Data-driven robust optimization approach: theory and human decision-makers
In practice, there is a lot of data available, but uncertainty is an indispensable issue for decision-makers. This project focuses on uncertainties in supply chains with highly volatile demand. We aim to develop a computationally efficient and human decision-compatible data-driven optimization method in inventory control. Robust optimization, in many cases, is a computationally efficient method that solves an optimization problem under multi-dimensional uncertainty.
Besides, our approach is human decision-compatible, which means the method can assist human decisions. First, since there is a bias toward the solutions obtained by robust optimization, we will calibrate the degree of the conservativeness of the method by observing human decisions. Second, we will use a lab experiment to evaluate the performance of our method as a decision support tool for human decisions.
June 2023 – Annual AI Planner of the Future Event. Presentation.
More info: firstname.lastname@example.org