104. On Model Generation and Decomposition for Global Optimization and Machine Learning using the Generate-and-Solve Approach
Invited abstract in session TB-7: Global optimization I, stream Global optimization.
Thursday, 10:05 - 11:20Room: M:I
Authors (first author is the speaker)
| 1. | Ivo Nowak
|
| HAW Hamburg |
Abstract
We present decomposition-based generate-and-solve methods for generating reformulations of complex optimization problems that can be solved easier than the original problem. The methods are implemented in the open-source frameworks Decogo and Decolearn. Numerical results for nonconvex MINLPs and machine learning problems are presented.
Keywords
- Global optimization
- Optimization for learning and data analysis
- Mixed integer nonlinear optimization
Status: accepted
Back to the list of papers