Abstract:Operations Research (OR) relies on expert-driven modeling-a slow and fragile process ill-suited to novel scenarios. While large language models (LLMs) can automatically translate natural language into optimization models, existing approaches either rely on costly post-training or employ multi-agent frameworks, yet most still lack reliable collaborative error correction and task-specific retrieval, often leading to incorrect outputs. We propose MIRROR, a fine-tuning-free, end-to-end multi-agent framework that directly translates natural language optimization problems into mathematical models and solver code. MIRROR integrates two core mechanisms: (1) execution-driven iterative adaptive revision for automatic error correction, and (2) hierarchical retrieval to fetch relevant modeling and coding exemplars from a carefully curated exemplar library. Experiments show that MIRROR outperforms existing methods on standard OR benchmarks, with notable results on complex industrial datasets such as IndustryOR and Mamo-ComplexLP. By combining precise external knowledge infusion with systematic error correction, MIRROR provides non-expert users with an efficient and reliable OR modeling solution, overcoming the fundamental limitations of general-purpose LLMs in expert optimization tasks.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2602.03318 [cs.CL] |
| (or arXiv:2602.03318v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2602.03318 arXiv-issued DOI via DataCite |
Submission history
From: Yifan Shi [view email]
[v1]
Tue, 3 Feb 2026 09:46:56 UTC (1,651 KB)
[v2]
Wed, 4 Feb 2026 08:04:51 UTC (1,652 KB)
[v3]
Mon, 1 Jun 2026 09:24:08 UTC (1,824 KB)