Metraipolis › archive › dc_9efcba9ed045
Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control — arXiv:2609.34695
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34695
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T21:02:50.961Z
- SHA-256 of the archived text
6cea7b29d8f60404808510c3cdee8569ab7391ab486ee9a376cc706d9fed8195- Stable URL of this page
https://metraipolis.com/documento/dc_9efcba9ed045
How to cite
Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control — arXiv:2609.34695 — arXiv (en), https://arxiv.org/abs/2609.34695 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_9efcba9ed045 (Metraipolis archive dc_9efcba9ed045, sha256 6cea7b29d8f60404808510c3cdee8569ab7391ab486ee9a376cc706d9fed8195)
BibTeX
@misc{dc9efcba9ed045,
title = {Sufficiency of Zeroth-Order Reward Shaping for Policy Gradient in Stabilization Control — arXiv:2609.34695},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34695},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_9efcba9ed045; sha256 6cea7b29d8f60404808510c3cdee8569ab7391ab486ee9a376cc706d9fed8195},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34695. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Reward shaping is fundamental to modern robotic control with deep reinforcement learning (RL), yet practitioners still rely heavily on heuristic principles borrowed from classical optimal control and trajectory optimization. Existing methods rarely distinguish reward terms that are intrinsic to the control objective from numerical regularizers, leading to brittle hyperparameter tuning. To determine which quantities a reward must contain, we study the stabilization control problem with a focus on zeroth-order (configuration) and first-order (velocity) information. We theoretically and empirically demonstrate that policy gradient methods can successfully solve stabilization tasks without first-order reward terms, adding such terms can instead introduce severe sensitivity as their scale grows. Conversely, our findings confirm that reward functions must be zeroth-order complete over goal-relevant coordinates, while the first-order state remains necessary in the policy observation under our low-dissipation assumptions. Overall, these results provide actionable and principled guidance for reward design in robotic RL. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.34695 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T20:29:03.134Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).