Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

From Large Language Models and Optimization to Decision Optimization CoPilot: A Research Manifesto (2402.16269v1)

Published 26 Feb 2024 in cs.AI, cs.LG, and math.OC

Abstract: Significantly simplifying the creation of optimization models for real-world business problems has long been a major goal in applying mathematical optimization more widely to important business and societal decisions. The recent capabilities of LLMs present a timely opportunity to achieve this goal. Therefore, we propose research at the intersection of LLMs and optimization to create a Decision Optimization CoPilot (DOCP) - an AI tool designed to assist any decision maker, interacting in natural language to grasp the business problem, subsequently formulating and solving the corresponding optimization model. This paper outlines our DOCP vision and identifies several fundamental requirements for its implementation. We describe the state of the art through a literature survey and experiments using ChatGPT. We show that a) LLMs already provide substantial novel capabilities relevant to a DOCP, and b) major research challenges remain to be addressed. We also propose possible research directions to overcome these gaps. We also see this work as a call to action to bring together the LLM and optimization communities to pursue our vision, thereby enabling much more widespread improved decision-making.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (60)
  1. INFORMS Journal on Applied Analytics. https://pubsonline.informs.org/loi/ijaa, 2024. Accessed: 2024-01-21.
  2. Optimus: Optimization modeling using MIP solvers and large language models. arXiv preprint arXiv:2310.06116, 2023.
  3. Almonacid, B. Towards an automatic optimisation model generator assisted with generative pre-trained transformer. arXiv preprint arXiv:2305.05811, 2023.
  4. AI-copilot for business optimisation: A framework and a case study in production scheduling. arXiv preprint arXiv:2309.13218, 2023.
  5. Neural machine translation by jointly learning to align and translate. In International Conference on Learning Representations (ICLR), 2015.
  6. Graph of thoughts: Solving elaborate problems with large language models. arXiv preprint arXiv:2308.09687, 2023.
  7. Language models are few-shot learners. In Neural Information Processing Systems, 2020.
  8. Sparks of artificial general intelligence: Early experiments with gpt-4. arXiv preprint arXiv:2303.12712, 2023.
  9. Diagnosing infeasible optimization problems using large language models. arXiv preprint arXiv:2308.12923, 2023.
  10. Learning to optimize: A primer and a benchmark. J. Mach. Learn. Res., 23(1), Jan 2022. ISSN 1532-4435.
  11. A deep reinforcement learning framework for column generation. In Neural Information Processing Systems (NeurIPS), 2022.
  12. Scaling instruction-finetuned language models. arXiv preprint arXiv:2210.11416, 2022.
  13. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168, 2021.
  14. Semi-supervised sequence learning. In Neural Information Processing Systems, 2015.
  15. Alibaba realizes millions in cost savings through integrated demand forecasting, inventory management, price optimization, and product recommendations. INFORMS Journal on Applied Analytics, 53(1):32–46, 2023.
  16. Fine-tuning pretrained language models: Weight initializations, data orders, and early stopping. arXiv preprint arXiv:2002.06305, 2020.
  17. Artificial intelligence for operations research: Revolutionizing the operations research process. arXiv preprint arXiv:2401.03244, 2024.
  18. Highlighting named entities in input for auto-formulation of optimization problems. In International Conference on Intelligent Computer Mathematics, pp.  130–141. Springer, 2023.
  19. Pal: Program-aided language models. In International Conference on Machine Learning, pp.  10764–10799. PMLR, 2023.
  20. GitHub Inc. GitHub copilot, 2021. URL https://github.com/features/copilot. Accessed: 21 January 2024.
  21. Google. Gemini AI, 2024. URL https://blog.google/technology/ai/google-gemini-ai/. Accessed: 25 January 2024.
  22. Lookback for learning to branch. Transactions of Machine Learning Research (TMLR), 2022. URL https://openreview.net/forum?id=EQpGkw5rvL.
  23. Gurobi Optimization, LLC. Gurobi Optimizer Reference Manual, 2024. URL https://www.gurobi.com.
  24. IBM. IBM ILOG CPLEX Optimizer, 2024. URL https://www.ibm.com/products/ilog-cplex-optimization-studio/cplex-optimizer.
  25. Jang, S. Tag embedding and well-defined intermediate representation improve auto-formulation of problem description. arXiv preprint arXiv:2212.03575, 2022.
  26. Mixtral of experts. arXiv preprint arXiv:2401.04088, 2024.
  27. Mip-gnn: A data-driven framework for guiding combinatorial solvers. In The AAAI Conference on Artificial Intelligence, 2022.
  28. Decomposed prompting: A modular approach for solving complex tasks. arXiv preprint arXiv:2210.02406, 2022.
  29. BART: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. In Jurafsky, D., Chai, J., Schluter, N., and Tetreault, J. (eds.), Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pp.  7871–7880, Online, July 2020. Association for Computational Linguistics. doi: 10.18653/v1/2020.acl-main.703. URL https://aclanthology.org/2020.acl-main.703.
  30. Large language models for supply chain optimization. arXiv preprint arXiv:2307.03875, 2023a.
  31. Synthesizing mixed-integer linear programming models from natural language descriptions. arXiv preprint arXiv:2311.15271, 2023b.
  32. Let’s verify step by step. arXiv preprint arXiv:2305.20050, 2023.
  33. Long, J. Large language model guided tree-of-thought. arXiv preprint arXiv:2305.08291, 2023.
  34. A survey in mathematical language processing. arXiv preprint arXiv:2205.15231, 2022.
  35. A novel approach for auto-formulation of optimization problems. arXiv preprint arXiv:2302.04643, 2023.
  36. OpenAI. Chatgpt: Language model for conversational agents, 2023a. URL https://www.openai.com/. Accessed: 1 January 2024.
  37. OpenAI. Introducing GPTs, 2023b. URL https://openai.com/blog/introducing-gpts. Accessed: 1 January 2024.
  38. GPT-4 technical report, 2023.
  39. Training language models to follow instructions with human feedback. In Advances in Neural Information Processing Systems, 2022.
  40. Un world food programme: Toward zero hunger with analytics. Informs Journal on Applied Analytics, 52(1):8–26, 2022.
  41. Augmenting operations research with auto-formulation of optimization models from problem descriptions. arXiv preprint arXiv:2209.15565, 2022.
  42. Nl4opt competition: Formulating optimization problems based on their natural language descriptions. In NeurIPS 2022 Competition Track, pp.  189–203. PMLR, 2023.
  43. Multitask prompted training enables zero-shot task generalization. arXiv preprint arXiv:2110.08207, 2021.
  44. Learning to summarize from human feedback. In Neural Information Processing Systems, 2020.
  45. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv: 2307.09288, 2023.
  46. Solving olympiad geometry without human demonstrations. Nature, 625(7995):476–482, 2024.
  47. Holy grail 2.0: From natural language to constraint models. arXiv preprint arXiv:2308.01589, 2023.
  48. Attention is all you need. Advances in neural information processing systems, 30, 2017.
  49. Stochastic Programming and Robust Optimization, pp.  395–447. Springer US, Boston, MA, 1997. ISBN 978-1-4615-6103-3. doi: 10.1007/978-1-4615-6103-3˙12. URL https://doi.org/10.1007/978-1-4615-6103-3_12.
  50. Finetuned language models are zero-shot learners. In International Conference on Learning Representations, 2022a. URL https://openreview.net/forum?id=gEZrGCozdqR.
  51. Chain-of-thought prompting elicits reasoning in large language models. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A. (eds.), Advances in Neural Information Processing Systems, volume 35, pp.  24824–24837. Curran Associates, Inc., 2022b.
  52. Williams, H. P. Model Building in Mathematical Programming, 5th Edition. Wiley, 2013. ISBN 978-1-118-44333-0. doi: 10.2307/253935.
  53. Crisp-dm: Towards a standard process model for data mining. In Proceedings of the 4th international conference on the practical applications of knowledge discovery and data mining, volume 1, pp.  29–39. Manchester, 2000.
  54. Integer and Combinatorial Optimization. Wiley Series in Discrete Mathematics and Optimization. Wiley, 1988. ISBN 9780471828198. URL https://books.google.com.tr/books?id=uG4PAQAAMAAJ.
  55. Autoformalization with large language models. In Koyejo, S., Mohamed, S., Agarwal, A., Belgrave, D., Cho, K., and Oh, A. (eds.), Neural Information Processing Systems, 2022.
  56. Large language models as optimizers. arXiv preprint arXiv:2309.03409, 2023.
  57. Tree of thoughts: Deliberate problem solving with large language models. In Neural Information Processing Systems, 2023.
  58. Automated and clinically optimal treatment planning for cancer radiotherapy. INFORMS Journal on Applied Analytics, 52(1):69–89, 2022. doi: 10.1287/inte.2021.1095. URL https://doi.org/10.1287/inte.2021.1095.
  59. The gap of semantic parsing: A survey on automatic math word problem solvers. IEEE transactions on pattern analysis and machine intelligence, 42(9):2287–2305, 2019.
  60. Fine-tuning language models from human preferences. CoRR, abs/1909.08593, 2019.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Segev Wasserkrug (7 papers)
  2. Leonard Boussioux (12 papers)
  3. Dick den Hertog (19 papers)
  4. Farzaneh Mirzazadeh (6 papers)
  5. Ilker Birbil (3 papers)
  6. Jannis Kurtz (18 papers)
  7. Donato Maragno (6 papers)
Citations (1)
View on Semantic Scholar