Abstract
Let f : X → X be an invertible Lipschitz transformation on a compact metric space X. Given a Hölder continuous invertible operator cocycles on a Banach space and an f-invariant ergodic measure, this paper establishes the Hölder continuity of Oseledets subspaces over a compact set of arbitrarily large measure. This extends a result in [V Araujo, A I Bufetov and S Filip, On Hölder-continuity of Oseledets subspaces, J. Lond. Math. Soc. 2016, 93 : 194–218]. for invertible operator cocycles on a Banach space. This paper also proves the Hölder continuity in the non-invertible case. Finally, some applications are given in the end of this paper.
Export citation and abstract BibTeX RIS
1. Introduction
The multiplicative ergodic theorem plays a fundamental role in the modern theory of dynamical systems, which says that the Lyapunov exponents exist almost everywhere with respect to every given invariant measure, see [1] for more details of Lyapunov exponents. Before Oseledets' work on multiplicative ergodic theorem [2], there are previous work on random multiplication of matrices by Furstenberg and Kesten [3].
Roughly speaking, the multiplicative ergodic theorem generalizes the notion of eigenvalues and eigenvectors for a single matrix to a product of matrices A(fn−1(x)) ⋯ A(f(x))A(x), where A : X → GL(d) is an invertible matrix valued function on a probability space (X, μ) and f : X → X is a measure-preserving map with respect to μ. Under some mild reasonable assumptions, there exists a finite set of numbers and subspaces of which are called Lyapunov exponents and Oseledets subspaces respectively, form a decomposition or a filtration of (depending on whether f is invertible or not). These exponents define the corresponding subspaces of vectors having the same exponential growth rate under the action of the cocycle generated by A.
For infinite dimensional dynamical systems, Ruelle [4] first proved the multiplicative ergodic theorem for compact linear operators on a separable Hilbert space by using the operator theory of Hilbert spaces, the main difficulty in this case are that the non-compactness of the phase space and the non-invertibility of the transformation. Mané [5] overcame the lack of inner product structure and extended the multiplicative ergodic theorem to compact operators on Banach spaces. Later, Thieullen [6] obtained the multiplicative ergodic theorem for bounded linear operators on Banach spaces. Lian and Lu [7] established the multiplicative ergodic theorem for strongly measurable operator cocycles on separable Banach spaces, see [8] and [9] for other versions of the multiplicative ergodic theorem for random dynamics in infinite dimensional spaces as mentioned in [7]. Froyland et al [10] established the multiplicative ergodic theorem in finite dimensional spaces for an arbitrary matrix cocycle over an invertible ergodic measure-preserving transformation of a probability space. Later, the authors in [11] extended it to a continuous semi-invertible operator on Banach spaces, and Tokman et al [12] set up the multiplicative ergodic theorem for a strongly measurable semi-invertible operator on Banach spaces. Recently, Varzaneh and Riedel [13] proved the multiplicative ergodic theorem for a strongly measurable semi-invertible operator on fields of Banach spaces, and constructed the stable and unstable manifolds theorem. Comparing with previous results on the multiplicative ergodic theorem, the fundamental difference is that the authors do not put any measurable (or topological) structure on the random bundle.
The dependence of the exponents and the corresponding Oseledets subspaces on the orbit is usually measurable. The stronger regularity of the dependence has been investigated. Brin [14] studied the special case of a partially hyperbolic C1+ε diffeomorphism on a compact manifold. To be more precisely, for every ergodic measure he proved that the subspace given by the direct sum of the Oseledets subspaces corresponding to strictly negative Lyapunov exponents depend Hölder continuously on the chosen orbit (see also [1], Chapter 5). If f : X → X is a Lipschitz map on a compact metric space and A : X → GL(d) is Hölder continuous, for every ergodic measure Araujo et al [15] established the Hölder continuity of the Oseledets subspaces over a compact set of arbitrarily large measure. Recently, Dragičević et al [16] proved the same result in the setting of possibly non-invertible cocycles, which, in addition, may take values in the space of compact operators on a Hilbert space.
Let f : X → X be a Lipschitz map on a compact metric space X, for an invertible operator cocycles on Banach spaces and an ergodic measure we show that the Oseledets subspaces depend Hölder continuously on a compact set of measure arbitrarily close to 1. In section 2, we recall some basic concepts and the multiplicative ergodic theorem for cocycles on Banach spaces which was proved by Thieullen [6], we also present the statements of the main results here. In section 3, we give the detailed proofs of the main result. For an ergodic measure, using the Lyapunov norm to constructing a regular set which is compact and the measure arbitrarily close to 1, we estimate the distance of the Oseledets subspaces, which depend Hölder continuously on points of the regular set. In section 4, we prove the Hölder continuity of Oseledets subspaces in the non-invertible case. Finally, we add some examples to illustrate our main results in section 5.
2. Preliminaries and statement of the main result
Throughout this paper, unless otherwise specified, let f be a homeomorphism on a compact metric space (X,d), and let μ be an f-invariant ergodic Borel probability measure, and denotes an infinite dimensional Banach space with norm ∥ · ∥.
2.1. Linear cocycles
Let denote the group of invertible bounded linear operators on , the metric ρ on is defined as follows
then is a complete metric space.
Let be a continuous operator valued function, we have that since X is compact. Consequently, we have that . A map is called a linear cocycle over f generated by A, if
for every x ∈ X. Clearly, . Let
whenever the limit exists.
Define the index of compactness (or Kuratowski measure of non-compactness) of a continuous linear operator as the number
where is the unit ball of . It is easy to show that ∥T∥α ≤ ∥T∥ and ∥T◦S∥α ≤ ∥T∥α ∥S∥α for every . Given a linear cocycle over f generated by , the index of compactness at point x is defined as
whenever the limit exists.
By the subadditive ergodic theorem, α(x) and λ(x) is well defined for μ-almost every x, and the function x ↦ α(x) and x ↦ λ(x) are measurable and f-invariant. Since μ is an f-invariant ergodic measure, α(x) and λ(x) are constants for μ-almost every x, denote the constants by α(A, μ) and λ(A, μ) respectively. Note that α(A, μ) ≤ λ(A, μ), and we call A is quasi-compact if α(A, μ) < λ(A, μ).
The following Oseledets multiplicative ergodic theorem of continuous cocycles on Banach spaces was proved by P. Thieullen [6].Multiplicative ergodic theorem.
Theorem 2.1 Let f be a homeomorphism on a compact metric space (X, d), and let be an f-invariant ergodic Borel probability measure. Given a linear cocycle over f generated by the quasi-compact and continuous operator valued function , where is a Banach space. Then there exists a f-invariant subset of full -measure such that one of the following cases hold: Case (1): There exist finite numbers
with and , and for every x ∈ X0 there is a splitting
such that
- (a)For each , is finite and constant. Moreover, , and for any we have
- (b)The distribution is closed and finite-codimensional, satisfies and
- (c)The maps are measurable;
- (d) Writing for the projection of onto via the splitting at x, we have
Case (2): There exist infinite numbers
with such that the following properties hold: for each x ∈ X0 and each positive integer there is a splitting
such that
- (a)For each , is finite and constant for μ-almost every x. Moreover, and for every we have
- (b)The distribution is closed, finite-codimensional and and
- (c)The maps are measurable;
- (d) Writing for the projection of onto via the splitting at x, we have
The number λi in the above theorem is called the i-th Lyapunov exponents of the cocycle with respect to μ and mi are called multiplicities of λi for every i. Moreover, the splitting is called the Oseledets splitting and Ei (x) is called the Oseledets subspaces.
2.2. Gaps and distance between closed linear subspaces
We gather in this subsection some facts that are relevant to Banach spaces. The definitions of gap and distance are taken from Kato [17] (see also [7] and [18]).
Let E and F be two non-trivial closed linear subspaces of the Banach space , and let SE denote the unit ball of E. Put
where , and define the gap between E and F as follows
If is a Hilbert space, then is a metric and coincides with the operator norm of the difference between orthogonal projections. However, in the case that is a Banach space, is not a metric since it does not satisfies the triangle inequality in general [17].
The topology on the set of closed linear subspaces of the Banach space is the metric topology defined by the Hausdorff distance between unit spheres
where . The gap and the distance are related by the following inequality (see [17]):
Hence, in the following we will work with since it is more convenient.
The following lemma gives conditions under which complementation persists, see [18], lemma 3.3 for detailed proofs.
Lemma 2.1. Assume that E is a finite dimensional subspace of , F is a closed subspace of and . Let be the projection operator . If is a finite dimensional subspace of such that , then .
2.3. Statement of the main result
For a cocycle generated by an invertible bounded linear operator on the Banach space and an ergodic measure, this paper proves that the corresponding Oseledets subspaces varies continuously on a compact set of arbitrarily large measure. More precisely, if f : X → X is bi-Lipschitz on a compact metric space and is Hölder continuous, then the Oseledets subspaces Ei (x), Fi (x) and F(x) are Hölder continuous on a compact set of measure close to 1. This extends the main results in [15] for linear cocycles on a Banach space. Dragičević et al [16] also proved the same result for possibly non-invertible cocycles on as well as compact operator cocycles on Hilbert spaces.
Since there is no inner product in our case, the tools in the previous papers do not work here. For the invertible case, we use the methods in [18] to construct coordinate changes between two Oseledets space (see lemma 3.7), which play a key role in our proof. For the non-invertible case, we use the methods in [12] to overcome the lack of orthogonal complement, and get a new complement of Oseledets space and show that it has necessary properties for our purpose in proving the main results (see lemma 4.1).
Theorem 2.2. Let be a bi-Lipschitz homeomorphism on a compact metric space X and an ergodic Borel probability measure on X, and let be a linear cocycle over f generated by a quasi-compact and ν-Hölder continuous function . Let denote the distinct Lyapunov exponents, corresponding to the splitting defined for -almost every . Then, for every , there exist a compact subset of X with , and constants , and such that for all with , we have that for
Remark 2.1. In the above theorem, k is the number of all finite dimensional subspaces and in Case (1), and k is any given positive integer in Case (2).
3. Proofs
We only prove theorem 2.2 for the Case (1), that is, there are finite Lyapunov exponents :
correspoding to the splitting defined on an f-invariant subset X0 of full μ-measure. The Case (2) can be proven in a similar fashion.
3.1. Construction of the regular set
We first recall the definition of the Lyapunov norm. Given a sufficiently small number ε > 0, for each x ∈ X0 and every , the Lyapunov norm is defined as
where for i = 1, ⋯ ,k and
For every n > 0 and every ui ∈ Ei (x)(i = 1, ⋯ ,k), one can easily show that
and for every uk+1 ∈ F(x)
The following lemma provides some fundamental properties of the above Lyapunov norm, see [19], theorem 7.2.3 for detailed proofs.
Lemma 3.1. Given a small number , there exists a measurable function such that for every
and for each
In the setting of the space or the Hilbert space, we have that the angels between two Oseledets subspaces decay sub-exponentially along orbits of x. However, in the case of Banach spaces, there is not such statement since the lack of reasonable definition of the 'angels' between two closed linear subspaces of Banach space. The fact that the norm of each projection operator πi (x), i = 1, ⋯ ,k is temperate (see (d) of theorem 2.1) helps us to overcome this difficulty. For the multiplicative ergodic theorem of semi-invertible operator on Banach spaces (see [11]), it is also valid by the following lemma proved by Dragičević et al [16], lemma 1.
Lemma 3.2. Let be an f-invariant set and let E(x) and be two families of closed subspaces of . Assume that there exist numbers with and measurable functions such that
- (1)and for every
- (2)For every and ,
- (3)For every and ,
- (4)For every and ,
- (5)For every and ,
Then, there exists a measurable function satisfies
and such that
for and .
Remark 3.1. Dragičević et al [16] proved the above lemma for the Euclidean space , it is also valid for the case of Banach space , with a minor modification of the proof. Use the lemma above with some inductions, one can easily to show that (d) of theorem 2.1 is valid for semi-invertible cocycle on Banach space.
Theorem 3.1. Let be a cocycle over f with Lyapunov exponents as in Case (1), for each . Let
Then there exists a full μ-measure, f-invariant subset Λ of X such that, for each small enough with , there are measurable functions with
so that for every :
- (1)For each and ,
- (2)For each and ,
Proof. Following the proof of proposition 3.2 in [20], we give the detailed proofs as follows. By (3), (4) and lemma 3.1, there exist a set Λ and a measurable function with so that the first statement hold. To complete the proof of the theorem, it suffices to prove the second statement.
Let and denote the projections of onto and via the splitting . By (d) of theorem 2.1, for every we have that
Therefore, define the function as follows :
One can easily to show that K is a well defined function on Λ. Moreover, one has that and for every .
Finally, for each and each one has that
and
This completes the proof of the theorem. □
Fix a sufficiently small ε > 0. For every , we define the regular set Λℓ by
One can easily show that each Λℓ is compact, Λℓ ⊂ Λℓ+1 and ⋃ℓ>0Λℓ = Λ. Thus, for every γ > 0, we may choose a subset Λℓ with μ(Λℓ ) > 1 − γ.
3.2. Hölder continuity of maps and
Fix i ∈ {1, ⋯ ,k}, and let and be as in theorem 3.1, we now prove that the map and are (locally) Hölder continuous on Λℓ .
The following two lemmas is useful in the proof of the main result. See [14] for the original versions of the finite dimensional case, we also refer the reader to lemmas 5.3.4 and 5.3.5 in [1] or lemmas 2.1 and 2.2 in [15].
Lemma 3.3. Assume that is ν-Hölder continuous with Hölder constant a1 and is bi-Lipschitz with constant , then there exists a constant such that
for every and .
Proof. We follow the proof of lemma 2.2 in [15] and argue by induction on n.
For k = 1, by the Hölder continuity of the map and (1) we have
Assume that there exists so that (8) hold for . For , we have that
where we use the fact that f is Lipschitz in the last inequality. To find the number a, all we need is that
that is
This can be easily achieved by taking a sufficiently large a such that
The case for can be proven in a similar fashion, this completes the proof. □
Lemma 3.4. Let be two sequences of operators in , such that, for some and , there exist closed subspaces and of satisfying such that for some fixed n
- (i)and for every ,
- (ii) and for every ,
- (iii) for each or .
Then for every satisfying
we have that
Proof. With a minor modification of the proof of lemma 2.1 in [15], we give the proof of the result in the following.
Let us define the cone . For each , one has
In consequence, and this implies that .
For each , write where and . Recall that , one has
This together with the fact that imply that
Since , by the definition of one has
Symmetrically, one can show that . Hence, using (2) we conclude that
□
Now, we estimate and for each x, y ∈ Λl with d(x, y) < 1, where is the same as in theorem 3.1.
Lemma 3.5. For each with , we have
where and are two constants.
Proof. Given with , let and . By theorem 3.1 and the definition of , the conditions (i), (ii), (iii) of lemma 3.4 hold for every . Set . By lemma 3.3, there exists a sufficiently large constant a such that and
for each . Moreover, there exists such that
It follows from lemma 3.4 that
where and □
Remark 3.2. Notice the hypotheses that A(x) is invertible is not used in the estimate of , and theorem 3.1 is also valid for semi-invertible cocycles (see [20], proposition 3.2). Thus, the conclusion of the above lemma is true for semi-invertible operators on a Banach space.
Lemma 3.6. For each with , we have
where and are two constants.
Proof. In order to estimate for every with , set for every , and . Take , since , it follows from theorem 3.1 that
where we use the facts that and . Consequently, one has
Similarly, one can show that for every . Thus, the condition (i) of lemma 3.4 holds. Replace x with y, one can show that the condition (ii) of lemma 3.4 holds in similar fashion, and the condition (iii) follows from (2) of theorem 3.1. Let and let the constant a be as lemma 3.4. It follows from lemma 3.4 that
where and □
3.3. Hölder continuity of the map x ↦ Ei (x)
In this section, we will give the proof of the main result of this paper.
For every γ > 0, fix Λℓ so that μ(Λℓ ) > 1 − γ. Take a point x ∈ Λℓ , then , here are the same as in theorem 3.1. By (10), choose a small number δ1 ∈ (0, 1) such that
for each x, y ∈ Λℓ with d(x, y) < δ1. By the definition of Λℓ and (7), the norm of the projection operator is no larger than ℓ, thus
By lemma 2.1, this yields that . So, there exists a linear operator such that the graph of Lx,y is equivalent to the subspace , that is,
Lemma 3.7. For each with , we have
Proof. For simplicity of presentation, denote by L. First, for each , the fact that implies
Then, we have that . Next, given , there exist such that . Since , it follows from (2) of theorem 3.1 that . Thus, one has
Therefore, . This together with (2) yield that
To prove the other inequality in (11), note that for each and for every , one has
Fix such a , there exists such that . Since , and , by (2) of theorem 3.1 we have that
Thus, we conclude that
for every . This implies that
By the arbitrariness of β and (2), one has
This completes the proof of the lemma. □
Next, we shall prove theorem 2.2.
Proof of theorem 2.2. By lemma 3.7, we have that
Hence, there exists with such that for any with .
Fix with . Let be the isomorphism from to . It is easy to see that . Since , we obtain that . We write it as , where and one can show that .
For simplicity of notations, denote and by and respectively. Note that
By the triangle inequality we have
We first estimate in (12). For each with , we have that
Similarly, for each with one has that
Consequently, we obtain that . By (2) and lemma 3.7, we have
where the last inequality use the fact that . Combining the above inequality and (10), we have that
Next we estimate . Let and .
For each and each , by (1) of theorem 3.1
where we use the fact that in the last inequality.
For each and each , by (1) of theorem 3.1 we have
where the third inequality uses the fact that .
For every , then and, by (2) of theorem 3.1 one has that
This yields that
Since , we have that . Therefore, one has that
Consider and replace ℓ by . Clearly, the sequence satisfies condition (i) of lemma 3.4, and it follows from (14), (15) and (16) that the condition (ii) and (iii) of lemma 3.4 hold for every .
At last, consider a constant as in lemma 3.3, one can show that
Let , since there exist such that
It follows from lemma 3.4 that
where and . By lemma 3.7 and (10), there exist constants and so that
Since implies , this together with (17) one has
Combining (12), (13) and (18), we have that
where and . This completes the proof. □
4. The non-invertible case
In this section, assume that f is a non-invertible Lipschitz map on a compact metric space (X, d), and μ is an f-invariant ergodic Borel probability measure on X.
Let be a separable Banach space, and let denote the space of all bounded linear operators on . Let be a ν-Hölder continuous operator valued function, i.e., there exist constants C and ν such that . In addition, suppose that α(A, μ) < λ(A, μ), i.e., A is quasi-compact (see the definitions in section 2.1).
The one-sided Oseledets theorem was obtained by Doan [19], theorem 7.1.7 that we will recall in below, see [7] for the two-sided case.
Theorem 4.1. Let f be a continuous map on a compact metric space (X, d), and let μ be an f-invariant ergodic Borel probability measure. Given a linear cocycle over f generated by a quasi-compact and strongly measurable operator valued function . Then there exists an f-invariant subset of full μ-measure such that there exist k () numbers , for every x ∈ X0 there is a filtration
with the following properties :
- (a)If , let , for each , Vi (x) is a closed, finite-codimensional subspace and co- is a finite constant. Moreover, , and for every we have that
- (b)If then , and Vi (x) is a closed, finite-codimensional and co- is a finite constant for each . Moreover, , and for every we have that
Remark 4.1. Notice that the continuity of implies that A is strongly measurable, and the map is measurable (see [12], remark 2.9).
In the following, we will show that the map x ↦ Vi (x) is (locally) Hölder continuous on a compact set of arbitrarily large measure provided that f is Lipschitz and is Höder continuous.
Theorem 4.2. Let f be a Lipschitz map on a compact metric space (X ,d), and let μ be an f-invariant ergodic Borel probability measure. Given a linear cocycle over f generated by a quasi-compact and ν-Höder continuous operator valued function , where is a separable Banach space. Then, for every and every subspace Vi (x) as in theorem 4.1, there exists a compact subset with so that the map is (locally) Hölder continuous on .
If k is a finite positive integer, we will show that the previous theorem for every i = 1, 2, ⋯ ,k + 1. If k is infinite, we will prove the same result for every . In the following, we only prove the first case, the latter can be proven in a similar fashion.
The following lemma proved in [12], lemma 2.11 allows us to choose a 'good' complementary space for every space Vi (x). In the setting of the space or the Hilbert space, one may usually choose the orthogonal complement .
Lemma 4.1. Let the filtration be as in theorem 4.1. Then, for every , there exists a finite-dimensional subspaces such that
- (1)For μ-almost every x, and the map is measurable;
- (2)Let (where ), and let and denote the projections of onto Ui (x) and Vi (x) via the splitting respectively, then
Since , the statement in theorem 4.2 clearly holds. In the following, fix i ∈ {2 ⋯ , k + 1}. By (2) of lemma 4.1, on can choose ℓ > 0, such that
for μ-almost every x ∈ X. Reducing X0 by a zero measure set such that X0 remains f-invariant and for every x ∈ X0 the above inequality holds for every x ∈ X0. Then, for every x ∈ X0, one has
for every u ∈ Ui (x) and every v ∈ Vi (x). By theorem 4.1, for every x ∈ X0 and every u ∈ Ui (x)⧹{0} we have that
Fix x ∈ X0, let , and . Since by the invariance of Vi (x), we have that
Similarly, define Bn (x), Cn (x) and Dn (x) as above, such that . As in [12], lemma 2.12 (see also [21], Sect. 4.2.5), One can show that
and
Note that , and . For every u ∈ Ui (x), by (20) we have that
This implies that
Moreover, if there exists u ∈ Ui (x)⧹{0} such that B(x)u = 0, then
for every n > 1. Since C(x)u ∈ Vi (f(x)), one has
This yields a contradiction with (21). Hence, one has B(x)u ≠ 0 for each u ∈ Ui (x)⧹{0}. Therefore, B(x) is an isomorphism from Ui (x) to Ui (fx). Note that , restricted on {Ui (x)}, {Bn (x)} is a cocycle on X with respect to f, applying the multiplicative ergodic theorem of the finite dimensional case (e.g., see [21]), one can obtain that for μ-almost every x ∈ X the following limit
exist for every u ∈ Ui (x)⧹{0}. Moreover, by [12], Sub-lemma 2.13 or [21], proposition 4.14, we have that for every u ∈ Ui (x)⧹{0}
This implies that
Without loss of generality, assume that (22) holds for every x ∈ X0.
Lemma 4.2. Fix x ∈ X0 and a small number , there exists such that for each with , we have for each .
Proof. For every , write . Note that and , it follows from (20) that
Therefore, we have that
This implies the desired result immediately. □
Next, we will show that the map x ↦ Vi (x) is (locally) Hölder continuous on a compact subset of arbitrarily large measure.
Proof of theorem 4.2. Fix and a sufficiently small number . For each , let
and
Clearly, the sequences of sets and are nested, and by theorem 4.1 and lemma 4.2 we have that
Therefore, for every there exists n0 such that for every .
Let . We may assume further that is compact since otherwise we can approximate it from within by a compact subset. For every with , where a is a sufficiently large constant as in lemma 3.3. Note that lemma 3.3 is also valid for in this case. Applying lemma 3.4 with and . By the construction of the set , for every , the conditions (i), (ii) and (iii) of lemma 3.4 hold. Set , there exists a number such that
It follows from lemma 3.4 that
where and This completes the proof. □
5. Applications
In this section, we give two applications of our main theorems, theorem 2.2 and theorem 4.2.
5.1. Derivative cocycles on Banach space
Let be a separable Banach space, and let be a C2 Fréchet differentiable map. We denote by Df(x) the derivative of f at point x and assume that the map x ↦ Df(x) is quasi-compact. This class of mapping includes time-t maps of semi-flows defined by periodically forced nonlinear dissipative parabolic PDEs (see [18] for details). We consider a compact attractor A, that is f(A) ⊂ A. Note that dynamical systems defined by dissipative PDEs are known to have attractors [22]. Let μ be an f-ergodic invariant Borel probability measure supported on A.
We consider the derivative cocycles on Banach space defined by
By theorem 4.1, for each ε > 0, there exists an f-invariant subset A0 of full μ-measure such that there exist numbers λ1 > ⋯ > λk > α(A, μ) + ε. For each x∈A0, there is a filtration
satisfying theorem 4.1.
Therefore, by theorem 4.2 we have that the maps x ↦ Vi (x) (i = 1, ⋯ ,k + 1) are (locally) Hölder continuous on a compact set of arbitrarily large μ-measure. Moreover, if f(A) = A and f is a C2 Fréchet diffeomorphisms on A. Then, for each x∈A0, there is a splitting
satisfying theorem 2.1. By theorem 2.2 we have that x ↦ Ei (x) (i = 1, ⋯ ,k) and x ↦ F(x) are (locally) Hölder continuous on a compact subset of arbitrarily large μ-measure.
5.2. Products of random operators
Let be a Banach space, and let be a sequence of operators in . We consider the compact metric space with the following metric d :
where 0 < β < 1 is a constant and {ai }, {bi } ∈ Σ. Let θ be the left shift map on Σ, and let μ be a θ-ergodic probability measure on Σ. Define the operator valued function by . Note that A is Hölder continuous since it is locally constant. Assume further that A is quasi-compact.
For each ε > 0, by theorem 2.2 there exists a θ-invariant subset Σ0 of full μ-measure and numbers λ1 > ⋯ > λk > α(A, μ) + ε such that for each x ∈ Σ0, there is a splitting
satisfying theorem 2.1. Applying theorem 2.2 one can obtain that the maps x ↦ Ei (x) (i = 1, ⋯ ,k) and x ↦ F(x) are (locally) Hölder continuous on a compact subset of arbitrarily large μ-measure.
Acknowledgments
The authors are grateful to the anonymous referees for valuable comments which helped to improve the manuscript greatly. They also would like to thank Professor Jianyu Chen for his suggestions.
Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.
Footnotes
- *
This work is partially supported by NSFC (12271386, 11871361), and the second author is partially supported by Qinglan project from Jiangsu Province.