A Survey of Multi-agent Systems based on Large Langueage Models

Chengze Deng; Shixing Li

doi:10.54691/6775er03

Authors

Chengze Deng
Shixing Li

DOI:

https://doi.org/10.54691/6775er03

Keywords:

Multi-Agent Systems(MAS); Large Language Models(LLM); Collaboration Models; Communication Module.

Abstract

Traditional MAS demonstrate significant limitations in handling dynamic environments and generalization capabilities. The emergence of LLM, however, has revolutionized the development of MAS through their ability to enhance individual agents' cognitive capacities while significantly improving inter-agent collaboration. This paper systematically examines the architecture and components of LLM-based MAS, detailing the collaborative mechanisms and communication frameworks. It concludes with an analysis of typical applications and current faced challenges, including insufficient decision-making interpretability and the lack of effective evaluation methodologies.

Downloads

Download data is not yet available.

References

[1] J. Zhang, Y. Wang, Z.L. Chen, et al.Multi-agent path finding algorithm, Journal of Zhejiang University:Engineering Science, vol.59(2025) No.8,p.1-9.

[2] P. Li, C.F. Chen, C.S. Yi, et al. Theories and methods of multi-agent decision under risk:a survey,Chinese Journal of Computers(2025),p.1-31.

[3] B. Luo, T.M. Hu, Y.H. Zhou, et al.Survey on multi agent reinforcement learning for control and decision-making, Acta Automatica Sinica, vol.51(2025) No.3, p.510-539.

[4] Z.R. Chen, Z.Y. Liu, L.P. Wan, et al. A review of multi-agent reinforcement learning theory and applications, Pattern Recognition and Artifical Intelligence,vol.37(2024) No.10, p.851-872.

[5] C.Y. Shan, S.F. Li, R.Y. Qi, et al. Fast task reassignment in multi-agent system emergency fault handling, Systems Engineering and Electronics(2025), p.1-18.

[6] J.H. Qin, L.C. Ma, M. Li, et al. Recent advances on multi-agent collaboration: A cross-perspective of game and control theory, Acta Automatica Sinica, vol.51(2025) No.3, p.489-509.

[7] Y.C. Li, Z.J. Liu, Y.T. Hong, et al. Multi-agent reinforcement learning based game: A survey, Acta Automatica Sinica, vol.51(2025) No.3, p.540-558.

[8] X. Wu, Y.H. Wang, B.H. Zhang, et al. Regional air defense and anti-missile weapon-target assignment based on multi-agent reinforcement learning(2025), p.1-12.

[9] S.H. Li, H. Liu, Z.M. Ren, et al. Task assignment in multi-agent games via reinforcement learning, Scientia Sinica Technologica, vol.55(2025) No.5, p.906-913.

[10] Y. Wang, S.H. Li, B.C. Xu, et al. Multi-agent planning recognition and game encirclement based on data driving, Transactions of Beijing Institute of Technology, vol.45(2025) No.5, p.521-530.

[11] R.N. Wang, Q. Dong: Multiagent game decision-making method based on the learning machanism, vol.46(2024) No.7, p.1251-1268.

[12] X.Y. Li, Y.J. Li: Priority-based replanning for multi-agent pathfinding with communication, Control Theory&Applications(2024), p.1-10.

[13] G.M. Mi, H. Zhang, J. Zhang, et al. Multi-agent cooperative confrontation with proximal policy optimization in urban environment, Journal on Communications, vol.46(2025) No.3, p.94-108.

[14] W.B. Gu, X. Zhou, J.G. Wang, et al. Multi-intelligent deep reinforcement learning algorithm for hybrid human-machine collaborative workshop scheduling, Computer Integrated Manufacturing Systems(2025), p. 1-29.

[15] R.N. Qi, Y.M. Quan, Y.N. Ni, et al. Comm-Cot: Standardized chain-of-thought communication framework for efficient LLM based multi-agent decision-making in real-time strategy games, 2025 IEEE 2nd International Coference on Electronics, Communications and Intelligent Science(ECIS)(Yueyang, China, 2025), p.1-8.

[16] J. Owotogbe:Assessing and enhancing the robustness of LLM-based multi-agent systems through chaos engineering, 2025 IEEE/ACM 4th International Conference on AI Engineering - Software Engineering for AI(CAIN)(Ottawa, ON, Canada,2025), p.250-252.

[17] Z.G. Yu, C.M. Lu, Q.Y. Liang, et al. Research on an LLM-empowered multi-agent integrated steel scheduling, Metallurgical Industry Automation, vol.49(2025) No.4, p.125-133.

[18] T.T. Shen: The research on legal question-answering method based on large language models with multi-knowledge bases and multi-agent systems(MS., Dissertation Submitted for Hangzhou Dianzi University, China 2025, p.27)

[19] T.T. Yang, P. Feng, Q.X. Guo, et al. AutoHMA-LLM: Efficient task coordination and execution in heterogeneous multi-agent systems using hybrid large language models, IEEE Transactions on Cognitive Communications and Networking, vol.11(2025) No.2, p.987-998.

[20] X.Y. Wu, J.Y. Diao, J.M. Deng. Research on the construction of AI-large-model-based multi-agent energy management system in the HVAC industry, Software Guide(2025), p.1-11.

[21] Z. Yuan: Research and implementation of multi-agent collaborative decision making method based on federated reinforcement learning(MS., Beijing University of Posts and Telecommunications, China 2024, p.57)

[22] X. Yuan, Y. Zhu, J.P. Qiang, et al. Clickbait detection via dual-layer multi-agent large language model, Computer Systems&Applications, vol.34(2025) No.5, p.116-123.

[23] Y.J. Han, W. Wang, Y.F. Zhang, et al. Research on collabarative decision-making method of large and small model under time varing environment, 1st National Conference on Large Language Models and Decision Intelligence(Hangzhou, China, 2024), p.262-271.

[24] Z.J. Ding, W.J. Zhang, T.Y. Xu, et al. Deepseek-based multi-intelligent search and quiz on food sagety laws and regulations, Journal of Chinese Institute of Food Science and Technology, vol.25(2025) No.6, p.449-464.

[25] P.L. Zhang, Y.X. Ma, H. Li, et al. Large language model driven multi-agents for network intent recognition framework, Journal of Chinese Computer Systems(2025), p.1-10.

[26] I. Chatzistefanidis, A. Leone and N. Nikaein. Maestro: LLM-driven collaborative automation of intent-based 6G networks, IEEE Networking Letters, vol.6(2024) No.4, p.227-231.

[27] Z. Duan and J. Wang. Enhancing multi-agent consensus through third-party LLM integration: Analyzing uncertainty and mitigating hallucinations in large language models, 2025 8th International Conference on Advanced Algorithms and Control Engineering(ICAACE)(Shanghai, China, 2025), p.2222-2227.

[28] Y. Sun, Y. Zheng, H.Y. Huang, et al. Multi-loop nested LLM-based multi-agent command and control processes, Journal of Command and Control, vol.10(2024) No.6, p.732-739.

[29] M. De Jesus, P. Sylvester, W. Clifford, et al. LLM-based multi-agent framework for troubleshooting distributed systems, 2025 IEEE Cloud Summit(Wasgington, DC, USA, 2025), p.110-115.

[30] Y.W. Zhang, H.R. Hong, D.S. Zhang, et al. A review of multi-agent-based digital twins and its application in industry, Control and Decision, vol.38(2023) No.8, p.2168-2182.

[31] Y.P. Zhao,J.M. Liang, B.Z. Chen, et al. Research progress and prospects of multi-agent large language models in agricultural applications, Smart Agricutural(2025), p.1-15.

[32] C. Sun, S. Huang, D. Pompili. LLM-based multi-agent decision-making: Challenges and future directions, IEEE Robotics and Automation Letters, vol.10(2025) No.6, p.5681-5688.

[33] T. Zou, X.L. Ding, H.S. Dai, et al. Research on swarm operations decision-making based on large language models, Journal od Naijing University of Aeronautics&Astronautics(Natural Science Edition)(2025), p.1-13.

[34] W.J. Zheng: A maintenance knowledge retrieval and multi-agent collaboration framework based on large language models(MS., Beijing University of Chemical Technology, China 2025, p.43)

[35] Z.N. Dong, Q.X. Zhang, J. Hu, et al. A multi-dimensional method for large language model-powered multi-ahent systems, Command Control&Simulation, vol.47(2025) No.2, p.121-131.