2000 character limit reached
Unifying Lyapunov exponents with probabilistic uncertainty quantification (2403.05085v2)
Published 8 Mar 2024 in math.DS, nlin.CD, and physics.data-an
Abstract: The Lyapunov exponent is well-known in deterministic dynamical systems as a measure for quantifying chaos and detecting coherent regions in physically evolving systems. In this Letter, we show how the Lyapunov exponent can be unified with stochastic sensitivity (which quantifies the uncertainty of an evolving uncertain system whose initial condition is certain) within a finite time uncertainty quantification framework in which both the dynamics and the initial condition of a continuously evolving $ n $-dimensional state variable are uncertain.
- J. Berner et al., Bull. Amer. Meteorol. Soc. 98, 565 (2017).
- L. Blake, J. Maclean, and S. Balasuriya (2023), https://arxiv.org/abs/2309.16334.
- D. Sanz-Alonso and A. M. Stuart, Comm. Math. Sci. 15, 2087 (2017).
- Y. N. Blagoveshchenskii, Theor. Prob. Appl. 7, 17 (1962).
- S. Balasuriya, SIAM Review 62, 781 (2020).
- S. C. Shadden, F. Lekien, and J. E. Marsden, Physica D 212, 271 (2005).
- S. Cheng et al., IEEE-CAA J. Auto. Sinica 10, 1361 (2023).
- S. MacDonald and D. Campbell, Statistics Surveys 15, 1 (2021).
- T. Berry and J. Harlim, SIAM-ASA J. Uncert. Quant. 3, 484 (2015).
- A. Badza, T. W. Mattner, and S. Balasuriya, Physica D 444, 133580 (2023).
- B. Kaszás and G. Haller, Chaos 30, 113144 (2020).
- S. Balasuriya, R. Kalampattel, and N. Ouellette, J. Fluid Mech. 807, 509 (2016).
- P. Bergamasco, G. Carlo, and A. Rivas, Phys. Rev. E 108, 024208 (2023).
- T. Xu, T. Scaffidi, and X. Cao, Phys. Rev. Lett. 124, 140602 (2020).
- J. Maldacena and D. Sandford, Phys. Rev. D 94, 106002 (2016).
- J. Maldacena, S. Shenker, and D. Stanford, J. High Energy Phys. 2016, 106 (2016).
- R. Jalabert and S. Pastawski, Phys. Rev. Lett. 86, 2490 (2001).
- C. Tsallis, Phil. Trans. R. Soc. A 381, 20220293 (2023).
- D. Zou, AIP Advances 13, 085330 (2023).
- E. Rozenbaum, S. Ganeshan, and V. Galitski, Phys. Rev. Lett. 118, 086801 (2017).
- S. Balasuriya and G. A. Gottwald, Phys. Rev. E 98, 013106 (2018).
- K. Itô, Nagoya Math. J. 1, 35 (1950).
- K. Law, A. Stuart, and K. Zygalakis, Data Assimilation: A Mathematical Introduction, Texts in Applied Mathematics No. volume 62 (Springer, Cham Heidelberg New York Dordrecht London, 2015).
- The result arises by using the whitening transformation 𝝂=𝚵0−1(𝝃−𝝃0)/2δ2𝝂superscriptsubscript𝚵01𝝃subscript𝝃02superscript𝛿2{\bm{\nu}}={\bm{\Xi}}_{0}^{-1}\left({\bm{\xi}}-{\bm{\xi}}_{0}\right)/\sqrt{2% \delta^{2}}bold_italic_ν = bold_Ξ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT start_POSTSUPERSCRIPT - 1 end_POSTSUPERSCRIPT ( bold_italic_ξ - bold_italic_ξ start_POSTSUBSCRIPT 0 end_POSTSUBSCRIPT ) / square-root start_ARG 2 italic_δ start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT end_ARG on the Gaussian variable 𝝃𝝃{\bm{\xi}}bold_italic_ξ.
- S. Balasuriya, N. T. Ouellette, and I. I. Rypina, Physica D 372, 31 (2018).
- The trace of a positive definite matrix is the sum of its eigenvalues (all of which are positive), and the norm of such a matrix is its largest eigenvalue.