EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Visualizing critic match loss landscapes for interpretation of online reinforcement learning control algorithms

Jingyi Liu, Jian Guo, Eberhard Gill

Acta Astronautica · 2026-09 · 246 · 909-920

DOI: 10.1016/j.actaastro.2026.04.045 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Acta Astronautica, 2026-09. Verfasst von Jingyi Liu und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 22 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Reinforcement learning has proven its power on various occasions. However, its performance is not always guaranteed when system dynamics change. Instead, it largely relies on users'empirical experience. For reinforcement learning algorithms with an actor-critic structure, the critic neural network reflects the approximation and optimization process in the RL algorithm. Analyzing the performance of the critic neural network helps to understand the mechanism of the algorithm. To support systematic interpretation of such algorithms in dynamic control problems, this work proposes a critic match loss landscape visualization method for online reinforcement learning. The method constructs a loss landscape by projecting recorded critic parameter trajectories onto a low-dimensional linear subspace. The critic match loss is evaluated over the projected parameter grid using fixed reference state samples and temporal-difference targets. This yields a three-dimensional loss surface together with a two-dimensional optimization path that characterizes critic learning behavior. To extend analysis beyond visual inspection, quantitative landscape indices and a normalized system performance index are introduced, enabling structured comparison across different training outcomes. The approach is demonstrated using the Action-Dependent Heuristic Dynamic Programming algorithm on cart-pole and spacecraft attitude control tasks. Comparative analyses across projection methods and training stages reveal distinct landscape characteristics associated with stable convergence and unstable learning. The proposed framework enables both qualitative and quantitative interpretation of critic optimization behavior in online reinforcement learning.

Zitieren

APA

Liu, J., Guo, J., & Gill, E. (2026). Visualizing critic match loss landscapes for interpretation of online reinforcement learning control algorithms. Acta Astronautica, 246, 909-920. https://doi.org/10.1016/j.actaastro.2026.04.045

Vancouver

Liu J, Guo J, Gill E. Visualizing critic match loss landscapes for interpretation of online reinforcement learning control algorithms. Acta Astronautica. 2026;246:909-920. doi:10.1016/j.actaastro.2026.04.045

BibTeX
@article{Liu2026,
  author = {Liu, Jingyi and Guo, Jian and Gill, Eberhard},
  title = {Visualizing critic match loss landscapes for interpretation of online reinforcement learning control algorithms},
  journal = {Acta Astronautica},
  year = {2026},
  volume = {246},
  pages = {909--920},
  doi = {10.1016/j.actaastro.2026.04.045}
}
Kostenlos starten →