Universal Averaging for Ergodic Flows

Basic Information

• Paper ID: 2511.09229
• Title: Universal Averaging for Ergodic Flows
• Author: Ryzhikov V.V.
• Classification: math.DS (Dynamical Systems)
• Submission Date: November 12, 2025 to arXiv
• Paper Link: https://arxiv.org/abs/2511.09229

Abstract

This paper investigates the universal averaging problem for ergodic flows. The author defines the universality of a sequence of normalized Borel measures $\nu_j$ on $\mathbb{R}$: for any $f\in L_1(X,m)$ and any ergodic flow, the following convergence holds: $\left\|\int_\mathbb{R} f(T_{r}x) d\nu_j(r) -\int f \, dm\right\|\to 0, \quad j\to \infty$

The main result demonstrates that if for some $n$, the convolution powers $\nu_j^{*n}$ form a universal sequence for ergodic flows, then $\nu_j$ itself is also universal. This generalizes an important result of Kozlov-Treschev.

Research Background and Motivation

Problem Background

1. Generalization of Classical Ergodic Theory: Classical ergodic theorems consider uniform averages (such as $h = \chi_{[0,1]}$), while Kozlov and Treschev's pioneering work in 2003 introduced the concept of weighted homothetic averages in the form: $P_t f(x) = \int_\mathbb{R} f(T_{rt}x)h(r)dr$
2. Limitations of Known Results: The Kozlov-Treschev Theorem 1.1 proves that for normalized measures $\nu$ absolutely continuous with respect to Lebesgue measure, the above average converges to a constant. However, this leaves a core question: Which singular measures also possess universality?
3. Research Motivation:
   • The "convolution roots" of absolutely continuous measures may be singular
   • Need to understand the stability of universality under convolution operations
   • Explore the role of singular measures in ergodic theory

Significance

This problem has foundational significance in dynamical systems theory because:

• It extends the applicability of ergodic theorems
• It reveals deep connections between the convolution structure of measures and ergodic properties
• It provides new tools for multiple mixing properties and non-uniform averaging theory

Core Contributions

1. Main Theoretical Result (Theorem 1.2/2.1): Proves the stability of universality under "extracting convolution roots"—if some convolution power $\nu^{*n}$ of a measure $\nu$ is universal, then $\nu$ itself is also universal
2. Constructive Results: Using spectral measures of Sidon automorphisms, provides concrete examples of singular universal measures—these measures are singular at low-order convolution powers but become absolutely continuous at order $d \geq n$
3. Characterization of Universal Flows:
   • Proves that all mixing flows are universal for the continuous measure class (Theorem 3.1)
   • Proves that flows with rigid factors are not universal (Theorem 3.2)
   • Constructs examples of non-mixing but universal flows (Remark 3.3)
4. Open Problems: Do there exist universal measures all of whose convolution powers are singular?

Methodology Details

Task Definition

Input:

• An ergodic flow $\{T_t\}_{t\in\mathbb{R}}$ on a probability space $(X, \mathcal{B}, m)$
• A normalized Borel measure $\nu$ on $\mathbb{R}$ (or a sequence of measures $\{\nu_j\}$)
• A function $f \in L_1(X,m)$

Output: Determine whether the measure (sequence) is universal, i.e., whether: $\left\|\int_\mathbb{R} f(T_r x) d\nu(r) - \int f \, dm\right\|_{L_1} \to 0$

Core Technical Framework

1. Operator-Theoretic Approach

Define the integral operator $P$: $Pf(x) = \int_\mathbb{R} f(T_r x) d\nu(r)$

Key observations:

• If $P$ corresponds to measure $\nu$, then $P^n$ corresponds to measure $\nu^{*n}$ (n-fold convolution)
• $P$ acts as the identity on constant functions
• Restricted to the space of zero-mean functions $L_1^0$, denoted as $Q$

2. Normal Operator Lemma (Core Technique)

Lemma: Let $Q_j$ be a sequence of normal operators in $L_2$ (i.e., $Q_j^* Q_j = Q_j Q_j^*$). If for all zero-mean functions $f \in L_2$, we have $\|Q_j^n f\| \to 0$, then $\|Q_j f\| \to 0$.

Proof Strategy (elementary argument independent of spectral representation):

1. From $\|Q_j^n f\| \to 0$ and operator commutativity, obtain: $(Q_j^* Q_j)^{2n} f \xrightarrow{w} 0$
2. This is equivalent to: $(Q_j^{2^n-1} f, Q_j^{2^n-1} f) \to 0$
3. Iteratively reduce powers: $(Q_j^* Q_j)^{2^{n-1}} f \xrightarrow{w} 0, \ldots, Q_j^* Q_j f \xrightarrow{w} 0$
4. The last convergence is equivalent to $\|Q_j f\| \to 0$

3. Proof Strategy for Theorem 2.1

1. Density Argument: $L_2$ functions in $L_1^0$ are dense
2. Norm Control: $\|Q_j g\|_1 \leq \|Q_j g\|_2 \to 0$ (for $L_2$ functions)
3. Approximation Argument: For any $f \in L_1^0$ and $\epsilon > 0$, choose $g \in L_2$ such that $\|f-g\|_1 < \epsilon$, then for sufficiently large $j$: $\|Q_j f\|_1 \leq \|Q_j(f-g)\|_1 + \|Q_j g\|_1 \leq \epsilon + o(1)$

Technical Innovations

1. Stability of Convolution Roots: First systematically establishes the stability of universality under convolution root extraction, a non-trivial result since singularity may disappear under convolution
2. Elementary Treatment of Normal Operators: Avoids using the spectral theorem, providing a completely elementary proof that makes the result more self-contained
3. Application of Sidon Automorphisms: Innovatively uses spectral properties of Sidon automorphisms to construct concrete examples, connecting abstract functional analysis results with concrete dynamical systems
4. Cantor-Type Measure Construction: In the proof of Theorem 3.2, constructs a Cantor-like measure that cleverly uses rigid sequences $t(i)$ and slowly growing sequences $m(i)$

Experimental Setup

Note: This is a pure mathematics theoretical paper with no numerical experiments or datasets. All results are rigorous mathematical proofs.

Theoretical Verification Methods

1. Constructive Examples:
   • Uses Sidon automorphisms from reference 4 to construct concrete singular universal measures
   • Transforms spectral measures on the circle to measures on $\mathbb{R}$ via special flows
2. Counterexample Construction:
   • For Theorem 3.2, constructs a refined Cantor-type measure that fails to converge to a constant at rigid times
3. Existence Proofs:
   • For Remark 3.3, uses methods from reference 5 to construct non-mixing but universal flows

Experimental Results

Main Theoretical Results

Theorem 1.2 (Single Measure Version)

If some convolution power $\nu^{*n}$ of a normalized Borel measure $\nu$ is universal for ergodic flows, then $\nu$ itself is universal.

Significance: This allows us to derive the existence of singular universal measures from known absolutely continuous universal measures (guaranteed by Theorem 1.1).

Theorem 2.1 (Measure Sequence Version)

If a sequence of measures $\{\nu_j\}$ has convolution power sequence $\{\nu_j^{*n}\}$ that is universal, then $\{\nu_j\}$ itself is universal.

Corollary: Combined with Kozlov-Treschev's result, any measure whose convolution power is absolutely continuous is universal.

Theorem 3.1 (Universality of Mixing Flows)

All mixing flows are universal for the continuous measure class.

Proof Highlights: Uses the mixing property $\mu(A \cap T_t B) \to \mu(A)\mu(B)$ through direct calculation.

Theorem 3.2 (Non-Universality of Rigid Flows)

Flows with rigid factors are not universal.

Construction Highlights:

• Uses rigid sequences $t(i)$ satisfying $\mu(A \cap T_{t(i)} A) \to \mu(A)$
• Extends to $\mu(A \triangle T_{m(i)t(i)} A) \to 0$ (for slowly growing $m(i)$)
• Constructs a Cantor-type measure whose support lies on a nested interval system
• At level $n$, defines $2^n$ intervals $I_0, I_1, \ldots$, each with measure $2^{-n}$
• Interval selection ensures $p(i)t(i) \in m(i)t(i)I_0$ and $\mu(A \triangle T_t A) < 0.001\mu(A)$

Result: For $f = \chi_A$, the average $\int_\mathbb{R} f(T_{m(i)t(i)r}x) d\nu(r)$ approaches $f$ rather than a constant.

Remark 3.3 (Existence of Non-Mixing Universal Flows)

There exist non-mixing flows that are universal for all continuous measures.

Construction Idea:

• Anti-mixing on fast-growing sequences $n_i$: $\mu(A \cap T_{n_i} A) \to \mu(A)/2$
• But intervals $I_k$ deviating from mixing are sufficiently small and sparse such that for all continuous measures $\nu$: $\sum_i \nu(J_i/t_k) \to 0, \quad k \to \infty$

Case Analysis

Example of Sidon Automorphisms:

• Starting from the Sidon automorphism $S$ in reference 4
• Its spectral measure is concentrated on the unit circle
• Construct special flow $T_t$ with constant return function 1
• The resulting spectral measure on $\mathbb{R}$ is singular (at low-order convolutions) but high-order convolution powers are absolutely continuous
• By Theorem 1.2, these singular measures are universal

Related Work

Main Research Trajectory

1. Kozlov-Treschev (2003) 1:
   • First introduced weighted homothetic averages
   • Proved universality of absolutely continuous measures (Theorem 1.1)
   • Pioneered the study of non-uniform averages
2. Ryzhikov (1991) 2:
   • Studied homothetic averages for multidimensional time flows
   • Explored connections between mixing properties and averaging
   • Proved $\int_\mathbb{R} \mu(A \cap T_{rt}B)h(r)dr \to \mu(A)\mu(B)$
3. Bogachev (2020) 3:
   • Systematically surveyed applications of Kozlov-Treschev averaging in ergodic theorems
   • Studied various generalizations of non-uniform averages
4. Ryzhikov (2026) 4:
   • Studied spectral properties of Sidon automorphisms
   • Provided examples of measures whose convolution powers gradually become absolutely continuous
   • Supplied key construction tools for this paper
5. Ryzhikov (2021) 5:
   • Studied mixing sets of rigid transformations
   • Proved that certain non-mixing transformations still mix most of the time
   • Provided theoretical foundation for Remark 3.3

Innovations and Advantages of This Paper

Compared to existing work, this paper:

1. First establishes convolution stability: The stability of universality under convolution root extraction is entirely new
2. Extends to singular measures: Breaks through the absolute continuity limitation of Kozlov-Treschev
3. Complete characterization of universal flows: Provides a complete picture of mixing flows, rigid flows, and intermediate cases
4. More elementary technically: Avoids deep spectral theory, making proofs more accessible

Conclusions and Discussion

Main Conclusions

1. Convolution Stability of Universality: Convolution roots of universal measures remain universal, providing a systematic method for constructing new universal measures
2. Existence of Singular Universal Measures: Through Sidon automorphisms, proves the existence of singular but universal measures
3. Structure Theory of Universal Flows:
   • Mixing $\Rightarrow$ universal (for continuous measures)
   • Rigid factor $\Rightarrow$ non-universal
   • Non-mixing universal flows exist

Limitations

1. Non-Explicit Constructions: While proving the existence of singular universal measures, concrete constructions depend on deep theory of Sidon automorphisms
2. Restriction to Continuous Measures: Theorems 3.1-3.3 only discuss the continuous measure class; the case of atomic measures is not addressed
3. Optimality Unknown: Whether the convolution power $n$ in the theorems is optimal is not discussed
4. Lack of Quantitative Estimates: Convergence rates are not quantitatively estimated

Future Directions

1. Open Problem: Do there exist universal measures all of whose convolution powers are singular? This is the core open problem posed in this paper
2. Quantitative Theory: Establish convergence rate estimates depending on the degree of singularity of the measure
3. Higher-Dimensional Generalization: Extend results to flows on $\mathbb{R}^d$
4. Application Exploration: Apply these results to concrete dynamical systems (e.g., hyperbolic flows, Hamiltonian flows)
5. Algorithmic Questions: Given a measure, how can one determine whether it is universal?

In-Depth Evaluation

Strengths

1. Theoretical Depth:
   • Reveals deep connections between measure convolution structure and ergodic properties
   • The elementary proof of the normal operator lemma demonstrates sophisticated technique
   • Elegantly combines abstract functional analysis with concrete dynamical systems
2. Completeness of Results:
   • Not only proves main theorems but also provides concrete constructions
   • Completely characterizes the landscape of universal flows through counterexamples (Theorem 3.2) and positive examples (Remark 3.3)
   • Posed open problems have clear research value
3. Innovation in Methods:
   • Elementary arguments avoiding the spectral theorem are more instructive
   • The Cantor-type measure construction technique is elegant
   • The convolution root idea may have broader applications
4. Clarity of Exposition:
   • Bilingual presentation (English and Russian) increases accessibility
   • Proof strategies are clear with rigorous logic
   • References are appropriate and historical context is clear

Weaknesses

1. Computability of Concrete Examples:
   • The construction of Sidon automorphisms is quite abstract, difficult to give explicit formulas
   • The Cantor measure construction in Theorem 3.2, while theoretically clear, is complex to compute
2. Sharpness of Results:
   • The optimality of "some $n$" in theorems is not discussed
   • Whether one can relax absolute continuity to other conditions (e.g., bounded variation) is unexplored
3. Limitations of Proof Techniques:
   • The normal operator lemma proof, while elementary, only applies to $L_2$ spaces
   • Extension from $L_2$ to $L_1$ relies on density arguments, potentially losing quantitative information
4. Unclear Application Directions:
   • As a pure theoretical work, applications to concrete problems are not discussed
   • Connections with other mathematical fields (e.g., harmonic analysis, number theory) are not fully explored

Impact

1. Contribution to the Field:
   • Significantly advances the theory of non-uniform ergodic theorems
   • Provides new perspective on the role of singular measures in dynamical systems
   • Open problems may stimulate subsequent research
2. Practical Value:
   • As a pure mathematics theoretical work, direct applications are limited
   • May have potential applications in long-time averaging, statistical physics, etc.
3. Reproducibility:
   • All proofs are complete mathematical arguments, fully verifiable
   • Theoretical constructions, though abstract, have clear logic
4. Technical Impact:
   • Elementary treatment of normal operators may be useful in other problems
   • The idea of convolution stability may inspire related research

Applicable Scenarios

1. Ergodic Theory Research: Provides tools for studying non-standard averages
2. Spectral Theory: Understands spectral properties of measures through dynamical systems perspective
3. Harmonic Analysis: Connections between convolution structure and ergodic properties may have harmonic analysis applications
4. Theoretical Physics: Non-uniform time averages may have physical significance in statistical mechanics

References

1 V.V. Kozlov, D.V. Treschev, On new forms of the ergodic theorem, J. Dynam. Control Systems, 9:3 (2003), 449-453

2 V.V. Ryzhikov, Connection between the mixing properties of a flow and the isomorphicity of the transformations that compose it, Math. Notes, 49:6 (1991), 621-627

3 V.I. Bogachev, Non-uniform Kozlov–Treschev averagings in the ergodic theorem, Russian Math. Surveys, 75:3 (2020), 393-425

4 V.V. Ryzhikov, Spectrum and joint dynamics of Poisson suspensions over rank-one transformations, Sbornik: Mathematics, 117:1 (2026)

5 V.V. Ryzhikov, Mixing Sets for Rigid Transformations, Math. Notes, 110:4 (2021), 565-570

Summary: This paper represents an important theoretical advance in ergodic theory. Through sophisticated functional analysis techniques, it establishes the convolution stability of universal measures and proves the existence of singular universal measures. While lacking direct applications as a pure theoretical work, its posed open problems and developed technical methods have important implications for subsequent research in the field. The paper's principal value lies in deepening our understanding of the relationship between measure structure and the ergodic properties of dynamical systems.