ReLoop: Structured Modeling and Behavioral Verification for Reliable LLM-Based Optimization

TL;DR

ReLoop enhances the reliability of LLM-generated optimization code by combining structured reasoning and behavioral verification, significantly reducing errors and improving accuracy across benchmarks.

Contribution

It introduces a novel two-mechanism approach—structured generation and behavioral verification—to address semantic errors in LLM-produced optimization code.

Findings

Structured generation improves accuracy on compositional problems (+8.5pp).

Behavioral verification detects and corrects localized defects (+4.4pp).

ReLoop achieves 100% executable code and improves accuracy across benchmarks.

Abstract

Large language models (LLMs) can translate natural language into optimization code, but silent failures pose a critical risk: code that executes and returns solver-feasible solutions may encode semantically incorrect formulations -- a feasibility-correctness gap reaching 90 percentage points on compositional problems. We introduce ReLoop, which addresses this gap through two complementary mechanisms. Structured generation decomposes code production into a four-stage reasoning chain (understand, formalize, synthesize, verify), preventing formulation errors at their source. Behavioral verification detects errors that survive generation by testing whether the formulation responds correctly to solver-based parameter perturbation -- an external semantic signal that bypasses LLM self-review and requires no ground truth. The two mechanisms are complementary by error structure: structured generation drives the largest gains on compositional problems (+8.5pp accuracy on RetailOpt-190 with Claude Opus 4.6), while behavioral verification dominates on localized defects (+4.4pp on MAMO-ComplexLP, its largest contribution across benchmarks). Combined with diagnostic execution recovery, ReLoop reaches 100% executable code on Claude Opus 4.6 and consistently improves accuracy on chat-tuned foundation models across three benchmarks; we further identify a known limitation of narrowly-tuned SFT models, whose learned output formats are brittle to chain-of-thought prompts -- an interaction we document and analyze. We release RetailOpt-190, 190 compositional retail optimization scenarios targeting the multi-constraint interactions where LLMs most frequently fail.