Dimensions and Dimension Spectra of Non-autonomous Iterated Function Systems

Basic Information

• Paper ID: 2508.20632
• Title: Dimensions and dimension spectra of Non-autonomous iterated function systems
• Authors: Jun Jie Miao, Tianrui Wang (East China Normal University)
• Classification: math.DS (Dynamical Systems)
• Publication Date: arXiv v2, October 25, 2025
• Paper Link: https://arxiv.org/abs/2508.20632

Abstract

This paper investigates dimension theory for non-autonomous conformal iterated function systems. The authors establish formulas for intermediate dimension spectra of non-autonomous conformal sets using upper and lower topological pressures, providing a unified framework connecting Hausdorff dimension and box-counting dimension. As corollaries, simplified forms of Hausdorff dimension, packing dimension, and box dimension are obtained. Furthermore, the paper explores Hausdorff dimension for non-autonomous infinite conformal iterated function systems containing countably many conformal maps, providing dimension formulas under specific conditions.

Research Background and Motivation

1. Core Problem

The core problem addressed in this paper is: How to compute various dimensions (Hausdorff dimension, box dimension, packing dimension, and intermediate dimension spectra) of non-autonomous conformal fractal sets.

2. Problem Significance

• Theoretical Value: Dimension theory is fundamental to fractal geometry, with Hausdorff dimension and box dimension being two basic concepts. For many non-typical fractal sets (such as non-autonomous self-affine sets and Moran sets), these two dimensions may differ, and understanding the transition between them has important theoretical significance.
• Unified Framework: The intermediate dimensions introduced by Falconer, Fraser, and Kempton in 2016 provide a continuous spectrum connecting Hausdorff dimension (θ=0) and box dimension (θ=1), but research on intermediate dimension spectra for non-autonomous systems remains insufficient.

3. Limitations of Existing Methods

• Known Results for Autonomous Systems: For classical iterated function systems (IFS), Mauldin and Urbański (1996) provided dimension formulas for self-conformal sets, where the critical value h defined by topological pressure satisfies dimₕE = dim_BE = h.
• Challenges for Non-autonomous Systems: Non-autonomous systems can have different sets of maps at each level, making the structure of attractors more complex. Existing work (such as Hua et al.'s research on Moran sets and Rempe-Gillen and Urbański's 2016 work) primarily addresses specific conditions, and these conditions cannot be mutually derived.

4. Research Motivation

• Generalize dimension theory of classical IFS to the non-autonomous setting
• Establish complete theory of intermediate dimension spectra for non-autonomous conformal sets
• Obtain dimension formulas under weaker conditions (such as condition (1.13) compared to condition (1.11))
• Extend to infinite non-autonomous systems

Core Contributions

The main contributions of this paper include:

1. Intermediate Dimension Spectra Formula (Theorem 2.1): For non-autonomous finite conformal sets E satisfying the open set condition (OSC) and condition (1.11), establish complete characterization of upper and lower intermediate dimensions:
   • dim_θ E = s_θ, dim^θ E = s^θ for θ ∈ 0,1 where s_θ and s^θ are defined through jump points of upper and lower pressure functions P(t,θ) and P̄(t,θ)
2. Simplified Formulas for Box and Packing Dimensions (Theorem 2.2): Prove that
   • dim_P E = dim_B E = s* where s* is defined by a simpler pressure function P(t), without requiring the complex cutting sets of intermediate dimensions
3. Hausdorff Dimension Formula (Theorem 2.3 and Corollary 2.4): Under the boundary measure condition L^d(∂J)=0, prove
   • dim_H E = s* and relax the condition from (1.11) to (1.13)
4. Dimension for Infinite Systems (Theorem 2.5): For non-autonomous infinite conformal sets, under condition (2.18) and technical assumptions on series convergence, prove dim_H E = s*
5. Generalization of Conditions (Example 1, Corollary 2.4, Corollary 7.3): Through concrete examples, clarify relationships between conditions (1.11), (1.12), and (1.13), and provide more general results under cone conditions

Methodology Details

Task Definition

Input: Non-autonomous (finite or infinite) conformal iterated function system Φ = {Φ_k}{k≥1}, where each Φ_k = {φ{k,i}}_{i∈I_k} is a family of conformal maps

Output: Various fractal dimensions of the attractor E:

• Hausdorff dimension dim_H E
• Box dimension dim_B E (upper and lower box dimensions)
• Packing dimension dim_P E
• Intermediate dimension spectra dim_θ E, dim^θ E for θ ∈ 0,1

Key Conditions:

• Open Set Condition (OSC): int(J_) ∩ int(J_) = ∅ for i≠j
• Contraction Rate Condition: lim_{k→∞} (log c_k)/(log M_k) = 0 (Condition 1.11) where c_k = inf_ ||Dφ_{k,j}||, M_k = sup_{u∈Σ_k} ||Dφ_u||

Core Technical Framework

1. Definition of Topological Pressure Functions

Upper and Lower Pressure Functions (for finite systems):

P̄(t) = lim sup_{k→∞} (1/k) log Σ_{u∈Σ_k} ||Dφ_u||^t
P(t) = lim inf_{k→∞} (1/k) log Σ_{u∈Σ_k} ||Dφ_u||^t

Pressure Functions for Intermediate Dimensions (θ ∈ (0,1]):

P̄(t,θ) = lim sup_{δ→0} (1/k_δ) log min_{M∈M(δ,θ)} {Σ_{u∈M} ||Dφ_u||^t}

where M(δ,θ) is the family of cutting sets satisfying specific scale constraints

Critical Value Definition:

s* = inf{t : P(t) < 0} = sup{t : P(t) > 0}
s_θ = inf{t : P̄(t,θ) < 0} = sup{t : P(t,θ) > 0}

2. Key Lemma System

Bounded Distortion Lemmas (Lemma 4.3-4.5):

• Derivative estimates for composite maps: C^{-1}||Dφ_u|| ||Dφ_v|| ≤ ||Dφ_|| ≤ C||Dφ_u|| ||Dφ_v||
• Relationship between set diameter and derivatives: C_1^{-1}||Dφ_u|| ≤ |J_u| ≤ C_1||Dφ_u||
• Scaling property of measures: C^{-d}||Dφ_u||^d L^d(A) ≤ L^d(Ψ_u(A)) ≤ C^d||Dφ_u||^d L^d(A)

Covering Estimate Lemma (Lemma 5.2): For any F⊂R^d intersecting E,

Σ_{k=k_F^-}^{k_F^+} c_k^d #D(F,k) ≤ C_2

This lemma controls the growth of the number of covering sets at different scales

Subsystem Approximation (Lemma 3.5, Proposition 3.6): For infinite systems, one can construct finite subsystems such that the pressure function remains invariant

Model Architecture (Proof Strategy)

Proof Sketch for Theorem 2.1 (Intermediate Dimension Spectra)

Lower Bound Proof (dim_θ E ≥ s_θ):

1. For α < s_θ, we have P̄(α,θ) > 0
2. For any covering F satisfying δ^{1/θ} ≤ |U| ≤ δ, construct associated cutting sets
3. Using condition (1.11): M_k^η < c_k^d, obtain

   Σ_{U∈F} Σ_{u∈A(U)} |J_u|^α ≥ C^{-α}_1 min_{M∈M(δ,θ)} {Σ_{u∈M} ||Dφ_u||^α} > C^{-α}_1
   

4. Derive Σ_{U∈F} |U|^{α-η} ≥ ε_0, therefore dim_θ E ≥ α-η

Upper Bound Proof (dim_θ E ≤ s_θ):

1. For β > γ > s_θ, we have P(γ,θ) < 0
2. Construct cutting set M_k such that Σ_{u∈M_k} ||Dφ_u||^γ < 1
3. Using condition (1.11) obtain c_k^{-β} M_k^{β-γ} < 1
4. Construct covering F_k = {U_u : u∈M_k}, prove

   Σ_{u∈M_k} |U_u|^β ≤ CC_1 Σ_{u∈M_k} ||Dφ_u||^γ < CC_1
   

Proof Sketch for Theorem 2.3 (Hausdorff Dimension)

Key Innovation: Probabilistic Measure Construction

1. For each n, define measure μ_n:

   μ_n(J_u) = ||Dφ_u||^t / Σ_{v∈Σ_n} ||Dφ_v||^t
   

2. For ball B(x,r), using covering estimates and pressure condition P(t) > 0:

   μ_n(B(x,r)) ≤ C_3 r^{t'} Σ_{k} #D(B(x,r),k) c_k^d ≤ C_2C_3 r^{t'}
   

3. Using measure compactness, {μ_n} weakly converges to measure μ
4. Apply mass distribution principle: μ(B(x,r)) ≤ C r^{t'} ⟹ dim_H E ≥ t'

Technical Innovations

1. Hierarchical Definition of Pressure Functions:
   • Introduce pressure functions P(t,θ) for intermediate dimensions, using cutting sets M(δ,θ) to finely control coverings at different scales
   • Establish monotonicity of pressure functions (Lemma 4.1) and existence of jump points
2. Weakening of Conditions:
   • From condition (1.11): lim (log c_k)/(log M_k) = 0
   • To condition (1.13): lim (log c_k - log #I_k)/(log M_k) = 0
   • Example 1 clearly demonstrates relationships between conditions (1.11), (1.12), and (1.13)
3. Subsystem Approximation Technique (Lemma 3.8):
   • For given sequence {α_n}, construct subsystem Φ' such that pressure remains invariant and

     max_{k,l∈I'_n} c_{n,k}/c_{n,l} ≤ α_n (#I_n)^{1/t_0}
     

   • This allows obtaining dimension information for infinite systems through finite subsystems
4. Application of Cone Conditions (Lemma 7.1-7.2):
   • Prove that cone conditions imply L^d(∂J) = 0
   • Use cone conditions to control the number of covering sets: #Λ ≤ M_1
5. Improved Measure Construction:
   • In Hausdorff dimension proof, avoid technical difficulties of traditional methods through normalized measures
   • Weak convergence of measures ensures good properties of the limit measure

Experimental Setup

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

Theoretical Verification Cases

Example 1 (Condition Comparison):

• E_1: Moran set, #I_k = 2^k, c_{k,j} = 1/3^{k+1}
  • Satisfies (1.11), dim_H E_1 = log2/log3
• E_2: #I_k = 2^k, c_{k,1} = 1/3^{k+1}, c_{k,j} = 1/(3^{k(k+1)}) (j≥2)
  • Does not satisfy (1.11) but satisfies (1.13), dim_H E_2 = 0
• E_3: Homogeneous Moran set, #I_k = 2, c_1=1/2, c_2=1/4, c_k = c_1...c_
  • Satisfies (1.12) but not (1.13), dim_H E_3 = 0

This example clearly demonstrates the independence and scope of the three conditions.

Experimental Results

Summary of Main Theoretical Results

Key Findings

1. Dimension Uniformity: Under conditions (1.11) and L^d(∂J)=0, the three main dimensions of non-autonomous finite conformal sets are equal:

   dim_H E = dim_P E = dim_B E = s*
   

2. Continuity of Intermediate Dimension Spectrum: By Proposition 1.1, dimension spectra dim_θ E and dim^θ E are continuous on θ∈(0,1], connecting to Hausdorff dimension at θ=0 and to box dimension at θ=1
3. Hierarchy of Conditions:
   • Strongest: Condition (1.11) ⟹ Complete intermediate dimension spectrum
   • Intermediate: Condition (1.13) ⟹ Hausdorff and box dimension lower bounds
   • Special: Cone condition + (7.45) ⟹ Hausdorff dimension (relaxing Rempe-Gillen and Urbański's conditions)
4. Tractability of Infinite Systems: Through subsystem approximation, dimension problems for infinite systems can be reduced to finite systems

Related Work

Classical IFS Theory

1. Self-similar Sets: For self-similar sets satisfying the open set condition, Hausdorff dimension and box dimension are equal, given by the zero of the pressure function
2. Self-conformal Sets (Mauldin & Urbański 1996):
   • Define critical value h such that P(h) = 0
   • Prove dim_H E = dim_B E = h
   • Extend to infinite conformal IFS
3. Self-affine Sets (Falconer 1988):
   • Introduce affine dimension d(T_1,...,T_n)
   • Under specific conditions, dim_H E = min{d, d(T_1,...,T_n)}

Non-autonomous Systems

1. Moran Sets (Moran 1946, Hua et al. 1994-2000):
   • Special case of non-autonomous similar systems
   • Under condition lim (log c_k)/(log M_k) = 0:

     dim_H E = s* = lim inf s_k
     dim_P E = dim_B E = s* = lim sup s_k
     

2. Rempe-Gillen & Urbański 2016:
   • Study non-autonomous conformal sets under condition lim (log #I_k)/k = 0
   • This paper's conditions (1.11) and (1.13) are incomparable with their condition (1.12)
3. Intermediate Dimension Theory (Falconer, Fraser, Kempton 2020):
   • Introduce general definition of intermediate dimensions
   • Banaji & Fraser 2023 study intermediate dimensions of infinitely generated attractors

Position of This Work

This paper advances the field in the following aspects:

• Unified Framework: Systematically apply intermediate dimension theory to non-autonomous conformal systems
• Condition Optimization: Relax conditions in multiple theorems compared to existing literature
• Infinite Systems: Provide systematic methods for handling non-autonomous infinite conformal systems
• Technical Innovation: Develop new techniques in subsystem approximation and probabilistic measure construction

Conclusions and Discussion

Main Conclusions

1. Complete Dimension Theory: This paper establishes complete dimension theory for non-autonomous conformal sets, including Hausdorff dimension, box dimension, packing dimension, and intermediate dimension spectra
2. Central Role of Pressure Functions: All dimensions can be characterized through critical values s* or s_θ of topological pressure functions
3. Essence of Conditions: Conditions (1.11) or (1.13) ensure "uniformity" of the system at different scales, and are key to obtaining dimension formulas
4. Bridge from Finite to Infinite: Through subsystem approximation technique, problems for infinite systems can be effectively reduced to finite systems

Limitations

1. Necessity of Conditions Not Discussed: The paper focuses on sufficient conditions, without exploring whether conditions like (1.11) and (1.13) are necessary
2. Restrictions of Cone Conditions: For Hausdorff dimension, the requirement L^d(∂J)=0 or cone conditions excludes certain fractal initial sets J
3. Technical Assumptions for Infinite Systems: Conditions (1)-(3) in Theorem 2.5 regarding series convergence are quite technical and may be difficult to verify in practical problems
4. Concrete Computation of Intermediate Dimension Spectra: Although theoretical formulas are provided, computing P(t,θ) for specific examples remains challenging
5. Role of Strong Separation Condition: The paper mentions that strong separation condition (SSC) can eliminate the requirement L^d(∂J)=0 (Remark 2), but does not elaborate

Future Directions

The paper implicitly suggests the following research directions:

1. Optimization of Conditions: Can condition (1.13) be further relaxed? Can optimal form of conditions be found?
2. Computational Methods: Develop numerical algorithms for computing pressure functions and dimensions
3. More General Map Classes: Generalize results to quasiconformal maps or more general contractions
4. Applications to Dynamical Systems: Apply theory to concrete dynamical system problems, such as Julia sets, hyperbolic sets, etc.
5. Multifractal Analysis: Study multifractal spectra of measures on non-autonomous systems

In-depth Evaluation

Strengths

1. Theoretical Depth:
   • Establishes complete theoretical framework for dimension theory of non-autonomous conformal systems
   • Rigorous proofs with careful technical handling (such as multiple applications of bounded distortion, weak convergence of measures, etc.)
   • Clear logical relationships between multiple theorems, forming a complete system
2. Innovation:
   • First systematic study of intermediate dimension spectra for non-autonomous conformal sets
   • Develops new techniques in subsystem approximation (Lemma 3.8, Proposition 3.6)
   • Relaxes conditions in multiple places compared to existing literature (such as condition (1.13) vs. (1.11))
3. Technical Contributions:
   • Hierarchical definition of pressure functions (P(t) and P(t,θ))
   • Covering estimate lemma (Lemma 5.2) provides key combinatorial control
   • Application of probabilistic measure construction method in Hausdorff dimension proof
4. Writing Quality:
   • Clear structure, progressing systematically from finite to infinite systems
   • Detailed introduction of background and related work
   • Example 1 effectively illustrates relationships between different conditions
   • Theorem statements are precise with clear assumptions
   • Definitions and notation are carefully introduced
5. Completeness:
   • Covers main types of fractal dimensions
   • Ranges from finite to infinite systems
   • Covers both general conditions and special cases (cone conditions)

Weaknesses

1. Readability Issues:
   • Notation system is quite complex (Σ^k_l, M(δ,θ), A(F), D(F,k), etc.)
   • Some proofs are lengthy and highly technical, potentially unfriendly to non-specialists
   • Lacks diagrams to intuitively illustrate the structure of non-autonomous fractals
2. Insufficient Examples:
   • Only one concrete example (Example 1), primarily used to illustrate condition relationships
   • Lacks examples showing actual computation of intermediate dimension spectra
   • No concrete instances of infinite systems provided
3. Theoretical Limitations:
   • Geometric meaning of conditions (1.11) and (1.13) is not sufficiently intuitive
   • Conditions (1)-(3) for infinite systems are overly technical
   • Does not discuss necessity or optimality of conditions
4. Missing Application Discussion:
   • No discussion of applications of these results to dynamical systems, probability theory, or other fields
   • Lacks connection to practical problems
5. Open Problems:
   • Paper does not explicitly list open problems
   • Limited discussion of future research directions

Impact Assessment

1. Theoretical Contribution:
   • High Impact: Provides systematic framework for dimension theory of non-autonomous systems
   • Fills gap in intermediate dimension theory for non-autonomous systems
   • Technical methods (particularly subsystem approximation) may inspire research on other problems
2. Practical Value:
   • Moderate: Primarily theoretical results, actual computation remains challenging
   • Provides theoretical guidance for understanding complex fractal structures
   • Potential applications in fractal image compression, dynamical system analysis, etc.
3. Reproducibility:
   • High: All results have complete mathematical proofs
   • Theorem conditions are clear and explicit
   • However, lacks algorithms and code for numerical computation
4. Follow-up Research:
   • Expected to stimulate further research on dimension theory of non-autonomous systems
   • May promote development of intermediate dimension theory
   • Provides tools for studying more complex fractal systems

Applicable Scenarios

1. Direct Application:
   • Dimension computation for non-autonomous Moran sets
   • Non-autonomous conformal fractals satisfying open set condition
   • Generalization of hierarchically constructed self-similar sets
2. Potential Applications:
   • Hyperbolic sets and Julia sets in dynamical systems
   • Random fractals and random iterated function systems
   • Multifractal analysis
   • Fractal antenna design and image compression
3. Inapplicable Scenarios:
   • Overlapping systems not satisfying open set condition
   • Sets generated by non-contractive maps
   • Systems where conditions (1.11) or (1.13) severely fail

Key References

The paper cites 42 references, with key ones including:

1. Falconer, Fraser, Kempton (2020): "Intermediate dimensions" - Original paper introducing intermediate dimensions
2. Mauldin & Urbański (1996): "Dimensions and measures in infinite iterated function systems" - Classical work on infinite conformal IFS
3. Rempe-Gillen & Urbański (2016): "Non-autonomous conformal iterated function systems and Moran-set constructions" - Work directly generalized by this paper
4. Hua et al. (1994-2000): Series of works on dimension theory of Moran sets
5. Banaji & Fraser (2023): "Intermediate dimensions of infinitely generated attractors" - Intermediate dimensions of infinitely generated attractors

Overall Assessment: This is a high-quality pure mathematics theoretical paper making important contributions to dimension theory of non-autonomous iterated function systems. The paper establishes a complete theoretical framework with rigorous proofs and significant technical innovations. Main weaknesses include lack of concrete computational examples and application discussions, as well as insufficient geometric intuition for certain conditions. This work has important theoretical value for fractal geometry and dynamical systems, and is expected to have significant academic impact.