Continued Fractions and Dynamical Systems: Three Observations from the FCFG Perspective

Author: Suhang Zhang, Luoyang, Henan

Abstract

Within the framework of FCFG (Fractal Continued-Fraction Geometry), this paper explores potential connections between the arithmetic recursion of continued fractions and dynamical systems. Setting aside grand systematic construction, we extract three concrete cross-cutting observations: first, the correspondence between the continued-fraction expansion of the rotation number of a circle map and the periodicity of orbits; second, the link between boundedness of continued-fraction coefficients and system stability in KAM theory; third, the relation between Lyapunov exponents and the growth rate of denominators of convergents of continued fractions. This work offers several heuristic ideas for understanding the homology between discrete arithmetic coding and continuous dynamical behaviour.

Keywords: continued fraction; dynamical system; KAM theorem; Lyapunov exponent; symbolic dynamics

1. Introduction

The FCFG framework has previously investigated correspondences between finite continued fractions and finite self-similar fractals, where arithmetic recursion of continued fractions matches geometric scaling recursion. A dynamical system is essentially an iterative mapping of states over time, which shares an intrinsic homology with recursive structures. Rather than formulating new dynamical axioms, this paper examines how arithmetic features of continued fractions (such as coefficient magnitude and periodicity) may correspond profoundly to dynamical invariants (rotation number, stability, separation rate), starting from three specific mathematical mechanisms.

2. Observation 1: Correspondence between rotation number \omega and continued-fraction periodicity

For circle maps (e.g., smooth circle diffeomorphisms), the rotation number \omega characterises the large-scale evolution of orbits. It is well known that \omega admits a continued-fraction expansion \omega=[a_1;a_2,a_3,\dots].

From the FCFG perspective, this arithmetic expansion directly classifies the dynamical behaviour of orbits:

- Rational rotation number (finite continued fraction): the system possesses exact periodic orbits, closing after finitely many iterations.

- Quadratic irrational number (periodic continued fraction): the system exhibits quasiperiodic orbits. Geometrically, this corresponds to limit fractal structures with strong self-similarity.

- Aperiodic continued fraction (e.g., transcendental numbers): orbits are dense on the torus, associated with more intricate dynamics and possibly chaos.

Insight: The periodicity of continued-fraction coefficients serves as an “arithmetic fingerprint” for orbit closure and ergodic properties in dynamical systems.

3. Observation 2: Bounded coefficients, KAM stability and geometric dimension

In KAM (Kolmogorov–Arnold–Moser) theory, the stability of a system is tightly linked to Diophantine approximation properties of the rotation number \omega.

- Bounded coefficients and stable KAM tori: If the coefficients a_k of \omega are bounded, \omega is poorly approximable by rationals and satisfies a Diophantine condition. Under such conditions, KAM invariant tori tend to persist under perturbation and the system remains stable.

- Unbounded coefficients, resonance and chaos: When a_k is unbounded, resonance becomes more likely, which may break KAM tori and drive the system toward chaos.

The FCFG framework supplies a geometric reading: bounded continued-fraction coefficients correspond to bounded scaling ratios in geometric recursion, yielding limit fractal sets of typically finite dimension with stable structure; unbounded coefficients, by contrast, may induce drastic deformation or divergent dimension of the geometric limit set.

Insight: Boundedness of continued-fraction coefficients provides an arithmetic clue for assessing dynamical stability, namely the persistence of KAM tori.

4. Observation 3: Lyapunov exponents and symbolic coding via continued fractions

The Lyapunov exponent \lambda is the central quantity quantifying the rate of orbital separation, a hallmark of chaos.

For the classical Gauss map, results from ergodic theory establish a direct relation between the Lyapunov exponent and the denominators q_n of convergents:

\lambda = \lim_{n \to \infty} \frac{1}{n} \ln q_n

Meanwhile, the coefficient sequence a_1,a_2,\dots of a continued fraction constitutes a natural symbolic-dynamics code. The complex evolution of orbits is faithfully recorded within this integer sequence.

From the FCFG viewpoint, the temporal separation rate \lambda of dynamical orbits is quantified by the growth of continued-fraction denominators; chaotic orbits in continuous space can be encoded entirely by these discrete coefficient sequences.

Insight: Continued-fraction sequences form a natural symbolic code for dynamical evolution, and the Lyapunov exponent measures the complexity of this coding.

5. Conclusion and Discussion

Working from the recursive-homology perspective of FCFG, this paper presents three cross-cutting observations linking continued fractions and dynamical systems:

1. The periodicity in the continued-fraction expansion of the rotation number shapes the behaviour of dynamical orbits;

2. Boundedness of continued-fraction coefficients is connected to the persistence or breakdown of KAM stability;

3. Lyapunov exponents correlate with the growth of continued-fraction denominators, furnishing symbolic coding for chaos.

Limitations: This is a conceptual exploratory paper. The above observations build on established classical results in dynamical systems (KAM theory, ergodic properties of the Gauss map). FCFG provides a unified interpretive language spanning arithmetic, geometry and dynamics. Rigorous geometric dimension estimates and explicit dynamical models remain open for subsequent numerical experiments and analytical verification.