LLM-Assisted Formulation of Robust and Adaptive Optimization Problems
A prompting and in-context-learning methodology that guides a large language model to translate a classical optimization problem into a robust or adaptive-robust formulation, incorporate user-specified uncertainty sets, derive the computationally tractable robust counterpart, and generate executable solver code. Demonstrates that specialized informational prompts can adapt an LLM to a niche modeling domain where its out-of-the-box performance is poor.
2501.00568
This paper investigates whether a large language model (ChatGPT) can make Robust Optimization (RO) and Adaptive Robust Optimization (ARO) accessible to non-experts. Robust Optimization seeks solution…