Abstract
We discuss the implementation of a novel demanding algorithm to compute contraction metrics for nonlinear systems. We compare MATLAB- and C++-implementations and study the effect of parallelizing the code. Further, we explore the use of different low-level math-libraries for the C++-implementation and the use of an undocumented code-path that considerably speeds up MATLAB code 011 AMD's Ryzen processors.
| Original language | English |
|---|---|
| Pages (from-to) | 310-316 |
| Number of pages | 7 |
| Journal | IFAC-PapersOnLine |
| Volume | 54 |
| Issue number | 9 |
| DOIs | |
| Publication status | Published - 1 Jun 2021 |
| Event | 24th International Symposium on Mathematical Theory of Networks and Systems, MTNS 2020 - Cambridge, United Kingdom Duration: 23 Aug 2021 → 27 Aug 2021 |
Bibliographical note
Publisher Copyright: Copyright © 2021 The Authors. This is an open access article under the CC BY-NC-ND license.Other keywords
- Contraction metric
- Lyapunov function
- Mesh-free collocation
- Numerical algorithm
- Riemannian metric