EUVIMEDEUVIMED entdecken
Crossref

OptimAI: Optimization from Natural Language Using LLM-Powered AI Agents

Raghav Thind, Youran Sun, Ling Liang, Haizhao Yang

Journal of Machine Learning · 29.09.2026

DOI: 10.4208/jml.260208 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Machine Learning, online erschienen am 29.09.2026. Verfasst von Raghav Thind und 3 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 46 Literaturangaben. Gefördert von: US National Science Foundation, Office of Naval Research, DOE.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Optimization plays a vital role in scientific research and practical applications. However, translating a concrete optimization problem described in natural language into a mathematical formulation and selecting a suitable solver require substantial domain expertise. We introduce OptimAI, a framework for solving optimization problems described in natural language by leveraging LLM-powered AI agents, and achieve superior performance over current state-of-the-art methods. Our framework is built upon the following key roles: (1) a formulator that translates natural language problem descriptions into mathematical formulations; (2) a planner that constructs a high-level solution strategy prior to execution; and (3) a coder and a code critic capable of interacting with the environment and reflecting to refine future actions. Ablation studies confirm that all roles are essential; removing the planner or code critic results in $5.8\times$ and $3.1\times$ drops in productivity, respectively. Furthermore, we introduce UCB-based debug scheduling to dynamically switch between alternative plans, yielding an additional $3.3\times$ productivity gain. Our design emphasizes multi-agent collaboration, and our experiments confirm that combining diverse models leads to performance gains. The best OptimAI configurations attain 88.1% accuracy on the NLP4LP dataset and 82.3% on the Optibench dataset, reducing error rates by 58% and 52%, respectively, over prior best results.

Förderung

  • US National Science Foundation
  • Office of Naval Research
  • DOE
  • DARPA

Zitieren

APA

Thind, R., Sun, Y., Liang, L., & Yang, H. (2026). OptimAI: Optimization from Natural Language Using LLM-Powered AI Agents. Journal of Machine Learning. https://doi.org/10.4208/jml.260208

Vancouver

Thind R, Sun Y, Liang L, Yang H. OptimAI: Optimization from Natural Language Using LLM-Powered AI Agents. Journal of Machine Learning. 2026. doi:10.4208/jml.260208

BibTeX
@article{Thind2026,
  author = {Thind, Raghav and Sun, Youran and Liang, Ling and Yang, Haizhao},
  title = {OptimAI: Optimization from Natural Language Using LLM-Powered AI Agents},
  journal = {Journal of Machine Learning},
  year = {2026},
  doi = {10.4208/jml.260208}
}
Kostenlos starten →