Augmenting Operations Research with Auto-Formulation of Optimization Models from Problem Descriptions
This work addresses the challenge of reducing manual effort in model formulation for operations researchers, though it appears incremental as it builds on existing controlled generation techniques.
The paper tackles the problem of simplifying optimization modeling in operations research by developing an augmented intelligence system that automatically suggests formulations from problem descriptions, and it evaluates this approach using a new dataset of linear programming problems across various domains.
We describe an augmented intelligence system for simplifying and enhancing the modeling experience for operations research. Using this system, the user receives a suggested formulation of an optimization problem based on its description. To facilitate this process, we build an intuitive user interface system that enables the users to validate and edit the suggestions. We investigate controlled generation techniques to obtain an automatic suggestion of formulation. Then, we evaluate their effectiveness with a newly created dataset of linear programming problems drawn from various application domains.