Elamur Research⌕
☰
home / papers / 2026 / Dual-Cluster Memory Agent: Resolving Multi-Paradigm Ambiguity in Optimization Problem Solving

Dual-Cluster Memory Agent: Resolving Multi-Paradigm Ambiguity in Optimization Problem Solving

Paper · 2026 · Xinyu Zhang, Yuchen Wan, Boxuan Zhang et al.

Summary

LLMs that turn text descriptions of optimization problems into solver code often get confused when one problem fits several competing formulations. For example, the same task might suggest integer linear programming, constraint programming and dynamic programming at once. The authors propose DCM-Agent, a training-free agent that reuses past solutions instead of fine-tuning.

Memory construction. Trajectories from a separate pool of problems are labelled by outcome:

  • Type A: always correct
  • Type B: recovered after retries
  • Type C: persistent failure

These trajectories are grouped into independent Modeling and Coding Clusters, linked by a weighted bipartite graph. Each cluster is condensed into three kinds of guidance: Approach, Checklist and Pitfall.

Inference. When it encounters an unseen problem, the agent looks up the clusters that match it. It then runs a loop that generates a formulation and code, checks them, fixes errors and switches to another candidate path if needed.

Results. Across seven benchmarks, the agent beats standalone LLMs by 11–21% on average. Memory built with larger models also transfers well to smaller ones. This suggests compact models can handle operations-research modeling more cheaply.

Key points

  • Evaluation data: seven optimization benchmarks, including NL4Opt, OptiBench and NLP4LP, plus the authors' own set of 500 problems. Models used include the Qwen, DeepSeek and GPT families.
  • "Knowledge inheritance": memory distilled by large models helps small models (around 8B) reach state-of-the-art accuracy, often beating larger models without memory.
  • Ablation (Qwen3-235B): dropping the Modeling Clusters hurts more than dropping the Coding Clusters. Correct mathematical logic is the main driver of success.
  • Hyperparameters:
    • Best retrieval size K = 3 and update threshold N = 5.
    • Gains grow with more planning candidates M, but M = 3 was chosen for cost reasons.
  • Main limitation: building the memory is a non-trivial one-time cost. In a case study, a baseline reported a cheaper answer ($60.80 vs. $70.00) that was actually infeasible because it ignored integrality constraints.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Xi'an Jiaotong UniversityXi'an · China42 works · 36,431 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,685 works

Cite

Cite as Xinyu Zhang, Yuchen Wan, Boxuan Zhang, Zesheng Yang, Lingling Zhang, Bifan Wei and Jun Liu. 2026. Dual-Cluster Memory Agent: Resolving Multi-Paradigm Ambiguity in Optimization Problem Solving. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.266

@inproceedings{zhang2026dual,
  title = {{Dual-Cluster Memory Agent: Resolving Multi-Paradigm Ambiguity in Optimization Problem Solving}},
  author = {Xinyu Zhang, Yuchen Wan, Boxuan Zhang, Zesheng Yang, Lingling Zhang, Bifan Wei and Jun Liu},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.266},
  url = {https://doi.org/10.18653/v1/2026.acl-long.266}
}

References

39 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ali AhmadiTeshnizi, Wenzhi Gao, and Madeleine Udell. 2024. Optimus: Scalable optimization modeling with (mi) lp solvers and large language models. In Forty-first International Conference on Machine Learning. find
  2. Nicolás Astorga, Tennison Liu, Yuanzhang Xiao, and Mihaela van der Schaar. 2025. Autoformulation of mathematical optimization models using llms. In Forty-second International Conference on Machine Learning. find
  3. Mariam Barry, Gaëtan Caillaut, Pierre Halftermeyer, Raheel Qader, Mehdi Mouayad, Fabrice Le Deit, Dimitri Cariolaro, and Joseph Gesnouin. 2025. Graphrag: Leveraging graph-based efficiency to minimize hallucinations in llm-driven rag for finance data. In Proceedings of the Workshop on Generative AI a… find
  4. S Belil, S Kemmoé-Tchomté, and N Tchernev. 2018. Milp-based approach to mid-term production planning of batch manufacturing environment producing bulk products. IFAC-PapersOnLine, 51(11):1689– 1694. find
  5. Yitian Chen, Jingfan Xia, Siyu Shao, Dongdong Ge, and Yinyu Ye. 2025a. Solver-informed rl: Grounding large language models for authentic optimization modeling. The Thirty-ninth Annual Conference on Neural Information Processing Systems. find
  6. Yuanyao Chen, Rongsheng Chen, Fu Luo, and Zhenkun Wang. 2025b. Improving generalization of neural combinatorial optimization for vehicle routing problems via test-time projection learning. The Thirty-ninth Annual Conference on Neural Information Processing Systems. find
  7. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948.
  8. Chenyu Huang, Zhengyang Tang, Shixi Hu, Ruoqing Jiang, Xin Zheng, Dongdong Ge, Benyou Wang, and Zizhuo Wang. 2025a. Orlm: A customizable framework in training large models for automated optimization modeling. Operations Research. find
  9. Jie Huang, Xinyun Chen, Swaroop Mishra, Huaixiu Steven Zheng, Adams Wei Yu, Xinying Song, and Denny Zhou. 2024. Large language models cannot self-correct reasoning yet. In The Twelfth International Conference on Learning Representations.
  10. Lei Huang, Weijiang Yu, Weitao Ma, Weihong Zhong, Zhangyin Feng, Haotian Wang, Qianglong Chen, Weihua Peng, Xiaocheng Feng, Bing Qin, et al. 2025b. A survey on hallucination in large language models: Principles, taxonomy, challenges, and open questions. ACM Transactions on Information Systems, 43(2)…
  11. Xuhan Huang, Qingning Shen, Yan Hu, Anningzhe Gao, and Benyou Wang. 2025c. LLMs for mathematical modeling: Towards bridging the gap between natural and mathematical languages. In Findings of the Association for Computational Linguistics: NAACL 2025, pages 2678–2710, Albuquerque, New Mexico. Associat… find
  12. Caigao Jiang, Xiang Shu, Hong Qian, Xingyu Lu, JUN ZHOU, Aimin Zhou, and Yang Yu. 2025a. Llmopt: Learning to define and solve general optimization problems from scratch. In The Thirteenth International Conference on Learning Representations. find
  13. Xia Jiang, Yaoxin Wu, Minshuo Li, Zhiguang Cao, and Yingqian Zhang. 2025b. Large language models as end-to-end combinatorial optimization solvers. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. find
  14. Aixin Liu, Aoxue Mei, Bangcai Lin, Bing Xue, Bingxuan Wang, Bingzheng Xu, Bochao Wu, Bowei Zhang, Chaofan Lin, Chen Dong, et al. 2025a. Deepseekv3. 2: Pushing the frontier of open large language models. arXiv preprint arXiv:2512.02556.
  15. Chunyang Liu, Houzhen Ma, Hengxu Zhang, Xiaohan Shi, and Fang Shi. 2023. A milp-based battery degradation model for economic scheduling of power system. IEEE Transactions on Sustainable Energy, 14(2):1000–1009. find
  16. Fan Liu, Zherui Yang, Cancheng Liu, Tianrui Song, Xiaofeng Gao, and Hao Liu. 2025b. Mm-agent: Llm as agents for real-world mathematical modeling problem. The Thirty-ninth Annual Conference on Neural Information Processing Systems. find
  17. Haoyang Liu, Jie Wang, Yuyang Cai, Xiongwei Han, Yufei Kuang, and Jianye HAO. 2025c. Optitree: Hierarchical thoughts generation with tree search for llm optimization modeling. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. find
  18. Hongliang Lu, Zhonglin Xie, Yaoyu Wu, Can Ren, Yuxuan Chen, and Zaiwen Wen. 2025. Optmath: A scalable bidirectional data synthesis framework for optimization modeling. In Forty-second International Conference on Machine Learning. find
  19. Mark Meerschaert. 2013. Mathematical modeling. Academic press. find
  20. OpenAI. 2024. Learning to reason with LLMs. find
  21. OpenAI. 2025. Gpt-5.1: A smarter, more conversational chatgpt. find
  22. Rindranirina Ramamonjison, Timothy Yu, Raymond Li, Haley Li, Giuseppe Carenini, Bissan Ghaddar, Shiqi He, Mahdi Mostajabdaveh, Amin Banitalebi-Dehkordi, Zirui Zhou, et al. 2023. Nl4opt competition: Formulating optimization problems based on their natural language descriptions. In NeurIPS 2022 compet… find
  23. Ankur Sinha, Shobhit Arora, and Dhaval Pujara. 2025. Autoopt: A dataset and a unified framework for automating optimization problem solving. In The Thirty-ninth Annual Conference on Neural Information Processing Systems Datasets and Benchmarks Track. find
  24. Yang Wang and Kai Li. 2025. Large language models and operations research: A structured survey. arXiv e-prints, pages arXiv–2509. find
  25. Zhiyuan Wang, Bokui Chen, Yinya Huang, Qingxing Cao, Ming He, Jianping Fan, and Xiaodan Liang. 2025. ORMind: A cognitive-inspired end-to-end reasoning framework for operations research. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 6: Industry Tra… find
  26. Zihao Wang, Anji Liu, Haowei Lin, Jiaqi Li, Xiaojian Ma, and Yitao Liang. 2024. Rat: Retrieval augmented thoughts elicit context-aware reasoning and verification in long-horizon generation. In NeurIPS 2024 Workshop on Open-World Agents. find
  27. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824–24837.
  28. Ziyang Xiao, Dongxiang Zhang, Yangjun Wu, Lilin Xu, Yuan Jessica Wang, Xiongwei Han, Xiaojin Fu, Tao Zhong, Jia Zeng, Mingli Song, et al. 2023. Chain-of-experts: When llms meet complex operations research problems. In The twelfth international conference on learning representations. find
  29. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, et al. 2025a. Qwen3 technical report. arXiv preprint arXiv:2505.09388.
  30. Zhicheng Yang, Yiwei Wang, Yinya Huang, Zhijiang Guo, Wei Shi, Xiongwei Han, Liang Feng, Linqi Song, Xiaodan Liang, and Jing Tang. 2025b. Optibench meets resocratic: Measure and improve llms for optimization modeling. In The Thirteenth International Conference on Learning Representations. find
  31. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R Narasimhan, and Yuan Cao. 2022. React: Synergizing reasoning and acting in language models. In The eleventh international conference on learning representations.
  32. Bowen Zhang and Pengcheng Luo. 2025. Or-llm-agent: Automating modeling and solving of operations research optimization problem with reasoning large language model. arXiv preprint arXiv:2503.10009. — [arXiv:2503.10009](https://arxiv.org/abs/2503.10009) find
  33. Lingling Zhang, Xiaojun Chang, Jun Liu, Minnan Luo, Zhihui Li, Lina Yao, and Alex Hauptmann. 2022. Tn-zstad: Transferable network for zero-shot temporal activity detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3):3848– 3861. find
  34. Xinyu Zhang, Yuxuan Dong, Yanrui Wu, Jiaxing Huang, Chengyou Jia, Basura Fernando, Mike Zheng Shou, Lingling Zhang, and Jun Liu. 2025a. PhysReason: A comprehensive benchmark towards physics-based reasoning. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (V…
  35. Xinyu Zhang, Yuxuan Dong, Lingling Zhang, Chengyou Jia, Zhuohang Dang, Basura Fernando, Jun Liu, and Mike Zheng Shou. 2025b. Cofft: Chain of foresight-focus thought for visual language models. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. find
  36. Xinyu Zhang, Lingling Zhang, Yanrui Wu, Muye Huang, and Jun Liu. 2025c. Cognitive predictive coding network: Rethinking the generalization in raven’s progressive matrices. In Proceedings of the 33rd ACM International Conference on Multimedia, pages 4097–4106. find
  37. Xinyu Zhang, Lingling Zhang, Yanrui Wu, Muye Huang, Wenjun Wu, Bo Li, Shaowei Wang, Basura Fernando, and Jun Liu. 2025d. Diagram-driven course questions generation. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 5995–6010. find
  38. Xinyu Zhang, Lingling Zhang, Yanrui Wu, Shaowei Wang, Wenjun Wu, Muye Huang, Qianying Wang, and Jun Liu. 2025e. Memory-enriched thought-by-thought framework for complex diagram question answering. Computer Vision and Image Understanding, page 104608. find
  39. Jie Zhao and Kang Hao Cheong. 2025. Structure-aware cooperative ensemble evolutionary optimization on combinatorial problems with multimodal large language models. In The Thirty-ninth Annual Conference on Neural Information Processing Systems. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhang-2026-dual-cluster-memory-agent-resolving/