Metraipolis › archive › dc_bd62d4285de2

Mimir: Physics-Grounded LLM Agents for Long-Horizon Irrigation Control — arXiv:2610.02038

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.02038
License (as archived)
CC0-1.0
Archived
2026-10-06T12:06:51.435Z
SHA-256 of the archived text
e581ba7aa72c56f971c48b3ba2aec97a2298b56b62402de7939e592ca2a09c8d
Stable URL of this page
https://metraipolis.com/documento/dc_bd62d4285de2

How to cite

Mimir: Physics-Grounded LLM Agents for Long-Horizon Irrigation Control — arXiv:2610.02038 — arXiv (en), https://arxiv.org/abs/2610.02038 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bd62d4285de2 (Metraipolis archive dc_bd62d4285de2, sha256 e581ba7aa72c56f971c48b3ba2aec97a2298b56b62402de7939e592ca2a09c8d)
BibTeX
@misc{dcbd62d4285de2,
  title = {Mimir: Physics-Grounded LLM Agents for Long-Horizon Irrigation Control — arXiv:2610.02038},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.02038},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bd62d4285de2; sha256 e581ba7aa72c56f971c48b3ba2aec97a2298b56b62402de7939e592ca2a09c8d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.02038. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model (LLM) agents increasingly combine reasoning, tool use, and action, but most evidence comes from episodic tasks with relatively immediate feedback and reset failures. Long-running physical control operates in a different regime: actions alter future states, errors compound across decisions, and an agent must improve from experience without being allowed to rewrite the physical rules that make execution safe. We study this regime through irrigation, where daily decisions interact with soil-water dynamics over entire growing seasons. We present Mimir, a physics-grounded LLM agent organized around two repair timescales. At the fast timescale, a structured physical interface and deterministic simulator turn an LLM output into a proposal that we numerically check, revise, and subject to bounded deterministic action selection before execution. At the slow timescale, recurrent failure patterns are consolidated into persistent contextual principles that condition future proposals, while the physical model, evaluator, and execution constraints remain immutable. Under a common retrospective evaluator across multiple sites, crops, and years, Mimir attains the lowest reported aggregate control cost among the evaluated references and uses about 51% less irrigation than the historical schedule replay. The ablation study show higher control cost when forward simulation, verified revision, or persistent context is removed; model-scale and model-family studies show no monotonic gain from increasing LLM size. The resulting lesson show that persistent physical agents can combine semantic reasoning with bounded, evidence-driven self-improvement while reserving physical truth and actuator authority for explicit numerical mechanisms.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.02038 — 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 2026-10-06T10:21:27.995Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).