Fix What's Wrong, Keep What's Right: Hallucination Repair for LLM-Based Optimization Modeling
Abstract
Large language models (LLMs) are increasingly being explored to automate operations research modeling, reduce modeling time and lower the expertise barrier to optimization based decision support. However, this process often introduces inconsistencies between generated artifacts and the original problem, which we refer to as optimization modeling hallucinations. Despite progress in general-purpose hallucination mitigation, optimization modeling poses a distinct repair challenge: restoring fidelity to the original problem through coordinated model-code updates. We introduce OPT-Kraken, a training-free hallucination repair framework. To the best of our knowledge, it is the first post-hoc framework for repairing hallucinations in LLM-based optimization modeling artifacts. It extracts problem requirements and instance data, localizes errors by checking problem-model fidelity, model-code consistency, and data grounding, and selects repair granularity by the affected scope. For local repair, category rules and instance evidence are compiled into contracts specifying permitted edits, expected effects, and protected content. Experiments on established optimization modeling benchmarks show positive gains across automatic modeling frameworks and fine-tuned models. On zero-shot artifacts from multiple base models, OPT-Kraken effectively repairs optimization modeling hallucinations while preserving initially correct artifacts. These results demonstrate effective and controlled repair for LLM-based optimization modeling artifacts.
Related
- A Two-Stage Optimization Framework for Real-World Large-Scale University Course Timetabling Based on Orthogonalized Hierarchical Modeling and Conflict-Driven ALNS Repair
- Right, Wrong, or Unfinished: A Competing Risks Framework for Language-Model Reasoning under Token Budgets
- Ranked Right, Sized Wrong: Ensemble Spread Fails Silently at Streamflow Extremes
- ORSMT: Unleashing the Potential of Satisfiability Modulo Theories for Semantic Verification and Repair in LLM-Based Optimization
- Immersogeometric fluid-structure interaction modeling of the human mitral valve and transcatheter edge-to-edge repair
BibTeX
@misc{wu2026whats,
title = {{Fix What's Wrong, Keep What's Right: Hallucination Repair for LLM-Based Optimization Modeling}},
author = {Jiageng Wu and Chao Shen and Zihan Guo and Junyan Shao and Zongyan Zhang and Jie Song and Zhong Li and Mingyang Sun},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.12529},
eprint = {2610.12529},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.12529},
}
RIS
TY - UNPB
TI - Fix What's Wrong, Keep What's Right: Hallucination Repair for LLM-Based Optimization Modeling
AU - Jiageng Wu
AU - Chao Shen
AU - Zihan Guo
AU - Junyan Shao
AU - Zongyan Zhang
AU - Jie Song
AU - Zhong Li
AU - Mingyang Sun
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.12529
UR - https://arxiv.org/abs/2610.12529
AB - Large language models (LLMs) are increasingly being explored to automate operations research modeling, reduce modeling time and lower the expertise barrier to optimization based decision support. However, this process often introduces inconsistencies between generated artifacts and the original problem, which we refer to as optimization modeling hallucinations. Despite progress in general-purpose hallucination mitigation, optimization modeling poses a distinct repair challenge: restoring fidelity to the original problem through coordinated model-code updates. We introduce OPT-Kraken, a training-free hallucination repair framework. To the best of our knowledge, it is the first post-hoc framework for repairing hallucinations in LLM-based optimization modeling artifacts. It extracts problem requirements and instance data, localizes errors by checking problem-model fidelity, model-code consistency, and data grounding, and selects repair granularity by the affected scope. For local repair, category rules and instance evidence are compiled into contracts specifying permitted edits, expected effects, and protected content. Experiments on established optimization modeling benchmarks show positive gains across automatic modeling frameworks and fine-tuned models. On zero-shot artifacts from multiple base models, OPT-Kraken effectively repairs optimization modeling hallucinations while preserving initially correct artifacts. These results demonstrate effective and controlled repair for LLM-based optimization modeling artifacts.
ER -