Novel Models and Methods for Optimization Problems with Tree Ensembles Embedded
Abstract
The Air Force faces numerous optimization problems that are complicated by the complex, ever-changing, and uncertain nature of its operations. Additionally, decision-makers must frequently factor in the actions of adversarial entities, which further complicates the evaluation of consequences. In such situations, it might be difficult to precisely specify the components of these models. Machine learning provides effective tools to infer these complex components from data. This proposal seeks to lay the mathematical foundations for the solution of families of optimization models whose components are trained machine learning models. Enabling the integration of machine learning components into optimization models can expand the range of applications of optimization. In turn, developing techniques to solve the resulting models can enhance the efficiency and effectiveness of many processes that are central to the operation of the Air Force.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Mar 06, 2024
- Source ID
- FA95502310451
Entities
People
- Jean-philippe Richard
Organizations
- Air Force Office of Scientific Research
- Regents of the University of Minnesota
- United States Air Force