EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Provable mean-field Langevin neural actor-critic algorithms for multi-objective reinforcement learning

Yichen Zhu, Huimin Gao, Junlong Zhu, Qingtao Wu, Meiyi Yang, Muhua Liu

Journal of King Saud University Computer and Information Sciences · 2026-11 · 38 (9)

DOI: 10.1007/s44443-026-01337-7 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of King Saud University Computer and Information Sciences, online erschienen am 28.10.2026. Verfasst von Yichen Zhu und 5 weiteren Autor:innen. Die Arbeit nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Henan Provincial Department of Science and Technology, Henan Provincial Department of Education, Science and Technology Department of Henan Province.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Actor-critic methods with theoretical convergence guarantees constitute an important class of algorithms for multi-objective reinforcement learning. To date, existing provable multi-objective actor-critic algorithms are limited to linear critic approximation and conventional stochastic actor updates. Consequently, a gap remains between the theoretical understanding of multi-objective actor-critic methods and the empirical success of deep neural actor-critic algorithms. To fill this gap, we propose a provable M ean- F ield L angevin M ulti- O bjective N eural A ctor- C ritic algorithm, referred to as MFL-MONAC, where mean-field Langevin dynamics is incorporated into both objective-wise critic evaluation and Pareto-guided actor improvement. Concretely, the critic performs regularized mean-field Langevin policy evaluation for each objective, while the actor uses MGDA-based gradient aggregation with momentum to construct a Pareto-guided update direction and applies mean-field Langevin dynamics to induce controlled particle-level exploration. We further establish a finite-time analysis for the MFL-MONAC algorithm, i.e., we show that the MFL-MONAC algorithm achieves ε -Pareto stationarity at a rate of $${{\mathcal{O}}}(1/\sqrt{T})$$ O ( 1 / T ) , where T is the number of iterations. Moreover, MFL-MONAC achieves a sufficient sample complexity of $${{\mathcal{O}}}({\varepsilon }^{-6})$$ O ( ε − 6 ) . Finally, we conduct experiments on representative MORL benchmarks to demonstrate the effectiveness of MFL-MONAC.

Förderung

  • Henan Provincial Department of Science and Technology
  • Henan Provincial Department of Education
  • Science and Technology Department of Henan Province

Zitieren

APA

Zhu, Y., Gao, H., Zhu, J., Wu, Q., Yang, M., & Liu, M. (2026). Provable mean-field Langevin neural actor-critic algorithms for multi-objective reinforcement learning. Journal of King Saud University Computer and Information Sciences, 38(9). https://doi.org/10.1007/s44443-026-01337-7

Vancouver

Zhu Y, Gao H, Zhu J, Wu Q, Yang M, Liu M. Provable mean-field Langevin neural actor-critic algorithms for multi-objective reinforcement learning. Journal of King Saud University Computer and Information Sciences. 2026;38(9). doi:10.1007/s44443-026-01337-7

BibTeX
@article{Zhu2026,
  author = {Zhu, Yichen and Gao, Huimin and Zhu, Junlong and Wu, Qingtao and Yang, Meiyi and Liu, Muhua},
  title = {Provable mean-field Langevin neural actor-critic algorithms for multi-objective reinforcement learning},
  journal = {Journal of King Saud University Computer and Information Sciences},
  year = {2026},
  volume = {38},
  number = {9},
  doi = {10.1007/s44443-026-01337-7}
}
Kostenlos starten →