OptiRepair: Closed-Loop Diagnosis and Repair of Supply Chain Optimization Models with LLM Agents

2026-02-22

AI Summary

This work proposes the first large language model (LLM)-based agent framework for automatically repairing infeasible supply chain optimization models, which often arise from modeling errors and traditionally require scarce operations research expertise to fix. The approach decomposes repair into two stages: a general feasibility phase that iteratively corrects linear constraints using an Irreducible Infeasible Set (IIS), and a domain validation phase that enforces five inventory-theoretic reasonableness checks. A novel self-teaching reasoning training mechanism is introduced, integrating solver feedback with verifiable operational rationality constraints. Experimental results demonstrate that the trained 8B-parameter model achieves a 97.2% success rate in restoring feasibility and an 81.7% rationality recovery rate, substantially outperforming existing API-based models, which average 21.3% and reach at best 42.2%.

Technology Category

• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Search and Optimization: Metareasoning and Metaheuristics
• Planning, Routing, and Scheduling: Replanning and Plan Repair

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Problem Definition. Supply chain optimization models frequently become infeasible because of modeling errors. Diagnosis and repair require scarce OR expertise: analysts must interpret solver diagnostics, trace root causes across echelons, and fix formulations without sacrificing operational soundness. Whether AI agents can perform this task remains untested. Methodology/Results. OptiRepair splits this task into a domain-agnostic feasibility phase (iterative IIS-guided repair of any LP) and a domain-specific validation phase (five rationality checks grounded in inventory theory). We test 22 API models from 7 families on 976 multi-echelon supply chain problems and train two 8B-parameter models using self-taught reasoning with solver-verified rewards. The trained models reach 81.7% Rational Recovery Rate (RRR) -- the fraction of problems resolved to both feasibility and operational rationality -- versus 42.2% for the best API model and 21.3% on average. The gap concentrates in Phase 1 repair: API models average 27.6% recovery rate versus 97.2% for trained models. Managerial Implications. Two gaps separate current AI from reliable model repair: solver interaction (API models restore only 27.6% of infeasible formulations) and operational rationale (roughly one in four feasible repairs violate supply chain theory). Each requires a different intervention: solver interaction responds to targeted training; operational rationale requires explicit specification as solver-verifiable checks. For organizations adopting AI in operational planning, formalizing what "rational" means in their context is the higher-return investment.

Problem

• supply chain optimization
• model infeasibility
• diagnosis and repair
• operational rationality
• OR expertise

Innovation

• LLM agents
• supply chain optimization
• infeasibility repair
• self-taught reasoning
• rational recovery rate