LLM-Driven Multi-Agent Decision Systems with Reinforcement Learning-Based Adaptive Communication
Fundamental Scientific Reports in Multidisciplinary Areas · 05.04.2026 · 2 (1) · 51-58
DOI: 10.66238/fsrma54 ↗Kurz gefasst
Fachartikel in der Zeitschrift Fundamental Scientific Reports in Multidisciplinary Areas, online erschienen am 05.04.2026. Verfasst von James Smith und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Effective communication among LLM-driven agents is critical for achieving high-quality collaborative decision-making. This study proposes an adaptive communication framework that leverages reinforcement learning to dynamically adjust message frequency, content granularity and routing strategies among agents. The system is trained on a multi-domain dataset comprising 9,800 tasks across negotiation, planning and data reasoning scenarios. A communication-aware reward function is designed to balance performance gains and communication cost. Experimental results demonstrate a 25.6% improvement in task success rate and a 30.2% reduction in communication overhead compared to fixed communication protocols. The proposed approach highlights the importance of adaptive communication policies in large-scale agent systems..
Zitieren
APA
Smith, J., Brown, E., & Williams, D. (2026). LLM-Driven Multi-Agent Decision Systems with Reinforcement Learning-Based Adaptive Communication. Fundamental Scientific Reports in Multidisciplinary Areas, 2(1), 51-58. https://doi.org/10.66238/fsrma54
Vancouver
Smith J, Brown E, Williams D. LLM-Driven Multi-Agent Decision Systems with Reinforcement Learning-Based Adaptive Communication. Fundamental Scientific Reports in Multidisciplinary Areas. 2026;2(1):51-58. doi:10.66238/fsrma54
BibTeX
@article{Smith2026,
author = {Smith, James and Brown, Emily and Williams, Daniel},
title = {LLM-Driven Multi-Agent Decision Systems with Reinforcement Learning-Based Adaptive Communication},
journal = {Fundamental Scientific Reports in Multidisciplinary Areas},
year = {2026},
volume = {2},
number = {1},
pages = {51--58},
doi = {10.66238/fsrma54}
}
EUVIMED