Adaptive Policy Alignment for Multi-Agent Large Language Models via Reinforcement Learning in Dynamic Task Environments
American Journal of Machine Learning · 01.04.2026 · 7 (1) · 173-179
DOI: 10.71465/ajml3665 ↗Kurz gefasst
Fachartikel in der Zeitschrift American Journal of Machine Learning, online erschienen am 01.04.2026. Verfasst von Oliver J. Thompson und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Large language model (LLM) agents have shown promising capabilities in collaborative decision-making, yet their performance often degrades under dynamically changing task constraints. This study proposes an adaptive policy alignment framework that integrates proximal policy optimization with prompt-conditioned policy modulation to enhance coordination among multiple LLM agents. The framework is evaluated on a synthetic decision-making benchmark comprising 12,000 task instances across three environments: resource allocation, negotiation and sequential planning. Experimental results show that the proposed method improves task success rate from 71.3% to 84.6% and reduces decision latency by 23.5% compared to static prompt-based agents. Furthermore, inter-agent consistency, measured via action divergence entropy, decreases by 31%, indicating improved collaborative coherence. The findings demonstrate that reinforcement learning-based alignment significantly enhances the adaptability and efficiency of LLM-driven multi-agent systems.
Zitieren
APA
Thompson, O. J., Hughes, D. P., & Clarke, A. R. (2026). Adaptive Policy Alignment for Multi-Agent Large Language Models via Reinforcement Learning in Dynamic Task Environments. American Journal of Machine Learning, 7(1), 173-179. https://doi.org/10.71465/ajml3665
Vancouver
Thompson OJ, Hughes DP, Clarke AR. Adaptive Policy Alignment for Multi-Agent Large Language Models via Reinforcement Learning in Dynamic Task Environments. American Journal of Machine Learning. 2026;7(1):173-179. doi:10.71465/ajml3665
BibTeX
@article{Thompson2026,
author = {Thompson, Oliver J. and Hughes, Daniel P. and Clarke, Amelia R.},
title = {Adaptive Policy Alignment for Multi-Agent Large Language Models via Reinforcement Learning in Dynamic Task Environments},
journal = {American Journal of Machine Learning},
year = {2026},
volume = {7},
number = {1},
pages = {173--179},
doi = {10.71465/ajml3665}
}
EUVIMED