Abstract
We establish an abstract quenched linear response result for random dynamical systems, which we then apply to the case of smooth expanding on average cocycles on the unit circle. In sharp contrast to the existing results in the literature, we deal with the class of random dynamics that does not necessarily exhibit uniform decay of correlations. Our techniques rely on the infinite-dimensional ergodic theory and in particular, on the study of the top Oseledets space of a parametrized transfer operator cocycle. Finally, we exhibit a surprising phenomenon: a random system and a smooth observable for which quenched linear response holds, but annealed response fails.
Similar content being viewed by others
Notes
Here and throughout the paper this means that for each \(\varepsilon \in I\), there exists a full measure set \(\Omega _\varepsilon \subset \Omega \).
We drop the \(\ell \) dependency in both \(\alpha , D\) to alleviate the notation.
See [16, Appendix A, Examples 1 and 2]. We note that the present example gives a smooth counterpart to the non-summability of integrated correlations phenomenon described in the aforementioned paper.
References
Arnold, L.: Random Dynamical Systems. Springer Monographs in Mathematics. Springer, Berlin (1998)
Backes, L., Dragičević, D.: Periodic approximation of exceptional Lyapunov exponents for semi-invertible operator cocycles. Ann. Acad. Sci. Fenn. Math. 44, 183–209 (2019)
Bahsoun, W., Saussol, B.: Linear response in the intermittent family: differentiation in a weighted \(C^0\)-norm. Discrete Contin. Dyn. Syst. 36, 6657–6668 (2016)
Bahsoun, W., Ruziboev, M., Saussol, B.: Linear response for random dynamical systems. Adv. Math. 364, 107011 (2020)
Baladi, V.: On the susceptibility function of piecewise expanding interval maps. Commun. Math. Phys. 275, 839–859 (2007)
Baladi, V.: Linear response, or else. ICM Seoul. In: Proceedings, vol. III, pp. 525–545 (2014)
Baladi, V.: Dynamical zeta functions and dynamical determinants for hyperbolic maps. A functional approach. Ergebnisse der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics, vol. 68. Springer, Cham, xv+291 (2018)
Baladi, V., Kondah, A., Schmitt, B.: Random correlations for small perturbations of expanding maps. Random Comput. Dyn. 4, 179–204 (1996)
Baladi, V., Smania, D.: Linear response formula for piecewise expanding unimodal maps. Nonlinearity 21, 677–711 (2008). (Corrigendum: Nonlinearity 25, 2203–2205 (2012))
Baladi, V., Smania, D.: Linear response for smooth deformations of generic nonuniformly hyperbolic unimodal maps. Ann. Sci. Éc. Norm. Supér. 45, 861–926 (2012)
Baladi, V., Todd, M.: Linear response for intermittent maps. Commun. Math. Phys. 347, 857–874 (2016)
Bogenschütz, T.: Stochastic stability of invariant subspaces. Ergod. Theory Dyn. Syst. 20, 663–680 (2000)
Bomfim, T., Castro, A., Varandas, P.: Differentiability of thermodynamical quantities in non-uniformly expanding dynamics. Adv. Math. 292, 478–528 (2016)
Butterley, O., Liverani, C.: Smooth Anosov flows: correlation spectra and stability. J. Mod. Dyn. 1, 301–322 (2007)
Butterley, O., Liverani, C.: Robustly invariant sets in fiber contracting bundle flows. J. Mod. Dyn. 7, 255–267 (2013)
Buzzi, J.: Exponential decay of correlations for random Lasota–Yorke maps. Commun. Math. Phys. 208, 25–54 (1999)
Crimmins, H.: Stability of hyperbolic Oseledets splittings for quasi-compact operator cocycles. Discrete Contin. Dyn. Syst. 42, 2795–2857 (2022)
Crimmins, H., Nakano, Y.: A spectral approach to quenched linear and higher-order response for partially hyperbolic dynamics. arXiv e-prints (2021). arXiv:2105.11188
Dolgopyat, D.: On differentiability of SRB states for partially hyperbolic systems. Invent. Math. 155, 389–449 (2004)
Dragičević, D., Froyland, G., Gonzàlez-Tokman, C., Vaienti, S.: A spectral approach for quenched limit theorems for random expanding dynamical systems. Commun. Math. Phys. 360, 1121–1187 (2018)
Dragičević, D., Sedro, J.: Statistical stability and linear response for random hyperbolic dynamics. Ergod. Theory Dyn. Syst. (2020). https://doi.org/10.1017/etds.2021.153
Dragičević, D., Sedro, J.: Quenched limit theorems for expanding on average cocycles. arXiv:2105.00548
Dragičević, D., Hafouta, Y., Sedro, J.: A vector-valued almost sure invariance principle for random expanding on average cocycles. arXiv:2108.08714
Froyland, G., Gonzalez-Tokman, C., Quas, A.: Stability and approximation of random invariant densities for Lasota–Yorke map cocycles. Nonlinearity 27, 647–660 (2014)
Galatolo, S., Giulietti, P.: A linear response for dynamical systems with additive noise. Nonlinearity 32, 2269–2301 (2019)
Galatolo, S., Sedro, J.: Quadratic response of random and deterministic dynamical systems. Chaos 30, 023113 (2020)
Gonzàlez-Tokman, C., Quas, A.: A semi-invertible operator Oseledets theorem. Ergod. Theory Dyn. Syst. 34, 1230–1272 (2014)
Gouëzel, S., Liverani, C.: Banach spaces adapted to Anosov systems. Ergod. Theory Dyn. Syst. 26, 123–151 (2006)
Hafouta, Y., Kifer, Yu.: Nonconventional Limit Theorems and Random Dynamics. World Scientific, Singapore (2018)
Hairer, M., Majda, A.: A simple framework to justify linear response theory. Nonlinearity 23, 909–922 (2010)
Kifer, Y.: Thermodynamic formalism for random transformations revisited. Stoch. Dyn. 8(1), 77–102 (2008)
Korepanov, A.: Linear response for intermittent maps with summable and nonsummable decay of correlations. Nonlinearity 29, 1735–1754 (2016)
Ruelle, D.: Differentiation of SRB states. Commun. Math. Phys. 187, 227–241 (1997)
Sedro, J., Rugh, H.H.: Regularity of characteristic exponents and linear response for transfer operator cocycles. Commun. Math. Phys. 383, 1243–1289 (2021)
Sedro, J.: A regularity result for fixed points, with applications to linear response. Nonlinearity 31, 1417–1440 (2018)
Acknowledgements
D.D. was supported in part by Croatian Science Foundation under the project IP-2019-04-1239 and by the University of Rijeka under the projects uniri-prirod-18-9 and uniri-pr-prirod-19-16. J.S. was supported by the European Research Council (E.R.C.) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 787304). P.G was partially supported by the PRIN Grant 2017S35EHN ”Regular and stochastic behaviour in dynamical systems” and by the INDAM - GNFM 2020 grant “Deterministic and stochastic dynamical systems for climate studies”. The authors have no conflict of interest to declare. Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by M. Hairer.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
An Example Where Quenched Response Holds and Annealed Response Fails
An Example Where Quenched Response Holds and Annealed Response Fails
Let us remind the reader that whenever quenched linear response holds (see e.g. Theorem 20), one has, for any smooth observable \(\phi \), that
In particular, the integral on the L.H.S is well-defined.
If one is interested in annealed linear response, the object of interest becomes the (double) integral \(\int _{\Omega }\int _{{\mathbb {S}}^1}\phi {\hat{h}}_\omega dm~d{\mathbb {P}}\): when it holds this integral is well-defined, and one has
Given that quenched results concerns a.e trajectory, and that annealed one are on average, in general one expects that a quenched result implies its annealed counterpart. This intuition is validated in the context of response theory by the existing results [18, 22, 34], where stochasticity is uniform (i.e. assumptions are uniform w.r.t \(\omega \)). However, when there is no stochastic uniformity, this intuition can fail, as the next example shows:
Consider \(({\tilde{\Omega }},{\mathcal {B}},\mathbb {Q}, S)\) the full-shift over \(\{1,2,\dots \}\), with probability vector
\((Z,Z/2^{2+\delta },\dots ,Z/n^{2+\delta },\dots )\), for some \(0\le \delta \le 1\), Z being the normalization constant.
Let \(h:{\tilde{\Omega }}\rightarrow {\mathbb {R}}\) be the (positive) observable defined by \(h(\omega )=\omega _0\) if \(\omega \,{:=}\,(\omega _n)_{n\in {\mathbb {Z}}} \in {\tilde{\Omega }}\). Note that
when \(0<\delta \le 1\).
Define \((\Omega ,{\mathcal {F}},{\mathbb {P}},\sigma )\) to be the suspension over S with roof function h, i.e. \(\Omega \,{:=}\,\{(\omega ,i)\in {\tilde{\Omega }}\times {\mathbb {N}},~ 0\le i<h(\omega )\}\), \(\sigma :\Omega \circlearrowleft \) is given by
and \({\mathbb {P}}(A)\,{:=}\,\left( \int _{{\tilde{\Omega }}} h~ d\mathbb {Q}\right) ^{-1}\sum _{i\ge 0}\mathbb {Q}\left( A\cap ({\tilde{\Omega }}\times \{i\})\right) \).
We can now define our random system: take \(T_0:{\mathbb {S}}^1\circlearrowleft \) to be the doubling map, i.e. \(T_0(x)=2x \ (\text {mod} \ 1)\), and let \(T_1\) be the identity map on \({\mathbb {S}}^1\). We consider the random circle map \(T_{(\omega ,i)}\), \((\omega ,i)\in \Omega \) defined by
Clearly, \((T_{(\omega ,i)})_{(\omega , i)\in \Omega }\) is an expanding on average cocycle, i.e.
Indeed, observe that \((T_{(\omega ,i)})_{(\omega , i)\in \Omega }\) is a particular case of cocycles \((T_\omega )_{\omega }\) introduced in Example 17. Moreover, \(\mu _{(\omega , i)}=m\) for \((\omega , i) \in \Omega \), where m denotes the Lebesgue measure on \({\mathbb {S}}^1\).
Let \(n_c\) be the (random) covering time for the interval [0, 1/2]:
Then, it is easy to see that \(n_c(\omega , i)=h(\omega )-i\). Observe that \(n_c\) is not integrable. Indeed, for each \(N\in {\mathbb {N}}\) we have that
for some constant \(C>0\), which easily implies that \(n_c\) is not integrable.
We now introduce the observable: consider a \(\psi \in C^\infty ({\mathbb {S}}^1)\), such that:
-
(i)
\(\text {supp}~\psi \subset I\), where \(I\subset {\mathbb {S}}^1\) is an interval such that \(I\cap \dfrac{1}{2}I=\emptyset \) (in particular, we can take any small enough \(I\not \ni 0\)).
-
(ii)
\(\int _{{\mathbb {S}}^1}\psi ~dm=0\) and \(\int _{{\mathbb {S}}^1}\psi ^2~dm=1\).
It is easy to see that \(\int _{{\mathbb {S}}^1}\psi \cdot \psi \circ T_0~dm=0\). Therefore, we have
In particular, it follows from the non-integrability of \(n_c\)Footnote 3 that
Let us now introduce the perturbed cocycle \(T_{(\omega ,i),\varepsilon }:= D_{\varepsilon }\circ T_{(\omega ,i)}\), where \(D_{\varepsilon }:{\mathbb {S}}^1\circlearrowleft \) is given by
Denote by \(S(x)\,{:=}\,-\int _0^x\psi ~dm\). Since this perturbation is smooth and deterministic (\(D_\varepsilon \) does not depend on \(\omega \)), we may apply Theorem 20, and obtain that for a.e. \((\omega ,i)\in \Omega \), the integral \(\int _{{\mathbb {S}}^1}\psi {\hat{h}}_{(\omega , i)} dm\) is well defined, and satisfies
where we used that \(\hat{{\mathcal {L}}}_{(\omega ,i)}f = -({\mathcal {L}}_{(\omega ,i)}(f)\cdot S)'\) (which can be proved by arguing as in [26, Sec 6.2]) and \(h_{(\omega ,i)}=\mathbbm {1}\) for our particular example.
Hence, from (61)
In particular, annealed response cannot hold.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Dragičević, D., Giulietti, P. & Sedro, J. Quenched Linear Response for Smooth Expanding on Average Cocycles. Commun. Math. Phys. 399, 423–452 (2023). https://doi.org/10.1007/s00220-022-04560-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00220-022-04560-1