betaFeedback

WorksMathematics

Conditional limit theorems in dynamical systems

Abstract

This paper develops a comprehensive theory of conditional limit theorems for open dynamical systems. We study the statistical properties of the dynamics within transient regimes by establishing a suite of limit theorems describing the behavior of trajectories under the condition that they remain outside a ``hole'', or region of no return. Under a spectral gap assumption for the associated open transfer operator, we first establish conditioning and Yaglom limits and identify the corresponding conditionally invariant and conditionally ergodic measures. We then establish conditional exponential decay of correlations and the Strong Law of Large Numbers for both deterministic and quenched random systems. We also obtain a conditional Central Limit Theorem, Large Deviation Principle, and Berry–Esseen theorem, with additional assumptions in the quenched random case. We also derive standard (non-conditional) limit theorems on the surviving set of points that never enter the hole.

Introduction

Open dynamical systems, in which trajectories are removed once they enter a designated “hole” or region of no return, provide a natural framework for studying transient and metastable behaviour in deterministic and random dynamics. The time-asymptotic behaviour of such systems has been investigated extensively through escape rates, conditionally invariant measures, quasi-stationary behaviour, and Yaglom limits [47, 38, 13, 14, 32, 18, 4].

In this work we turn to complementary statistical questions that remain much less developed, namely conditional limit laws for observables when the underlying law is conditioned on survival up to a large but finite time. We are partly motivated by the fact that the loss of mass in open dynamical systems means that observed ensembles depend on the observation horizon. In long transient simulations only those trajectories that have survived up to the time of observation are available for analysis. This point of view is naturally phrased in terms of NN-conditioned measures, obtained by conditioning an initial distribution on survival up to time NN. These measures describe the statistics seen in long transient experiments, and they are the primary objects of this paper. The resulting questions include conditional ergodicity, laws of large numbers, central limit theorems, large deviations, and Berry–Esseen estimates.

To the best of our knowledge, rigorous results addressing these questions remain remarkably scarce. In deterministic open dynamics, an early result is the conditional ergodicity and central limit theorem of Collet–Martínez [12] for subshifts of finite type with respect to the distinguished conformal measure. More recent results on conditional limit laws for open dynamical systems and absorbing Markov chains include [31, 5, 46]. The present paper develops a systematic abstract transfer operator framework for conditioning limits and conditional statistical laws. Our setting accommodates both deterministic systems and quenched random dynamical systems in the standard cocycle framework [1, 15], and allows a broad class of initial distributions that are absolutely continuous with respect to a conformal reference measure. Our abstract assumptions are verified for several standard classes of dynamics, including random piecewise expanding interval maps with random holes [2, 3, 4], multidimensional expanding maps with random holes [30, 41], and subshifts of finite type with holes [39].

Under a spectral gap assumption (G) for the normalized open transfer-operator cocycle, we first prove convergence to conditioning limits (Theorem A), Yaglom limits (Theorem B(1)–(2)), uniqueness of the conditionally invariant measure (Theorem B(3)), and conditional exponential mixing (Theorem C). Building on these estimates, we establish conditional ergodic behaviour, a conditional weak law of large numbers, and a conditional strong law of large numbers (Theorem D). In particular, the conditional strong law in Theorem D(3) seems to be new already in its formulation. Under the stronger limit law condition (LL), and using a perturbative transfer-operator approach adapted from the literature for random closed systems [21, 22, 11, 4, 19], we further prove a conditional central limit theorem (Theorem E), a conditional large deviation principle (Theorem F), and a conditional Berry–Esseen theorem (Theorem G). The corresponding dynamical variance is identified via a Green–Kubo type formula (Proposition 8.1). Under suitable continuity and boundary assumptions, these estimates also yield standard dynamical properties and limit laws on the surviving set (Proposition 5.7 and Corollary 5.8). Furthermore, the method is not tied to these particular statements: with the appropriate perturbative or rare event estimates, the same conditional transfer operator framework should be useful for conditional versions of other statistical laws, such as functional central limit theorems, laws of the iterated logarithm, extreme value laws, Erdős–Rényi laws, and almost sure invariance principles.

It is important to distinguish these conditional limit laws from the ordinary limit laws for the restricted dynamics on X∞X_{\infty}, consisting of points whose forward orbits never enter the hole. Invariant measures on X∞X_{\infty} describe distributions of points that survive forever, whereas the conditioned laws we introduce describe the statistics of trajectories selected by survival up to a large finite time. This distinction is important in applications: the basin of the invariant measure on X∞X_{\infty} is frequently negligible with respect to natural initial distributions, such as Lebesgue measure, and therefore is only informative for a negligible set of initial conditions. In contrast, our finite-time conditioned measures and our conditional laws capture the statistics of non-negligible sets of initial conditions: those that are visible in long transient observations.

The paper is organized as follows. § 2 introduces (closed) random dynamical systems, and § 3 defines open random dynamical systems and the associated open random transfer operator cocycle. In § 4 we define the conditioning limit, Yaglom limit, and conditionally ergodic limit for random (and deterministic) probability measures. § 5 contains statements of all of our main results. § 6 verifies our assumptions for several example classes of deterministic and random dynamics. § 7 proves Theorems A–C, § 8 justifies the existence of a limiting conditional dynamical variance, and § 9 proves Theorem D. § 10 contains crucial derivative information on the leading Lyapunov exponent of the twisted transfer operator cocycle, and § 11 contains the proofs of Theorems E–G.

Random dynamical systems

Let (Ω,F,m)(\Omega, \mathcal{F}, m) be a probability space and σ:Ω→Ω\sigma: \Omega\to\Omega an invertible, ergodic map which preserves the measure mm. Let (Ξ,d)(\Xi, d) be a complete, compact metric space equipped with the Borel σ\sigma-algebra B\mathcal{B} and let {Xω}ω∈Ω\{X_\omega\}_{\omega\in\Omega} be a closed random subset of Ξ\Xi (i.e. XωX_\omega is a closed subset of Ξ\Xi for each ω∈Ω\omega\in\Omega and ω↦d(x,Xω)\omega\mapsto d(x, X_\omega) is measurable for every x∈Ξx \in\Xi), in particular,

Xω⊆Ξfor each ω∈Ω.X_\omega\subseteq\Xi\qquad\text{for each } \omega\in\Omega.

We let

X:=⋃ω∈Ω{ω}×Xω⊂Ω×Ξ.X := \bigcup_{\omega\in\Omega} \{\omega\} \times X_\omega\subset\Omega\times\Xi.

Further, we consider a family of surjective maps {Tω:Xω→Xσω}ω∈Ω\{T_\omega: X_\omega\to X_{\sigma\omega}\}_{\omega\in\Omega} and define the induced skew-product map T:X→XT : X \to X given by

T(ω,x)=(σω,Tω(x)).T(\omega, x) = (\sigma\omega, T_\omega(x)).

We make the following assumption:

(M1) The map T:X→XT : X \to X is measurable with respect to product σ\sigma-algebra F⊗B\mathcal{F} \otimes\mathcal{B}.

Random iteration of our maps TωT_\omega is denoted by Tωn:Xω→XσnωT_\omega^n : X_\omega\to X_{\sigma^n\omega} and is given by

Tωn:=Tσn−1ω∘Tσn−2ω∘⋯∘Tω(1)T_\omega^n := T_{\sigma^{n-1}\omega} \circ T_{\sigma^{n-2}\omega} \circ\cdots\circ T_\omega \tag*{(1)}

for each n≥1n \ge1. Notice that Tn(ω,x)=(σnω,Tωn(x))T^n(\omega, x) = (\sigma^n\omega, T_\omega^n(x)).

A probability measure ζ\zeta on (Ω×Ξ,F⊗B)(\Omega\times\Xi, \mathcal{F} \otimes\mathcal{B}) is said to be a random probability measure relative to mm if it has marginal mm (i.e. ζ∘π1−1=m\zeta\circ\pi_1^{-1} = m, where π1(ω,x)=ω\pi_1(\omega, x) = \omega). This property holds if and only if the disintegrations {ζω}ω∈Ω\{\zeta_\omega\}_{\omega\in\Omega} of ζ\zeta with respect to the partition {{ω}×Ξ}ω∈Ω\{\{\omega\} \times\Xi\}_{\omega\in\Omega} satisfy the following properties:

  1. For every B∈BB \in\mathcal{B}, the map Ω∋ω↦ζω(B)∈[0,1]\Omega\ni\omega\mapsto\zeta_\omega(B) \in[0,1] is measurable,

  2. For mm-a.e. ω∈Ω\omega\in\Omega, the map B∋B↦ζω(B)∈[0,1]\mathcal{B} \ni B \mapsto\zeta_\omega(B) \in[0,1] is a Borel probability measure.

Refer to [1, 15]. Thus, we often identify a random probability measure ζ\zeta with {ζω}ω∈Ω\{\zeta_\omega\}_{\omega\in\Omega} by writing ζ={ζω}ω∈Ω\zeta= \{\zeta_\omega\}_{\omega\in\Omega}. Given a measurable A⊆XA \subseteq X, we let M1(A)\mathcal{M}_1(A) denote the set of all random probability measures supported in AA. If A=XA = X, we simply write M1\mathcal{M}_1 for M1(X)\mathcal{M}_1(X). We also define a random measure ζ={ζω}ω∈Ω\zeta= \{\zeta_\omega\}_{\omega\in\Omega} by (1) and (2) with “measure” instead of “probability measure”.

Closed random transfer operator

We consider a Banach space (B,∥⋅∥)(\mathcal{B},\|\cdot\|) of complex-valued Borel measurable functions on Ξ\Xi and a random weight function g:X→(0,∞)g:\mathsf{X}\to(0,\infty) satisfying the following properties. We define the (closed) random transfer operator Lω,c:B→B\mathcal{L}_{\omega,c}:\mathcal{B}\to\mathcal{B} associated with the weight gω:=g(ω,⋅)g_\omega:=g(\omega,\cdot) as follows: for each ω∈Ω\omega\in\Omega, h∈Bh\in\mathcal{B}, and x∈Xσωx\in X_{\sigma\omega},

Lω,c(h)(x):=∑y∈Tω−1(x)gω(y)h(y).(2)\mathcal{L}_{\omega,c}(h)(x):=\sum_{y\in T_\omega^{-1}(x)}g_\omega(y)h(y). \tag*{(2)}

The same formula defines Lω,c(h)\mathcal{L}_{\omega,c}(h) for every bounded Borel measurable function h:Ξ→Ch:\Xi\to\mathbb{C}. We assume that the weight gg is chosen so that Lc\mathcal{L}_c is a well-defined, bounded linear operator. In fact we will assume that there is C1>0C_1>0 such that Lω,ch∈B\mathcal{L}_{\omega,c}h\in\mathcal{B} and

(C0) ∥Lω,ch∥≤C1∥h∥\|\mathcal{L}_{\omega,c}h\|\le C_1\|h\|

for almost every ω∈Ω\omega\in\Omega and h∈Bh\in\mathcal{B}, and that

(M2) the map (ω,x)↦(Lch)ω(x)(\omega,x)\mapsto(\mathcal{L}_c h)_\omega(x) is measurable for every B\mathcal{B}-valued random variable hh, where

(Lch)ω:=Lσ−1ω,chσ−1ωfor ω∈Ω.(\mathcal{L}_c h)_\omega:=\mathcal{L}_{\sigma^{-1}\omega,c}h_{\sigma^{-1}\omega}\qquad\text{for }\omega\in\Omega.

Remark 2.1. Note that if the random maps TωT_\omega are finite-to-one and the weight functions gωg_\omega are bounded, then the transfer operator Lω,c\mathcal{L}_{\omega,c} is always well-defined. On the other hand if the maps TωT_\omega are countable-to-one, then we require that the weight functions are summable. See [2] for details.

We make the following assumption on the closed system. Moreover, for a measurable set A⊂ΞA\subset\Xi, let A^\widehat{A} denote its indicator function. For a probability measure μ\mu on Ξ\Xi and a μ\mu-integrable measurable function h:Ξ→Ch:\Xi\to\mathbb{C}, we use the shorthand μ(h):=∫h dμ\mu(h):=\int h\,\mathrm{d}\mu.

(C1) There exists a random non-atomic Borel probability νc\nu_c such that

νσω,c(L~ω,ch)=νω,c(h)(3)\nu_{\sigma\omega,c}(\widetilde{\mathcal{L}}_{\omega,c}h)=\nu_{\omega,c}(h) \tag*{(3)}

for mm-a.e. ω∈Ω\omega\in\Omega and every bounded Borel function h:Ξ→Ch:\Xi\to\mathbb{C}, where

L~ω,ch:=λω,c−1Lω,ch,andλω,c:=νσω,c(Lω,cX^ω).\widetilde{\mathcal{L}}_{\omega,c}h:=\lambda_{\omega,c}^{-1}\mathcal{L}_{\omega,c}h,\qquad\text{and}\qquad \lambda_{\omega,c}:=\nu_{\sigma\omega,c}(\mathcal{L}_{\omega,c}\widehat{X}_\omega).

Notice that

Lω,c(f∘Tω⋅h)=f⋅Lω,c(h)\mathcal{L}_{\omega,c}(f\circ T_\omega\cdot h)=f\cdot\mathcal{L}_{\omega,c}(h)

for every pair of functions f,h∈Bf,h\in\mathcal{B} by the definition of Lc\mathcal{L}_c in (2). Therefore, we get the following important consequence of (3) regarding duality:

νω,c(f∘Tω⋅h)=νσω,c(f⋅L~ω,ch).(4)\nu_{\omega,c}(f\circ T_\omega\cdot h)=\nu_{\sigma\omega,c}(f\cdot\widetilde{\mathcal{L}}_{\omega,c}h). \tag*{(4)}

In fact, below we will only use (4) instead of (3).

Remark 2.2. The random probability measure νc\nu_c is known as the random (closed) conformal measure. Such measures have been shown to exist in many random settings, e.g. [7, 9, 34, 33, 35, 2, 3].

Open random dynamical systems

Let HH be a measurable set in XX. The fibers HωH_\omega are determined by the condition that {ω}×Hω=H∩({ω}×Xω)\{\omega\} \times H_\omega= H \cap(\{\omega\} \times X_\omega). For each ω∈Ω\omega\in\Omega and n≥1n \ge1 we define

Xω,n:=⋂j=0nTω−j(Xσjω,0)(5)X_{\omega,n} := \bigcap_{j=0}^{n} T_\omega^{-j}(X_{\sigma^j\omega,0}) \tag*{(5)}

References

  1. [1]L. Arnold. Random Dynamical Systems. Springer Monographs in Mathematics. Springer Berlin Heidelberg, Berlin, Heidelberg, 1998. 3, 4
  2. [2]J. Atnip, G. Froyland, C. González-Tokman, and S. Vaienti. Thermodynamic formalism for random weighted covering systems. Communications in Mathematical Physics, July 2021. 3, 5, 21, 22, 23DOI
  3. [3]J. Atnip, G. Froyland, C. González-Tokman, and S. Vaienti. Equilibrium states for non-transitive random open and closed dynamical systems. Ergodic Theory and Dynamical Systems, pages 1–23, Oct. 2022. 3, 5, 12, 21, 22, 23DOI
  4. [4]J. Atnip, G. Froyland, C. González-Tokman, and S. Vaienti. Thermodynamic formalism and perturbation formulae for quenched random open dynamical systems. arXiv:2307.00774, to appear in Dissertationes Mathematicae, 2023. 2, 3, 7, 10, 12, 14, 15, 18, 20, 22, 23, 25, 26, 28, 54arxiv.org/abs/2307.00774
  5. [5]B. Bassols Cornudella and M. Castro. Exact stable laws and rare-event limits for absorbing markov chains, 2026. arXiv:2609.22499. 3
  6. [6]P. Billingsley. Convergence of Probability Measures. Wiley, 2 edition, 1999. 19
  7. [7]T. Bogenschütz and V. M. Gundlach. Ruelle’s transfer operator for random subshifts of finite type. Ergod. Th. & Dynam. Sys, 15:413–447, 1995. 5, 21DOI
  8. [8]L. A. Breyer and G. O. Roberts. A quasi-ergodic theorem for evanescent processes. Stochastic Processes and their Applications, 84(2):177–186, Dec. 1999. 17DOI
  9. [9]R. Brück and M. Büger. Generalized iteration. Computational Methods and Function Theory, 3(1-2, [On table of contents: 2004]):201–252, 2003. 5, 21
  10. [10]J. Buzzi. Absolutely continuous S.R.B. measures for random Lasota–Yorke maps. Transactions of the American Mathematical Society, 352(7):3289–3303, 2000. 14DOI
  11. [11]Y. M. Chung, Y. Nakano, and J. Wittsten. Quenched limit theorems for random u (1) extensions of expanding maps. Discrete and Continuous Dynamical Systems-Series A, 43(1):338–377, 2023. 3, 15, 54, 55arxiv.org/abs/2104.01606
  12. [12]P. Collet and S. Martínez. Diffusion coefficient in transient chaos. Nonlinearity, 12(3):445–450, 1999. 3DOI
  13. [13]P. Collet, S. Martínez, and J. San Martín. Quasi-stationary distributions. Probability and its Applications (New York). Springer, Heidelberg, 2013. Markov chains, diffusions and dynamical systems. 2, 9, 10, 12
  14. [14]P. Collet, S. Martínez, and V. Maume-Deschamps. On the existence of conditionally invariant probability measures in dynamical systems. Nonlinearity, 13(4):1263–1274, June 2000. 2, 12
  15. [15]H. Crauel. Random probability measures on Polish spaces, volume 11 of Stochastics Monographs. Taylor & Francis, London, 2002. 3, 4
  16. [16]H. Crimmins. Stability of hyperbolic Oseledets splittings for quasi-compact operator cocycles. arXiv:1912.03008 [math], Dec. 2019. arXiv: 1912.03008. 25, 28, 31arxiv.org/abs/1912.03008
  17. [17]M. F. Demers, C. J. Ianzano, P. Mayer, P. Morfe, and E. C. Yoo. Limiting distributions for countable state topological markov chains with holes. Discrete and Continuous Dynamical Systems, 37(1):105–130, 2017.DOI
  18. [18]M. F. Demers and L.-S. Young. Escape rates and conditionally invariant measures. Nonlinearity, 19(2):377–397, 2006.
  19. [19]D. Dragičević and Y. Hafouta. Limit theorems for random expanding or anosov dynamical systems and vector-valued observables. Annales Henri Poincaré, 21(12):3869–3917, 2020.DOI
  20. [20]D. Dragičević, G. Froyland, C. González-Tokman, and S. Vaienti. Almost sure invariance principle for random piecewise expanding maps. Nonlinearity, 31(5):2252–2280, May 2018.DOI
  21. [21]D. Dragičević, G. Froyland, C. González-Tokman, and S. Vaienti. A spectral approach for quenched limit theorems for random expanding dynamical systems. Communications in Mathematical Physics, Jan. 2018. arXiv: 1705.02130.DOI
  22. [22]D. Dragičević, G. Froyland, C. González-Tokman, and S. Vaienti. A spectral approach for quenched limit theorems for random hyperbolic dynamical systems. Transactions of the American Mathematical Society, 373(1):629–664, 2020.arxiv.org/abs/1812.07340
  23. [23]M. Engel, J. Lamb, and M. Rasmussen. Conditioned lyapunov exponents for random dynamical systems. Transactions of the American Mathematical Society, 372(9):6343–6370, 2019.DOI
  24. [24]G. Froyland, C. González-Tokman, and R. Murray. Quenched stochastic stability for eventually expanding-on-average random interval map cocycles. Ergodic Theory and Dynamical Systems, 39(10):2769–2792, 2019.DOI
  25. [25]G. Froyland, S. Lloyd, and A. Quas. A semi-invertible Oseledets theorem with applications to transfer operator cocycles. Discrete and Continuous Dynamical Systems, 33(9):3835–3860, Mar. 2013.arxiv.org/abs/1001.5313
  26. [26]C. González-Tokman and A. Quas. A semi-invertible operator Oseledets theorem. Ergodic Theory and Dynamical Systems, 34(4):1230–1272, 2014.
  27. [27]S. Gouëzel. Limit theorems in dynamical systems using the spectral method. Hyperbolic dynamics, fluctuations and large deviations, 89:161–193, 2015.
  28. [28]L. H. H. Hennion. Limit theorems for Markov chains and Stochastic Properties of Dynamical Systems by Quasicontractness, volume 1766. Springer-Verlag, 2001.
  29. [29]G. He, H. Zhang, and Y. Zhu. On the quasi-ergodic distribution of absorbing Markov processes. Statistics & Probability Letters, 149:116–123, June 2019.DOI
  30. [30]G. Keller. Generalized bounded variation and applications to piecewise monotonic transformations. Z. Wahr. verw. Geb., 69:461–478, 1985.DOI
  31. [31]J. Leppänen, Y. Nakajima, and Y. Nakano. Functional correlation bound for random Lasota–Yorke maps with holes and its applications to conditional normal approximations, 2026. arXiv:2604.17203.arxiv.org/abs/2604.17203
  32. [32]C. Liverani and V. Maume-Deschamps. Lasota–Yorke maps with holes: conditionally invariant probability measures and invariant probability measures on the survivor set. Annales de l’Institut Henri Poincaré (B) Probability and Statistics, 39(3):385–412, May 2003.
  33. [33]V. Mayer and M. Urbański. Countable alphabet random subshifts of finite type with weakly positive transfer operator. Journal of Statistical Physics, 160(5):1405–1431, 2015.DOI
  34. [34]V. Mayer, M. Urbański, and B. Skorulski. Distance Expanding Random Mappings, Thermodynamical Formalism, Gibbs Measures and Fractal Geometry, volume 2036 of Lecture Notes in Mathematics. Springer Berlin Heidelberg, Berlin, Heidelberg, 2011.DOI
  35. [35]V. Mayer and M. Urbański. Random dynamics of transcendental functions. Journal d’Analyse Mathématique, 134(1):201–235, Feb. 2018.arxiv.org/abs/1409.7179
  36. [36]S. Méléard and D. Villemonais. Quasi-stationary distributions and population processes. Probab. Surv., 9:340–410, 2012.
  37. [37]C. Nan. Invariant and conditionally invariant measures for random open interval maps with countably many branches, 2026. arXiv:2603.19586.arxiv.org/abs/2603.19586
  38. [38]G. Pianigiani and J. A. Yorke. Expanding maps on sets which are almost invariant. Decay and chaos. Trans. Amer. Math. Soc., 252:351–366, 1979.DOI
  39. [39]M. Pollicott and M. Urbanski. Open conformal systems and perturbations of transfer operators. Springer Berlin Heidelberg, New York, NY, 1st edition edition, 2017. 3, 12, 17, 33, 34, 35DOI
  40. [40]M. Rychlik. Bounded variation and invariant measures. Studia Mathematica, 76:69–80, 1983. 21DOI
  41. [41]B. Saussol. Absolutely continuous invariant measures for multidimensional expanding maps. Israel Journal of Mathematics, 116(1):223–248, 2000. 3, 26, 27, 28, 29, 30, 32DOI
  42. [42]H. Tanaka. Perturbation analysis in thermodynamics using matrix representations of ruelle operators and its application to graph ifs. Nonlinearity, 32:728–767, 2019. 34DOI
  43. [43]H. Tanaka. Perturbed finite-state markov systems with holes and perron complements of ruelle operators. Israel Journal of Mathematics, 236:91–131, 2020. 34DOI
  44. [44]H. Tanaka. Quasi-compactness of transfer operators for topological markov shifts with holes. Discrete and Continuous Dynamical Systems, 44(9):2464–2490, 2024. 34DOI
  45. [45]E. A. van Doorn and P. K. Pollett. Quasi-stationary distributions for discrete-state models. European J. Oper. Res., 230(1):1–14, 2013. 9DOI
  46. [46]C. Wang and Y. Su. Conditional limit laws for open intermittent dynamics, 2026. arXiv:2610.21175. 3arxiv.org/abs/2610.21175
  47. [47]A. M. Yaglom. Certain limit theorems of the theory of branching random processes. Doklady Akad. Nauk SSSR (N.S.), 56:795–798, 1947. 2
  48. [48]J. Zhang, S. Li, and R. Song. Quasi-stationarity and quasi-ergodicity of general Markov processes. Science China Mathematics, 57(10):2013–2024, Oct. 2014. 17DOI

Paper details

vlnc:W_htnv24f3
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

ChatGPT

View statement
Looking for related papers…
Contents