Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:B3L-C
Session:
Number:B3L-C-03
Analysis of Dynamical Systems Using Symbolic Regression
Soichiro Kanaya , Toma Takano , Satoshi Sunada , Tomoaki Niiyama,
pp.320-322
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.B3L-C-03
PDF download (1.2MB)
Summary:
We propose an identification method of dynamical systems' equations from the measurement data samples with the aid of symbolic regression and the AI-Feynman proposed by Tegmark et al. The symbolic regression based on a genetic algorithm is used for finding the system equations from measurement data, combined with neural networks and a physics-inspired model, whereas the AI- Feynman can be used for detecting simplified representations of the system equations, such as symmetry and separability. In this study, we apply our identification model for chaotic dynamical systems, such as Lorenz systems, and discuss the identification performance.