Decentralized Collaborative Reasoning with Communication-Constrained Multi-Agent Large Language Models for Multi-Domain Dialogue Systems

Large language model (LLM)-based multi-agent systems show strong potential for supporting complex reasoning and decision-making in dialogue tasks. However, existing systems often rely on centralized coordination or uncontrolled inter-agent communication, which can limit scalability and increase communication overhead in multi-domain task-oriented dialogue environments. In this paper, we propose a decentralized multi-agent framework that combines role-based collaboration with communication-efficient coordination, enabling agents to operate effectively under constrained token budgets. The proposed approach aims to improve scalability, reduce communication cost, and enhance task success across diverse dialogue domains. Experimental results demonstrate that the proposed method outperforms single-agent and centralized coordination baselines, achieving a 15% improvement in task success rate and a 20% reduction in communication cost. These findings indicate that communication-efficient decentralized coordination significantly enhances both efficiency and robustness. Overall, the proposed architecture provides a practical and scalable solution for multi-domain task-oriented dialogue systems, particularly in resource-constrained environments.

Decentralized multi-agent systems, Dialogue systems, Large language models, Multi-domain dialogue, Role-based collaboration, Scalable architectures.

1. N. S. Shashidhara, “How do machines understand language? A beginner’s guide to natural language processing,” World J. Adv. Res. Rev., vol. 26, no. 2, pp. 1691–1699, 2025, doi: 10.30574/wjarr.2025.26.2.1811. [Google Scholar] [Crossref]

2. N. S. Shashidhara, “How do machines understand language? A beginner’s guide to natural language processing,” World J. Adv. Res. Rev., vol. 26, no. 2, pp. 1691–1699, 2025, doi: 10.30574/wjarr.2025.26.2.1811. [Google Scholar] [Crossref]

3. S. Hu et al., “Multi3WOZ: A Multilingual, Multi-Domain, Multi-Parallel Dataset for Training and Evaluating Culturally Adapted Task-Oriented Dialog Systems,” Trans. Assoc. Comput. Linguist., vol. 11, pp. 1396–1415, Nov. 2023, doi: 10.1162/tacl_a_00609. [Google Scholar] [Crossref]

4. H. Lu, C. Yuan, and X. Wang, “STOD: Towards Scalable Task-Oriented Dialogue System on MultiWOZ-API,” Appl. Sci., vol. 14, no. 12, p. 5303, Jan. 2024, doi: 10.3390/app14125303. [Google Scholar] [Crossref]

5. T. Adimulam, “Enhancing Conversational AI with Generative Models: Exploring Advanced Dialogue Management in Chatbots,” vol. 13, no. 11, 2024. [Google Scholar] [Crossref]

6. H. Su, R. Fang, L. Jiang, X. Huang, and S. Song, “Multi-domain dialogue state tracking based on dynamic knowledge fusion,” Vicinagearth, vol. 3, no. 1, p. 6, Mar. 2026, doi: 10.1007/s44336-026-00037-0. [Google Scholar] [Crossref]

7. H. Qian et al., “MemoRAG: Boosting Long Context Processing with Global Memory-Enhanced Retrieval Augmentation,” in Proceedings of the ACM on Web Conference 2025, in WWW ’25. New York, NY, USA: Association for Computing Machinery, Apr. 2025, pp. 2366–2377. doi: 10.1145/3696410.3714805. [Google Scholar] [Crossref]

8. D. Maldonado, E. Cruz, J. Abad Torres, P. J. Cruz, and S. del P. Gamboa Benitez, “Multi-Agent Systems: A Survey About Its Components, Framework and Workflow,” IEEE Access, vol. 12, pp. 80950–80975, 2024, doi: 10.1109/ACCESS.2024.3409051. [Google Scholar] [Crossref]

9. S. Rasal, “LLM Harmony: Multi-Agent Communication for Problem Solving,” Jan. 02, 2024, arXiv: arXiv:2401.01312. doi: 10.48550/arXiv.2401.01312. [Google Scholar] [Crossref]

10. Z. Xu et al., “Towards Reasoning in Large Language Models via Multi-Agent Peer Review Collaboration,” Dec. 17, 2023, arXiv: arXiv:2311.08152. doi: 10.48550/arXiv.2311.08152. [Google Scholar] [Crossref]

11. X. Tang et al., “MedAgents: Large Language Models as Collaborators for Zero-shot Medical Reasoning,” Jun. 04, 2024, arXiv: arXiv:2311.10537. doi: 10.48550/arXiv.2311.10537. [Google Scholar] [Crossref]

12. J. Becker, “Multi-Agent Large Language Models for Conversational Task-Solving,” Nov. 01, 2024, arXiv: arXiv:2410.22932. doi: 10.48550/arXiv.2410.22932. [Google Scholar] [Crossref]

13. Y. Yang et al., “AgentNet: Decentralized Evolutionary Coordination for LLM-based Multi-Agent Systems,” May 29, 2025, arXiv: arXiv:2504.00587. doi: 10.48550/arXiv.2504.00587. [Google Scholar] [Crossref]

14. C. Qian et al., “Scaling Large Language Model-based Multi-Agent Collaboration,” Mar. 17, 2025, arXiv: arXiv:2406.07155. doi: 10.48550/arXiv.2406.07155. [Google Scholar] [Crossref]

15. A. Jin, Y. Ye, B. Lee, and Y. Qiao, “DeCoAgent: Large Language Model Empowered Decentralized Autonomous Collaboration Agents Based on Smart Contracts,” IEEE Access, vol. 12, pp. 155234–155245, 2024, doi: 10.1109/ACCESS.2024.3481641. [Google Scholar] [Crossref]

16. H. Zhang et al., “Building Cooperative Embodied Agents Modularly with Large Language Models,” Feb. 17, 2024, arXiv: arXiv:2307.02485. doi: 10.48550/arXiv.2307.02485. [Google Scholar] [Crossref]

17. F. Jiang et al., “Large Language Model Enhanced Multi-Agent Systems for 6G Communications,” IEEE Wirel. Commun., vol. 31, no. 6, pp. 48–55, Dec. 2024, doi: 10.1109/MWC.016.2300600. [Google Scholar] [Crossref]

18. Y. Talebirad and A. Nadiri, “Multi-Agent Collaboration: Harnessing the Power of Intelligent LLM Agents,” Jun. 05, 2023, arXiv: arXiv:2306.03314. doi: 10.48550/arXiv.2306.03314. [Google Scholar] [Crossref]

19. G. Zhang et al., “Cut the Crap: An Economical Communication Pipeline for LLM-based Multi-Agent Systems,” Oct. 03, 2024, arXiv: arXiv:2410.02506. doi: 10.48550/arXiv.2410.02506. [Google Scholar] [Crossref]

20. Y. Yu et al., “FINCON: A Synthesized LLM Multi-Agent System with Conceptual Verbal Reinforcement for Enhanced Financial Decision Making”. [Google Scholar] [Crossref]

21. P. He, Y. Lin, S. Dong, H. Xu, Y. Xing, and H. Liu, “Red-Teaming LLM Multi-Agent Systems via Communication Attacks,” in Findings of the Association for Computational Linguistics: ACL 2025, W. Che, J. Nabende, E. Shutova, and M. T. Pilehvar, Eds., Vienna, Austria: Association for Computational Linguistics, Jul. 2025, pp. 6726–6747. doi: 10.18653/v1/2025.findings-acl.349. [Google Scholar] [Crossref]

22. S. S. Kannan, V. L. N. Venkatesh, and B.-C. Min, “SMART-LLM: Smart Multi-Agent Robot Task Planning using Large Language Models,” Mar. 23, 2024, arXiv: arXiv:2309.10062. doi: 10.48550/arXiv.2309.10062. [Google Scholar] [Crossref]

23. “MultiWOZ 2.2.” Accessed: Apr. 07, 2026. [Online]. Available: https://www.kaggle.com/datasets/taejinwoo/multiwoz-22 [Google Scholar] [Crossref]

• The Role of Artificial Intelligence in Revolutionizing Library Services in Nairobi: Ethical Implications and Future Trends in User Interaction
• ESPYREAL: A Mobile Based Multi-Currency Identifier for Visually Impaired Individuals Using Convolutional Neural Network
• Comparative Analysis of AI-Driven IoT-Based Smart Agriculture Platforms with Blockchain-Enabled Marketplaces
• AI-Based Dish Recommender System for Reducing Fruit Waste through Spoilage Detection and Ripeness Assessment
• SEA-TALK: An AI-Powered Voice Translator and Southeast Asian Dialects Recognition