issue 10: 108 - 121

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.10.009

研究论文

基于大语言模型的图书馆知识服务多智能体协作模型研究

  • 吴畑 ,
  • 胡宜君 ,
  • 郑冰
展开
  • 1南昌大学公共政策与管理学院 南昌 330031;
    2南昌大学江西省哲学社会科学重点研究基地数字素养与技能提升研究中心 南昌 330031
吴畑，讲师，博士，硕士生导师，E-mail:wutian@ncu.edu.cn;胡宜君，本科生;郑冰，硕士研究生。

收稿日期: 2025-04-01

  修回日期: 2025-08-02

  网络出版日期: 2026-05-07

基金资助

本文系江西省职业早期青年科技人才培养项目“大模型安全能力的评测与自主增强方法研究”（项目编号：20244BCE52007）与江西省自然科学基金项目“面向社交媒体内容验证的大语言模型水印技术研究”（项目编号：20252BAC200613）研究成果之一。

收起

Research on Multi-Agent Collaborative Model for Library Knowledge Services Based on Large Language Models

  • Wu Tian ,
  • Hu Yijun ,
  • Zheng Bing
Expand
  • 1 School of Public Policy and Management, Nanchang University, Nanchang 330031;
    2 Digital Literacy and Skills Enhancement Research Center, Jiangxi Province Philosophy and Social Science Key Research Base, Nanchang University, Nanchang 330031
Wu Tian, Lecturer, PhD, Master’s Supervisor, E-mail: wutian@ncu.edu.cn; Hu Yijun, Undergraduate; Zheng Bing, Master’s Candidate.

Received date: 2025-04-01

  Revised date: 2025-08-02

  Online published: 2026-05-07

Supported by

This work is supported by Jiangxi Provincial Early-Career Youth Science and Technology Talent Cultivation Project, titled “Research on Evaluation and Autonomous Enhancement Methods for Large Model Security Capabilities” (Grant No. 20244BCE52007) and the Jiangxi Provincial Natural Science Foundation Project, titled “Research on Large Language Model Watermarking Technology for Social Media Content Verification” (Grant No. 20252BAC200613).

Fold

摘要

[目的/意义]设计基于大语言模型的图书馆知识服务多智能体协作模型，实现模拟环境下的图书馆核心职能和图书馆资源动态更新。为图书馆构建智能化服务生态系统提供理论基础与实践框架。[方法/过程]构建基于多个智能体群体的图书馆协作模型，结合循证图书馆学决策机制与读者决策采购模型，实现图书馆核心行为的全流程。利用多智能体的角色专业化与交互序列化协同工作能力，构建能感知、自主决策、动态更新资源、持续优化的图书馆智慧服务。[结果/结论]基于大语言模型的图书馆知识服务多智能体协作模型实现智能检索、动态更新与推荐优化功能，加强文献资源与用户需求的连接，促使图书馆从静态资源向自适应知识系统转型，为数智时代图书馆知识服务转型提供实践模型。

关键词： 大语言模型; 多智能体协作; 循证图书馆学; 读者决策采购; 智慧图书馆

本文引用格式

吴畑 , 胡宜君 , 郑冰 . 基于大语言模型的图书馆知识服务多智能体协作模型研究[J]. 图书情报工作, 2026 , 70(10) : 108 -121 . DOI: 10.13266/j.issn.0252-3116.2026.10.009

Abstract

[Purpose/Significance] This study designs a multi-agent collaborative model for library knowledge services based on large language models to implement core library functions and dynamic resource updates in a simulated environment. The research establishes theoretical foundations and practical frameworks for intelligent service ecosystem construction in libraries. [Method/Process] The study constructs a collaborative library model comprising multiple specialized agents, integrating Evidence-Based Librarianship (EBL) decision-making mechanisms with Patron-Driven Acquisition (PDA) models to realize comprehensive implementation of core library operations. The model leverages role specialization and sequential interaction protocols among agents to develop library services capable of environmental perception, autonomous decision-making, resource dynamic updating, and continuous service optimization. [Result/Conclusion] The multi-agent collaborative model for library knowledge services based on large language models successfully implements intelligent retrieval functionality, dynamic collection updating, and recommendation optimization. This approach strengthens the interconnection between bibliographic resources and user information needs, facilitating library transformation from static resource repositories toward adaptive knowledge systems, thus providing a viable implementation model for knowledge service transformation in the digital intelligence era.

Key words： large language models; multi-agent collaboration; evidence-based librarianship; patron-driven acquisition; smart library

参考文献

[1] DAS A, CHEN S C, SHYU M L, et al. Enabling synergistic knowledge sharing and reasoning in large language models with collaborative multi-agents[C]//2023 IEEE 9th international conference on collaboration and Internet computing. Atlanta: IEEE, 2023: 92-98.
[2] LI H, CHONG Y, STEPPUTTIS S, et al. Theory of mind for multi-agent collaboration via large language models[C]//Proceedings of the 2023 conference on empirical methods in natural language processing. Singapore: Association for Computational Linguistics, 2023: 180-192.
[3] 中国政府网. 中共中央办公厅、国务院办公厅印发《关于推进实施国家文化数字化战略的意见》[EB/OL]. [2025-07-29]. https://www.gov.cn/xinwen/2022-05/22/content_5691759.htm.(The State Council of the People’s Republic of China. General Offices of the CPC Central Committee and The State Council issue “Opinions on promoting the implementation of the national culture digitization strategy”[EB/OL]. [2025-07-29]. https://www.gov.cn/xinwen/2022-05/22/content_5691759.htm.)
[4] 储节旺, 樊鑫鑫. 智能体在智慧图书馆中的应用：特征能力、体系框架与构建路径[J]. 图书情报工作, 2025, 69(14): 52-60. (CHU J W, FANG X X. Application of agents in smart libraries: feature capabilities, system framework, and development path[J]. Library and information service, 2025, 69(14): 52-60.)
[5] 杨杰, 张新鹤. 数智场域下图书馆智慧化转型升级图景与实现路径[J]. 图书馆工作与研究, 2024(6): 5-12, 42. (YANG J, ZHANG X H. Landscape and implementation paths of intelligent transformation and upgrading of libraries in the field of digital intelligence[J]. Library work and study, 2024(6): 5-12, 42.)
[6] 王毅, 董怡婷. 类ChatGPT人工智能在图书馆智慧服务中的应用与思考[J]. 图书馆理论与实践, 2023(6): 129-136. (WANG Y, DONG Y T. Application and reflection of ChatGPT-like artificial intelligence in library intelligence service[J]. Library theory and practice, 2023(6): 129-136.)
[7] TRAN K T, DAO D, NGUYEN M D, et al. Multi-agent collaboration mechanisms: A survey of LLMs[J]. arXiv preprint arXiv:2501.06322, 2025.
[8] LU J, PANG Z, XIAO M, et al. Merge, ensemble, and cooperate! A survey on collaborative strategies in the era of large language models[J]. arXiv preprint arXiv:2407.06089, 2024.
[9] MNIH V, KAVUKCUOGLU K, SILVER D, et al. Playing atari with deep reinforcement learning[J]. arXiv preprint arXiv:1312.5602, 2013.
[10] MNIH V, KAVUKCUOGLU K, SILVER D, et al. Human-level control through deep reinforcement learning[J]. Nature, 2015, 518(7540): 529-533.
[11] HONG S, ZHUGE M, CHEN J, et al. MetaGPT: Meta programming for a multi-agent collaborative framework[C]//The twelfth international conference on learning representations. Vienna: ICLR, 2024.
[12] QIAN C, LIU W, LIU H, et al. Chatdev: Communicative agents for software development[J]. arXiv preprint arXiv:2307.07924, 2023
[13] DUAN Z, WANG J. Exploration of llm multi-agent application implementation based on langgraph+ crewai[J]. arXiv preprint arXiv:2411.18241, 2024.
[14] WU Q, BANSAL G, ZHANG J, et al. AutoGen: Enabling next-gen LLM applications via multi-agent conversations[C]//First conference on language modeling. Philadelphia: COLM, 2024.
[15] CHEN W, SU Y, ZUO J, et al. Agentverse: Facilitating multi-agent collaboration and exploring emergent behaviors in agents[J]. arXiv preprint arXiv:2308.10848, 2023.
[16] DU Y, LI S, TORRALBA A, et al. Improving factuality and reasoning in language models through multiagent debate[C]//Forty-first international conference on machine learning. Vienna: PMLR, 2024.
[17] LI G, HAMMOUD H, ITANI H, et al. Camel: Communicative agents for “mind” exploration of large language model society[J]. Advances in neural information processing systems, 2023, 36: 51991-52008.
[18] PARK J S, O’BRIEN J, CAI C J, et al. Generative agents: Interactive simulacra of human behavior[C]//Proceedings of the 36th annual ACM symposium on user interface software and technology. San Francisco: ACM, 2023: 1-22.
[19] 李文江, 陈诗琴. AIMLBot智能机器人在实时虚拟参考咨询中的应用[J]. 现代图书情报技术, 2012, 28(7): 127-132. (LI W J, CHEN S Q. Application of AIMLBot intelligent robot in real-time virtual reference service[J]. New technology of library and information service, 2012, 28(7): 127-132.)
[20] 刘泽, 徐潇洁, 邵波. 基于多策略混合问答系统模型的图书馆咨询机器人的设计与应用[J]. 新世纪图书馆, 2022(5): 43-49. (LIU Z, XU X J, SHAO B. Design and application of library consultation robot based on multi-strategy mixed question answering system model[J]. New century library, 2022(5): 43-49.)
[21] ROBERTSON S, ZARAGOZA H. The probabilistic relevance framework: BM25 and beyond[J]. Foundations and trends in information retrieval, 2009, 3(4): 333-389.
[22] LIU Y, OTT M, GOYAL N, et al. Roberta: A robustly optimized bert pretraining approach[J]. arXiv preprint arXiv:1907.11692, 2019.
[23] CHIANG W L, LIU X, SI S, et al. Cluster-gcn: An efficient algorithm for training deep and large graph convolutional networks[C]//Proceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. Anchorage: ACM, 2019: 257-266.
[24] HE X, LIAO L, ZHANG H, et al. Neural collaborative filtering[C]//Proceedings of the 26th international conference on World Wide Web. Perth: ACM, 2017: 173-182.
[25] 王珏, 李涵霄. 加快推进智慧图书馆建设与智慧服务能力建设——“第二届智慧图书馆与智慧服务学术研讨会”会议综述[J]. 图书情报工作, 2023, 67(10): 148-151. (WANG J, LI H X. Accelerating the construction of smart libraries and smart service capabilities—summary of the ‘Second symposium on smart libraries and smart services’[J]. Library and information service, 2023, 67(10): 148-151.)
[26] 邵波, 王怡, 王进. 智慧图书馆建设进程与发展方向——以南京大学为例[J]. 图书情报工作, 2023, 67(13): 14-20. (SHAO B, WANG Y, WANG J. Research on the construction process and development direction of smart library: taking Nanjing University as an example[J]. Library and information service, 2023, 67(13): 14-20.)
[27] 王翼虎, 白海燕, 孟旭阳. 大语言模型在图书馆参考咨询服务中的智能化实践探索[J]. 情报理论与实践, 2023, 46(8): 96-103. (WANG Y H, BAI H Y, MENG X Y. An intelligent practical exploration of large language model in library reference consulting service[J]. Information studies: theory & application, 2023, 46(8): 96-103.).
[28] 张晓林. Library-Inside: AI赋能图书馆新质生产力的一种基础模型[J]. 中国图书馆学报, 2024, 50(3): 4-16. (ZHANG X L. Library-Inside: a foundation model for AI-empowered new quality productive forces for library services[J]. Journal of library science in China, 2024, 50(3): 4-16.)
[29] 汪聪, 张春红, 高楠, 等. 面向人工智能生成内容时代的图书馆服务适应与创新[J]. 情报理论与实践, 2024, 47(7): 17-24. (WANG C, ZHANG C H, GAO N, et al. Adaptation and innovation of library services in the AIGC era[J]. Information studies: theory & application, 2024, 47(7): 17-24.)
[30] 丁遒劲, 苏静. 面向生成式人工智能（AIGC）的图书馆信息资源建设优化策略研究[J]. 图书情报工作, 2024, 68(18): 23-31. (DING Q J, SU J. Research on the optimization strategies for library information resources construction oriented to the development of AIGC[J]. Library and information service, 2024, 68(18): 23-31.)
[31] KHALAF B A, ALHAMMADI N A M, MOSTAFA S A, et al. Development of a multi-agent approach for a smart library management system[C]//20214th international symposium on agents, multi-agent systems and robotics. Batu Pahat: IEEE, 2021: 64-69.
[32] 金佳丽, 潘晶, 张紫徽, 等. AI驱动的未来图书馆范式重构与创新路径[J]. 中国图书馆学报, 2025, 51(5): 67-78. (JIN J L, PAN J, ZHANG Z H, et al. AI-driven paradigm shift and innovation pathways for future library[J]. Journal of library science in China, 2025, 51(5): 67-78.)
[33] 刘璇. 循证图书馆学(EBL)的发展及对图书馆学的启示[J]. 图书馆杂志, 2009, 28(1): 22-25. (LIU X. The development of evidence-based librarianship and revelation to library science[J]. Library journal, 2009, 28(1): 22-25.)
[34] 刘华. 以读者为主导的文献资源建设——美国学术图书馆读者决策采购(PDA)研究[J]. 图书情报工作, 2012, 56(5): 89-93, 127. (LIU H. Patron-initiated collection development-a study of patron-driven acquisition in American academic libraries[J]. Library and information service, 2012, 56(5): 89-93, 127.)
[35] 张彦静. 我国图书馆读者自主采购的探索与思考——以佛山市图书馆“新书借阅处”为例[J]. 图书馆学研究, 2013(15): 54-58, 76. (ZHANG Y J. An exploration and thinking on readers’ independent acquisition in Chinese library: a case on “new books loan desk” in Foshan Library[J]. Research on library science, 2013(15): 54-58, 76.)
[36] 张一翠, 魏青山. 循证图书馆学在读者决策采购模型建立和实践中的应用研究——以西安交通大学为例[J]. 图书馆学研究, 2021(5): 53-59. (ZHANG Y C, WEI Q S. Application research of evidence-based library science in the establishment and practice of the patron-driven acquisition model: a case study of Xi’an Jiaotong University Library[J]. Research on library science, 2021(5): 53-59.)
[37] 胡钦文, 陈韵如. 国内读者决策采购实践研究述评[J]. 图书情报工作, 2016, 60(16): 137-144, 148. (HU Q W, CHEN Y R. Research review on patron driven acquisition (PDA) in China based on the practice perspective[J]. Library and information service, 2016, 60(16): 137-144, 148.)
[38] REIMERS N, GUREVYCH I. Sentence-BERT: sentence embeddings using siamese BERT-networks[C]//INUI K, JIANG J, NG V, et al. Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing. Hong Kong: Association for Computational Linguistics, 2019: 3982-3992.
[39] STAMPE L, STOCKDIEK J L, GRIMME B, et al. Benchmarking sentence embeddings in textual stream clustering with applications to campaign detection[C]//2024 international joint conference on neural networks. Yokohama: IEEE, 2024: 1-8.
[40] OUYANG L, WU J, JIANG X, et al. Training language models to follow instructions with human feedback[C]//Proceedings of the 36th international conference on neural information processing systems. Red Hook: Curran Associates, 2022: 27730-27744.
[41] OPENAI, HURST A, LERER A, et al. GPT-4o system card[EB/OL]. [2025-07-29]. https://arxiv.org/abs/2410.21276.
[42] GRATTAFIORI A, DUBEY A, JAUHRI A, et al. The llama 3 herd of models[EB/OL]. [2025-07-29]. https://arxiv.org/abs/2407.21783.
[43] TEAM G, GEORGIEV P, LEI V I, et al. Gemini 1.5: unlocking multimodal understanding across millions of tokens of context[EB/OL]. [2025-07-29]. https://arxiv.org/abs/2403.05530.
[44] ANTHROPIC. Model card addendum: Claude 3.5 haiku and upgraded claude 3.5 sonnet[EB/OL]. [2025-07-29]. https://www.anthropic.com/claude/haiku.
Options
文章导航

/

〈 〉