Structure-Aware Skill Retrieval and Arbitration for Optimization Modeling
Abstract
Large Language Models (LLMs) can turn natural-language optimization tasks into solver-checked answers, but current agents remain brittle under narrative rewrites and often store reusable experience at the instance level. We introduce ArchOR, an archetype-centric skill agent that builds on OptSkills by adding two retrieval-focused modules. Centroid-Aware Representation Learning (CARL) refines the ordering of cached skill representations through a low-rank prototype residual. Signature-Guided Retrieval (S\textsuperscript{3}R) attaches open-dimensional structural signatures to skills, prioritizes candidates by weighted Jaccard overlap, and bounds semantic re-ranking to a top- set. A baseline-preserving arbitration rule determines when the structural route should replace the original selector. With the same DeepSeek-V4.1-Flash backbone and released skill library, ArchOR reaches 65.00% Micro-Avg.\ Pass@1 across five benchmarks, improving over OptSkills by 1.64pp Micro-Avg.\ and 2.68pp Macro-Avg.
Figures & tables
| Category | Models / Methods | Macro-Avg. | Micro-Avg. | Rank | OptiBench | Mamo.C | OptMATH | IndustryOR | ComplexOR |
| General Models | GPT-5.4 ( OpenAI 2026 ) | 51.71 | 57.82 | 7 | 69.09 | 47.39 | 46.39 | 29.00 | 66.67 |
| Gemini-3.1-Pro ( Google DeepMind 2026 ) | 53.58 | 57.88 | 6 | 66.56 | 46.45 | 54.22 | 34.00 | 66.67 | |
| Qwen3-235B ( Yang et al. 2025a ) | 42.63 | 52.36 | 8 | 63.80 | 41.71 | 39.76 | 29.00 | 38.89 | |
| DeepSeek-V3.2 ( DeepSeek-AI 2025 ) | 43.21 | 49.45 | 10 | 62.64 | 27.49 | 40.36 | 30.00 | 55.56 | |
| Agent-Based Methods | Chain-of-Experts ( Xiao et al. 2024 ) | 38.84 | 52.36 | 8 | 62.31 | 51.66 | 35.54 | 28.00 | 16.67 |
| OptiMUS ( AhmadiTeshnizi et al. 2024 ) | 34.14 | 45.18 | 14 | 60.99 | 27.49 | 22.89 | 26.00 | 33.33 |
| Configuration | Macro-Avg. | Micro-Avg. | OptiBench | Mamo.C | OptMATH | IndustryOR | ComplexOR |
| OptSkills | 57.93 | 63.36 | 71.24 | 53.08 | 65.66 | 33.00 | 66.67 |
| Direct CARL+S 3 R | 57.96 | 61.45 | 68.76 | 51.18 | 62.65 | 35.00 | 72.22 |
| ArchOR (ours) | 60.61 | 65.00 | 72.40 | 53.55 | 69.88 | 35.00 | 72.22 |
| Configuration | OptMATH Pass@1 (%) |
| OptSkills | 65.66 |
| CARL only | 61.45 |
| S 3 R only | 61.45 |
| Direct CARL+S 3 R | 62.65 |
| ArchOR (ours) | 69.88 |
| Representation | Hit@1 | Hit@3 | Hit@5 | MRR |
| Cached skill vectors | 0.607 | 0.732 | 0.821 | 0.682 |
| CARL | 0.804 | 0.893 | 0.893 | 0.839 |
| Case | Benchmark / trigger | Baseline structural |
| S1 | OptiBench #579; , | knapsack fixed-charge ✓ |
| S2 | IndustryOR #43; , | inventory dispatch ✓ |
| F1 | Mamo.C #133; , | transportation flow |
LLM Agent Skill Retrieval
Large language model agent skill retrieval finds and selects reusable procedures to help agents complete tasks.
See all 109 papers