A version of Oseledets for proximal random dynamical systems on the circle

Basic Information

• Paper ID: 2503.16026
• Title: A version of Oseledets for proximal random dynamical systems on the circle
• Authors: Jamerson Bezerra (Universidade Federal do Ceará), Graccyela Salcedo (Centre de Physique Théorique, CNRS, Institut Polytechnique de Paris)
• Classification: math.DS (Dynamical Systems)
• Publication Date: November 15, 2025 (arXiv v2)
• Paper Link: https://arxiv.org/abs/2503.16026v2

Abstract

This paper studies proximal random dynamical systems on the circle without common fixed points. The authors prove the existence of two random points that control the behavior of forward and backward orbits of the system. Under the assumption of differentiability of the maps, these random points are characterized through extremal Lyapunov exponents. As an application, the exact dimensionality properties of stationary measures in this setting are established.

Research Background and Motivation

Research Problem

This paper aims to generalize the classical Oseledets theorem from linear cocycles to nonlinear random dynamical systems generated by general homeomorphisms on the circle.

Importance of the Problem

1. Theoretical Significance: The Oseledets theorem is a cornerstone of dynamical systems theory, describing the existence of stable and unstable directions in linear systems. Extending it to the nonlinear case has significant theoretical value.
2. Statistical Properties: Understanding the asymptotic behavior of random dynamical systems is crucial for validating numerical simulations and establishing statistical properties such as laws of large numbers and central limit theorems.
3. Geometric Structure: Dimension theory of stationary measures reveals the geometric and probabilistic structure of the system, which is important for understanding system complexity.

Limitations of Existing Methods

1. Linear Assumption: The classical Oseledets theorem primarily applies to actions of GL₂(ℝ) matrices on S¹, relying on linear structure.
2. Strong Hypotheses: Existing work such as Kleptsyn and Nalskii KN04 requires stronger assumptions (minimality of forward and backward systems, and existence of "north pole/south pole" type maps).
3. Orientation-Preserving Requirement: Many results, such as He et al. HJX23, assume that maps preserve the orientation of the circle, which limits the scope of applications.

Research Motivation

The authors bypass the orientation-preserving and strong minimality assumptions through proximality, a topological condition, establishing Oseledets-type results in a more general framework.

Core Contributions

1. Main Theorem (Theorem 1): Proves that proximal random dynamical systems without common fixed points admit two random points π(ω) and θ(ω), serving as "attracting direction" and "repelling direction" respectively, controlling the behavior of forward and backward orbits.
2. Lyapunov Exponent Characterization (Theorem 3): In the differentiable case, the random directions π and θ are characterized using extremal Lyapunov exponents Λ(ν) and λ(ν).
3. Uniqueness of Stationary Measures (Corollary 2): Proves uniqueness of stationary measures η and η⁻, expressing them as distributions of random directions.
4. Exact Dimension (Theorem 5): Establishes the exact dimension formula for stationary measure η: $\dim(η) = -\frac{h_F(η,ν)}{λ(ν)}$ where h_F is the Furstenberg entropy.
5. Methodological Innovation: Combines Malicet's Mal17 results on local contraction with Hochman-Solomyak's HS17 dimension analysis techniques.

Detailed Methodology

Task Definition

Input:

• Probability measure ν on the group of circle homeomorphisms Hom(S¹)
• Satisfying proximality and without common fixed points

Output:

• Random "stable" direction θ: X_ν^ℕ → S¹
• Random "unstable" direction π: X_ν^ℕ → S¹
• Characterization of extremal Lyapunov exponents
• Exact dimension of stationary measures

Core Concepts

1. Proximality

The semigroup Γ_ν is proximal if for all x,y ∈ S¹, there exists a sequence (g_n) ⊂ Γ_ν such that: $\lim_{n→∞} d(g_n(x), g_n(y)) = 0$

This means the system can synchronize orbits from different initial points.

2. Absence of Common Fixed Points

There does not exist x ∈ S¹ such that f(x) = x for all f ∈ Γ_ν. This is a natural condition ensuring nontrivial dynamics, analogous to the "gap condition" in Oseledets theorem.

3. Extremal Lyapunov Exponents

For measure ν on Diff^{1+τ}(S¹), define: $Λ(ν) := \lim_{n→∞} \frac{1}{n}\log\sup_{z∈S¹}|(f_n ∘ ··· ∘ f_1)'(z)|$$λ(ν) := \lim_{n→∞} \frac{1}{n}\log\inf_{z∈S¹}|(f_n ∘ ··· ∘ f_1)'(z)|$

Technical Architecture

Step One: Existence of Repelling Points (Proposition 13)

Through local contraction properties, for ν^ℕ-almost every ω, construct point θ(ω) such that:

• For any interval I not containing θ(ω), $\limsup_{n→∞} \frac{1}{n}\log\text{diam}(f_n^ω(I)) ≤ \log q < 0$

Proof Strategy:

1. Use Proposition 12: For any x,y, at least one connecting arc x,y or y,x contracts at exponential rate q
2. For a dense set Q, define sets I⁺ and I⁻
3. θ(ω) serves as the boundary point of these two sets

Step Two: Existence of Attracting Points (Proposition 15)

Prove existence of π(ω) such that: $\lim_{n→∞} \bar{f}_n^ω * η = δ_{π(ω)}$

Proof Strategy:

1. Apply Guivarc'h-Raugi GR86 result: sequence $(\bar{f}_n^ω * η)$ converges in weak* topology
2. Through subsequence arguments and orientation/reverse analysis, prove the limit must be a Dirac measure
3. Use Corollary 14 (existence of sequence (g_n) such that (g_n)_*η → δ_z) to complete proof

Step Three: Synchronization (Proposition 17)

Prove for all x ∈ S¹: $\lim_{n→∞} \bar{f}_n^ω(x) = π(ω)$

This is completed through proof by contradiction and density arguments.

Technical Innovations

1. Flexible Use of Proximality

Proximality ensures both orbit synchronization (similar to contraction) and prevents existence of invariant measures (Proposition 6-7), a key duality.

2. Symmetric Treatment of Forward and Backward

Through construction of inverse measure ν⁻ (formula 13): $ν^-(·) = ν(\{f : f^{-1} ∈ ·\})$ and proving it also satisfies proximality (Proposition 23), achieving unified treatment of forward/backward.

3. Characterization of Extremal Points (Proposition 21-22)

Proves that points where derivatives achieve maximum values must converge to θ(ω), establishing connection between geometry (θ point) and analysis (extremal Lyapunov exponent): $Λ(ν) = \lim_{n→∞} \frac{1}{n}\log|(f_n^ω)'(θ(ω))|$

4. Distortion Control (Lemma 19)

On closed intervals away from θ(ω), establishes uniform distortion estimates: $e^{-nδ} ≤ \frac{|(f_n^ω)'(x)|}{|(f_n^ω)'(y)|} ≤ e^{nδ}$

This is crucial for dimension analysis.

Experimental Setup

This is a pure theoretical mathematics paper without numerical experiments. However, Section 5 provides theoretical examples.

Theoretical Examples (Example 27)

Constructs proximal systems that cannot be topologically conjugate to linear projective maps:

Case 1:

• Map f has 4 fixed points (2 attractors, 2 repellers), with two repelling points very close to an attracting point
• Map g is an irrational rotation

Case 2:

• Map f has N>2 fixed points, with at least one being an attractor
• Map g has unique fixed point, topologically parabolic type

These examples demonstrate the nontriviality and broad applicability of the theory.

Experimental Results

Main Theoretical Results

Significance of Theorem 1

For proximal systems without common fixed points, there exist measure maps π, θ: X_ν^ℕ → S¹ satisfying:

1. Global Attractivity: For all x ∈ S¹, $\lim_{n→∞} f_1 ∘ ··· ∘ f_n(x) = π(ω)$
2. Exponential Contraction: For closed set A ⊂ S¹{θ(ω)}, $\lim_{n→∞} \text{diam}(f_n ∘ ··· ∘ f_1(A)) = 0$ with contraction rate q (local contraction rate)
3. Equivariance: $f_1(π(σω)) = π(ω), \quad f_1(θ(ω)) = θ(σω)$

Precise Characterization of Theorem 3

In the differentiable case, Lyapunov exponent values at different points:

This completely characterizes the hyperbolic structure of the system.

Integral Representation of Corollary 4

$Λ(ν) = -\int\log|f'(x)|d(η^- ⊗ ν)(x,f)$$λ(ν) = \int\log|f'(x)|d(η ⊗ ν)(x,f)$

This connects extremal exponents with stationary measures.

Dimension Formula of Theorem 5

Stationary measure η has exact dimension, and: $\dim(η) = -\frac{h_F(η,ν)}{λ(ν)} ∈ (0,1]$

where Furstenberg entropy: $h_F(η,ν) := \int\log\frac{df_*η}{dη}(f(x))d(η⊗ν)(x,f)$

Role of Key Lemmas

Lemma 11: Proves forward synchronization implies backward synchronization, which is nontrivial since forward and backward compositions have the same finite-time distribution but different asymptotic behavior.

Proposition 12: Refines results from GS23, not only ensuring synchronization but also guaranteeing that some connecting arc between two points contracts at exponential rate.

Proposition 26: Critical for dimension analysis, controls distance to π(ω) under the map for points away from set Z_n(ω,ε).

Related Work

Classical Theoretical Foundations

1. Oseledets Theorem Ose68: Existence of Lyapunov exponents and characterization of stable/unstable directions in linear cocycles
2. Furstenberg Fur73: Introduced proximality concept, studied Lie group boundaries
3. Guivarc'h-Raugi GR86: Proved typicality of proximality in random dynamical systems induced by linear maps

Nonlinear Generalizations

1. Kleptsyn-Nalskii KN04: Requires bidirectional minimality and special maps
2. Malicet Mal17: Establishes local contraction theory and invariance principles
3. Gelfert-Salcedo GS23, GS24: Study statistical properties of contracting systems

Dimension Theory

1. Hochman-Solomyak HS17: Exact dimension in linear case
2. He-Jiao-Xu HJX23: Dimension theory for orientation-preserving diffeomorphisms (stronger assumptions)

Advantages of This Paper

• Weaker Assumptions: No need for orientation-preserving or bidirectional minimality
• More General Framework: Applies to general homeomorphisms, not limited to diffeomorphisms
• Innovative Methods: Combines topological (proximality) and analytical (Lyapunov exponents) techniques

Conclusions and Discussion

Main Conclusions

1. Proximality + absence of common fixed points ⇒ existence of random Oseledets directions
2. In the differentiable case, these directions can be characterized using extremal Lyapunov exponents
3. Stationary measures have exact dimension given by the ratio of entropy to Lyapunov exponent

Theoretical Significance

• Extends Oseledets theorem from linear to nonlinear
• Reveals deep structure of proximality
• Establishes unified framework connecting topology, measure theory, and differential properties

Limitations

1. Proximality Assumption: While typical in linear case, its universality in general nonlinear case is unknown
2. Circle Restriction: Methods highly depend on one-dimensional topology of S¹, generalization to higher dimensions is difficult
3. Hölder Continuity: Theorems 3 and 5 require Diff^{1+τ} regularity
4. Non-constructive: Existence of random directions π and θ is non-constructive

Future Directions

While not explicitly stated in the paper, potential directions include:

1. Study typicality conditions for proximality in nonlinear systems
2. Generalize to higher-dimensional manifolds (may require additional structure)
3. Study weakened versions for non-Hölder case
4. Explore deeper connections with ergodic theory (e.g., SRB measures)

In-Depth Evaluation

Strengths

1. Theoretical Depth

• Major Generalization: Successfully extends classical Oseledets theorem to nonlinear setting, addressing a long-standing important problem
• Unified Framework: Organically combines topology (proximality), measure theory (stationary measures), differential properties (Lyapunov exponents), and dimension theory
• Technical Sophistication: Skillfully applies Malicet's local contraction theory and Guivarc'h-Raugi's convergence theorem

2. Reasonableness of Assumptions

• Proximality is a natural topological condition satisfied in many systems
• Absence of common fixed points ensures nontrivial dynamics
• Significantly weaker assumptions than existing work (e.g., KN04)

3. Completeness of Results

• Forms complete chain from existence (Theorem 1) to characterization (Theorem 3) to applications (Theorem 5)
• Symmetric treatment of forward and backward systems (via ν⁻) reflects deep insight
• Corollaries 2 and 4 provide operationalizable measure characterizations

4. Proof Rigor

• Each step has detailed proofs
• Key lemmas (e.g., Lemma 11, 18, 19) are well-prepared
• Counterexamples and examples (Example 10, 27) demonstrate necessity of assumptions

Weaknesses

1. Readability Issues

• Complex notation system (f_n^ω vs f̄_n^ω, π vs θ, η vs η⁻)
• Lacks intuitive diagrams explaining geometric meaning of random directions
• Main theorems stated in technical form, lacking accessible explanations

2. Limited Applications

• Theoretical examples (Example 27) are somewhat artificial
• No discussion of applications to actual physical or engineering systems
• Absence of numerical verification (though this is pure theory, numerical examples would strengthen credibility)

3. Technical Limitations

• Highly dependent on one-dimensional topological structure of circle
• Hölder continuity assumption may be too strong for some applications
• Verification of proximality in practical systems may be difficult

4. Comparison with Existing Work

• Insufficient detailed comparison with HJX23 (though stronger assumptions are mentioned)
• No discussion of whether results are stronger in orientation-preserving case
• Relationship with Barrientos-Malicet BM24 "mostly contracting" framework could be deeper

Impact Assessment

Academic Impact

• High Impact: Resolves fundamental questions in random dynamical systems theory
• Pioneering: Lays foundation for Lyapunov theory of nonlinear RDS
• Citation Potential: Expected to become important reference in the field

Practical Value

• Theoretical Tool: Provides mathematical framework for analyzing complex random systems
• Statistical Applications: Supports establishment of statistical properties like laws of large numbers and central limit theorems (e.g., GS24)
• Numerical Simulation: Provides theoretical benchmark for validating numerical simulations

Reproducibility

• Theoretical Reproducibility: Complete proofs, clear logic
• Computational Implementation: Though no code provided, theoretical framework can guide numerical method development
• Generalizability: Methods may apply to other one-dimensional compact manifolds

Applicable Scenarios

1. Theoretical Research

• Ergodic theory of random dynamical systems
• Hyperbolicity theory of nonlinear systems
• Fractal geometry and dimension theory

2. Potential Applications

• Climate Models: Circulation systems under random perturbations
• Neuroscience: Synchronization of random neural oscillators
• Financial Mathematics: Stochastic volatility models
• Control Theory: Stability of random switching systems

3. Methodological Contributions

• Proximality as alternative to contractivity
• Geometric interpretation of extremal Lyapunov exponents
• Generalization of dimension formulas to nonlinear case

In-Depth Analysis of Technical Highlights

1. Cleverness of Proposition 6

Proves proximality + invariant measure ⇒ common fixed point. Key steps:

• Using proximality, at least one of any two positive-measure arcs is contracted
• Through measure preservation, contracted arc's measure concentrates at single point
• Applying proximality again, this point must be common fixed point

This elegant argument avoids using Mal17's complex structure theorem.

2. Non-triviality of Lemma 11

Forward synchronization ⇒ backward synchronization proof uses probabilistic techniques:

• Construct map ϕ reordering sequences
• Using product structure of ν^ℕ, prove set measures are equal
• Monotone convergence theorem completes proof

This reveals deep symmetry between forward and backward dynamics.

3. Geometric Insight of Proposition 21

Proof that maximum derivative points converge to θ(ω) combines:

• Local contraction (Lemma 19)
• Integral representation of Lyapunov exponent (Lemma 20)
• Proof by contradiction

This provides analytical characterization of "repelling direction."

4. Proof Strategy of Theorem 5

Exact dimension proof adapts HS17, key modifications:

• Replace linear structure with proximality
• Use Proposition 26 to control points away from Z_n(ω,ε)
• Maker theorem handles accumulated distortion terms

This demonstrates how to generalize linear techniques to nonlinear case.

In-Depth Comparison with Related Literature

Open Problems

Though not explicitly listed in the paper, naturally arising questions include:

1. Universality of Proximality: Under what conditions are random systems of circle homeomorphisms proximal?
2. Higher-Dimensional Generalization: Can this be generalized to higher-dimensional tori T^n or more general manifolds? What additional structure is needed?
3. Regularity: Can Hölder continuity be weakened to merely C¹ or Lipschitz?
4. Multiple Stationary Measures: What is the structure of stationary measures for non-proximal systems?
5. Ergodicity: What are conditions for ergodicity of stationary measure η?
6. SRB Measures: Does random direction π(ω) always correspond to some SRB-type measure?
7. Numerical Algorithms: How to effectively compute extremal Lyapunov exponents and random directions?

Selected References

The paper cites 22 references, key ones include:

1. Ose68 Oseledets - Multiplicative ergodic theorem (original)
2. Fur73 Furstenberg - Proximality concept and boundary theory
3. GR86 Guivarc'h-Raugi - Convergence theorem for random matrix products
4. Mal17 Malicet - Random walks of circle homeomorphisms
5. HS17 Hochman-Solomyak - Dimension of Furstenberg measures
6. Via14 Viana - Lyapunov exponents lectures (modern reference)
7. GS23, GS24 Gelfert-Salcedo - Statistical properties of mean contracting systems

Summary

This is a high-quality theoretical mathematics paper that successfully generalizes the classical Oseledets theorem to nonlinear random dynamical systems. Main contributions include:

1. Theoretical Breakthrough: Establishes nonlinear Oseledets theorem under weak assumptions
2. Methodological Innovation: Skillfully combines proximality and local contraction theory
3. Complete Results: Forms complete system from existence to characterization to applications

Paper's main limitations are:

• Highly dependent on one-dimensional topology
• Lacks application examples and numerical verification
• Technical presentation could improve readability

Nevertheless, this work makes important contributions to random dynamical systems theory and is expected to have lasting impact on the field. For scholars researching nonlinear random systems, fractal geometry, or ergodic theory, this is essential reading.