Numerical third-order Hopf form for ordinary differential equations
Abstract A data-driven approach has been developed to obtain the numerical third-order Hopf form for two-dimensional ODEs. By using a single numerical solution, the flow on the 2D phase-plane is recovered by a least-squares fit. We then calculate the Poincaré-Lyapunov constant. Sensitivity analysis is carried out to determine the accuracy of the calculated Poincaré-Lyapunov constant; a main result is a formula to compute the relative error of the Poincaré-Lyapunov constant from the errors of the identified coefficients in the third-order Hopf form. We also provide a sufficient condition for correctly identifying the criticality of the system (the sign of the Poincaré-Lyapunov constant). A performance metric is defined for our identification algorithm and this metric is computed by Monte Carlo simulations. Our method achieves good accuracy.
Authors
- Tamás Kalmár‐Nagy (ORCID: https://orcid.org/0000-0003-1374-2620)
- Dávid András Horváth (ORCID: https://orcid.org/0000-0002-3534-0457)
Institutions
- Budapest University of Technology and Economics (HU)
Publication Details
- Journal
- Nonlinear Dynamics
- Published
- 2025-06-12
- DOI
- https://doi.org/10.1007/s11071-025-11351-5
- Primary Topic
- Model Reduction and Neural Networks
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Budapesti Műszaki és Gazdaságtudományi Egyetem
- Nemzeti Kutatási Fejlesztési és Innovációs Hivatal
- Nemzeti Kutatási, Fejlesztési és Innovaciós Alap