Legat, Benoît
[UCL]
Weisser, Tillmann
[LANL]
With the modeling power of Sum-of-Squares programming comes significant challenges for efficiently generating equivalent semidefinite programs that can be solved efficiently and accurately. We discuss in this talk the approach we used to implement a reformulation flexible, feature-rich and extensible while still being as efficient as specialized code. We then detail the different features implemented in SumOfSquares to improve the numerical conditioning and decrease the size of the semidefinite problem in order to improve scalability.
Bibliographic reference |
Legat, Benoît ; Weisser, Tillmann. SumOfSquares: A Julia package for Polynomial Optimization.INFORMS Annual Meeting (Virtual, du 07/11/2020 au 13/11/2020). |
Permanent URL |
http://hdl.handle.net/2078.1/240499 |