There are not many periodic orbits in bunches for iteration of complex quadratic polynomials of one variable
Feliks Przytycki · 2025
Read There are not many periodic orbits in bunches for iteration of complex quadratic polynomials of one variable by Feliks Przytycki (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-03738.txt.
Introduction
Feliks Przytycki proves bounds on how periodic orbits can bunch for complex quadratic maps. His paper is from two thousand twenty-five. Such a map is run again and again, and a periodic orbit is a loop that comes back to its start after a set count of runs. He looks at maps with a special fixed point named for Cremer.
Near that point, he shows there is at most one orbit of each large period in a tiny disc. He also bounds how many orbits of a set period can sit in one bunch that stays tight through the whole loop. The count grows no faster than a set rate. With that, a key pressure value from orbit points matches the one from preimages.
Catalog record
| CREATOR | Feliks Przytycki |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-03738
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
There are not many periodic orbits in bunches for iteration of complex quadratic polynomials of one variable Feliks Przytycki Abstract . It is proved that for every complex quadratic polynomial f f with Cremer’s fixed point z 0 z_{0} (or periodic orbit) for every δ > 0 \delta>0 , there is at most one periodic orbit of minimal period n n for all n n large enough, entirely in the disc (ball) B ( z 0 , exp − δ n ) B(z_{0},\exp-\delta n) (and at most p p for a Cremer orbit of period p p ; in fact at most one, by a renormalization remark by C.L.Petersen). Next, it is proved that the number of periodic orbits of period n n in a bunch 𝒫 n \mathcal{P}_{n} , that is for all x , y ∈ 𝒫 n x,y\in\mathcal{P}_{n} , | f j ( x ) − f j ( y ) | ≤ exp − δ n |f^{j}(x)-f^{j}(y)|\leq\exp-\delta n for all j = 0 , … , n − 1 j=0,...,n-1 , does not exceed exp δ n \exp\delta n . We conclude that the geometric pressure defined with the use of periodic points coincides with the one defined with the use of preimages of an arbitrary typical point. I. Binder, N. Makarov and S. Smirnov (Duke Math. J. 2003) proved this for all polynomials but assuming all periodic orbits were hyperbolic, and asked about general situations. We prove here a positive answer for all quadratic polynomials. 2000 Mathematics Subject Classification Primary: 37F20; Secondary: 37F10 1. Introduction In this paper we consider holomorphic maps f:U\to\overline{\mathbb{C}} , where U\subset\overline{\mathbb{C}} is an open domain in the Riemann sphere \overline{\mathbb{C}} , where \mathbb{C} is the complex plane. Usually f will be a rational function on U=\overline{\mathbb{C}} or just a quadratic polynomial f(z)=f_{c}(z)=z^{2}+c for a complex number c . The following notions of geometric pressure for a rational function f:\overline{\mathbb{C}}\to\overline{\mathbb{C}} are of interest, see e.g. [ PRS2 ] and [ P-ICM18 ] . Let us start with just variational pressure for continuous f:X\to X for a compact set X and a real-valued continuous function \phi:X\to\mathbb{R} (called a potential ) function. Definition 1.1 (variational pressure) . (1.1) P_{\rm var}(f,\phi)=\sup_{\mu\in{\mathcal{M}}(f)}\left(h_{\mu}(f)+\int_{X}\phi\,d\mu\right), where {\mathcal{M}}(f) is the set of all f -invariant Borel probability measures on X and h_{\mu}(f) is measure theoretical entropy. For f:\overline{\mathbb{C}}\to\overline{\mathbb{C}} a rational function with and \phi=-t\log|f^{\prime}| we consider f restricted to X=J(f) being Julia set for f , and any real t . Then we call the variational pressure above the variational geometric pressure and \phi the geometric potential. Though \phi is not real continuous if f has a critical point in J(f) , the definition still makes sense, since \log|f^{\prime}| is integrable for each \mu , [ P-Lyapunov ] . Definition 1.2 (Tree pressure) . For every z\in J(f) (or just z\in\overline{\mathbb{C}} ) and t\in\mathbb{R} define (1.2) P_{\tree}(z,t)=\limsup_{n\to\infty}\frac{1}{n}\log\sum_{f^{n}(x)=z,\,x\in J(f)}|(f^{n})^{\prime}(x)|^{-t}. These notions follow adequate notions in equilibrium statistical physics, e.g. free energy for say generalizations of Ising model with a hamiltonian potential on the space of configurations over \mathbb{Z}^{d} , see e.g. [ Ruelle ] and [ Bowen ] . Another definition is also natural: Definition 1.3 (Periodic pressure) . For every t\in\mathbb{R} define (1.3) P_{\per}(t)=\limsup_{n\to\infty}\frac{1}{n}\log\sum_{f^{n}(z)=z,\,z\in J(f)}|(f^{n})^{\prime}(x)|^{-t}. It was proved e.g. in [ PRS2 ] , see also [ P-conical ] , that for every z\in\overline{\mathbb{C}} except in a zero Hausdorff dimension subset of \overline{\mathbb{C}} the tree pressure and variational geometric pressure coincide. It was proved in [ PRS2 , Theorem C] , that for t\geq 0 they coincide with the periodic pressure for every rational function f if the following Hypothesis H holds, see [ PRS2 ] and [ P-ICM18 ] . Denote the standard spherical metric on the Riemann sphere \overline{\mathbb{C}} by \rho . We assume that the diameter of \overline{\mathbb{C}} in \rho is equal to 1. For rational f denote by \Per_{n} the set of all periodic points of minimal period n in the Julia set J(f) . Hypothesis H . For every rational function f:\overline{\mathbb{C}}\to\overline{\mathbb{C}} , for every \delta>0 and all n large enough, if for a set \mathcal{P}\subset\Per_{n} , for all x,y\in\mathcal{P} and all i:0\leq i<n \rho(f^{i}(x),f^{i}(y))<\exp-\delta n , then \#\mathcal{P}\leq\exp\delta n . We prove here, in Section 5 , that Hypothesis H holds for all quadratic polynomials. Therefore we conclude with Theorem 1.4 . For every quadratic polynomial f and real all t\geq 0 geometric pressures for f acting in \mathbb{C} , variational and tree ones, are equal to the periodic pressure. A condition stronger than Hypothesis H has been formulated in [ BMS ] . Namely Hypothesis BMS . For every \epsilon>0 there exists r=r(f,\epsilon) such that if n\geq n(f,\epsilon) and x\in\Per_{n} , then \#\{y\in\Per_{n}:\rho(f^{i}(x),f^{i}(y))<r\forall 0\leq i<n\}\leq\exp\epsilon n. The authors of [ BMS ] proved it for all polynomials, but under the assumption REP saying that all periodic orbits of f in J(f) are hyperbolic repelling, that is |(f^{n})^{\prime}(x)|>1 where n is a period of x . They used external rays, as we do here, and Yoccoz’s puzzle structure (which we do not use). They asked whether the assumption REP can be skipped. See [ BMS , Lemma 7 and the comment preceding it] . As we already mentioned, we succeed here, proving the equality of pressures in Theorem 1.4 for all quadratic polynomials, without assuming REP, via proving a weaker than Hypothesis BMS, but sufficient for us, Hypothesis H. Hypothesis BMS, hence Theorem 1.4 , hold also for all generalized multimodal maps of interval, with all periodic orbits hyperbolic repelling, see [ PR , Section 5] . Let now f:U\to\overline{\mathbb{C}} be a holomorphic mapping as above. Let z_{0}\in U be a fixed point for f . We call it indifferent if |f^{\prime}(z_{0})|=1 . Then the derivative f^{\prime}(z_{0}) is either a root of unity in which case we call z_{0} parabolic, or not a root of unity when we call z_{0} irrationally indifferent. In the latter case if f is not holomorphically linearizable in a neighbourhood of z_{0} we call z_{0} Cremer. See e.g. [ Milnor-book ] for an introduction to this theory. We use the same language for any periodic orbit of period n , replacing f by f^{n} . For any x of period n (in particular for z_{0} as above) the derivative (f^{n})^{\prime}(z_{0}) is sometimes called multiplier . If z_{0} is a Cremer fixed point and f^{\prime}(z_{0})=\exp\alpha 2\pi i , where \alpha is irrational then its convergents p_{n}/q_{n} in the continued fraction algorithm satisfy (1.4) \sum_{n\geq 1}\frac{\log q_{n+1}}{q_{n}}=\infty. (So convergence of the series, Bryuno condition, implies the existence of a so-called Siegel disc around z_{0} , that is a rotation in a disc by the angle 2\pi\alpha in respective holomorphic coordinates.) If additionally (1.5) \sum_{n\geq 1}\frac{\log\log q_{n+1}}{q_{n}}<\infty then f has a sequence of periodic orbits converging to z_{0} of periods being a subsequence of (q_{n}) . see [ Perez-Marco1 , Perez-Marco2 ] . A question arises how many at most such periodic orbits converging to z_{0} in full, may happen. In this paper we prove the following Theorem 1.5 . Set f=f_{c}(z)=z^{2}+c . Let z=z_{0}\in\mathbb{C} be a fixed point for f . Then for every \delta>0 there exists r_{0}>0 such that for every integer