EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Graph-Based Policy Optimization for Collaborative Agents under Reinforcement Learning

Wei Zhang, Jun Li, Yong Tan

Global Media and Social Sciences Research Journal · 05.04.2026 · 7 (1) · 304-310

DOI: 10.71465/gmssrj177 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Global Media and Social Sciences Research Journal, online erschienen am 05.04.2026. Verfasst von Wei Zhang und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study introduces a graph-based reinforcement learning framework to model interaction structures among collaborative agents in complex decision environments. Each agent is represented as a node in a dynamic interaction graph, where edge weights capture communication intensity and dependency relationships. The framework employs graph attention networks to update agent policies, combined with policy gradient methods for optimization. The approach is evaluated on a dataset containing 8,900 collaborative decision instances derived from logistics planning and distributed scheduling problems. Results show a 22.4% improvement in overall task efficiency and a 19.7% reduction in communication redundancy compared to non-graph-based approaches. The findings suggest that explicitly modeling interaction topology can significantly enhance coordination performance.

Zitieren

APA

Zhang, W., Li, J., & Tan, Y. (2026). Graph-Based Policy Optimization for Collaborative Agents under Reinforcement Learning. Global Media and Social Sciences Research Journal, 7(1), 304-310. https://doi.org/10.71465/gmssrj177

Vancouver

Zhang W, Li J, Tan Y. Graph-Based Policy Optimization for Collaborative Agents under Reinforcement Learning. Global Media and Social Sciences Research Journal. 2026;7(1):304-310. doi:10.71465/gmssrj177

BibTeX
@article{Zhang2026,
  author = {Zhang, Wei and Li, Jun and Tan, Yong},
  title = {Graph-Based Policy Optimization for Collaborative Agents under Reinforcement Learning},
  journal = {Global Media and Social Sciences Research Journal},
  year = {2026},
  volume = {7},
  number = {1},
  pages = {304--310},
  doi = {10.71465/gmssrj177}
}
Kostenlos starten →