On the localization of the poles of the best Möbius approximations of f f
Hugo Arbeláez, Martin Chuaqui, Rodrigo Hernández, Willy Sierra · 2025
Read On the localization of the poles of the best Möbius approximations of f f by Hugo Arbeláez, Martin Chuaqui, Rodrigo Hernández, Willy Sierra (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-2510-20065.txt.
Introduction
Hugo Arbelaez and three co-authors study where the poles of the best Möbius approximations sit for locally univalent functions in the unit disk. Locating those poles matters for understanding how well such approximations can represent a function. The team establishes sharp geometric bounds for the pole function using Pommerenke linear invariant orders.
The analysis refines classical criteria for convexity and concavity and then examines special families in detail, including starlike mappings, convex functions of a given order, Janowski functions, and the Robertson class. For polygonal mappings they describe exactly which regions the poles can cover and prove precise multiplicity results. New convexity conditions based on bounds for the Schwarzian derivative follow as well.
Catalog record
| CREATOR | Hugo Arbeláez, Martin Chuaqui, Rodrigo Hernández, Willy Sierra |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 29 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-20065
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On the localization of the poles of the best Möbius approximations of f f Thanks: The first author was supported by the Universidad Nacional de Colombia. The fourth author thanks the Universidad del Cauca for providing time for this work through research project VRI ID 6235. Hugo Arbeláez, Martin Chuaqui, Rodrigo Hernández, Willy Sierra Abstract . We study the localization of the poles of the best Möbius approximations for locally univalent functions in the unit disk. Sharp geometric bounds for the pole function are established in terms of Pommerenke’s linear invariant orders, refining classical criteria for convexity and concavity. The behavior of poles is further analyzed for starlike mappings, convex functions of order α \alpha , Janowski functions, and Robertson’s class. For polygonal mappings, we describe the regions covered by the poles and obtain exact multiplicity results. We also derive new convexity conditions based on bounds of the Schwarzian derivative. 2020 Mathematics Subject Classification 30C45, 30C80, 30A10, 30J10 1. Introduction and preliminaries This paper is motivated by [ 8 ] where geometric properties of a conformal mapping of the unit disk were found to be determined by the location of the poles of the family of its best Möbius approximations (BMAs). An unexpected result was that a mapping is convex iff all the poles lie outside the disk, while it is concave iff all poles lie inside the disk [ [ 8 ] , Theorems 2.2, 2.3]. We will offer here a simple geometric proof. In our study, we investigate further connections between a conformal map and the family of such poles. In Theorems 3.1 and 3.4, the upper and lower order of a function as defined by Pommerenke are shown to give sharp estimates for the location of the poles, while stronger forms of convexity are shown to give refinements of the main results in [ 8 ] . Proposition 4.10, and Theorems 4.6 and 4.8, provide sharp estimates for the location of the poles for starlike mappings, and for the Janowski and Robertson subclasses of conformal mappings. In this same section of the paper, we give a fairly complete description of the regions covered by the poles when mapping the disk onto a general polygon or its complement. The final section is devoted to related results on convexity that bring in the Schwarzian derivative in connection with the values of the second coefficient a_{2} . 1.1. Best Möbius Approximation (BMA) The best Möbius approximation (BMA) to a locally injective analytic function f(z) at a point \zeta is the unique Möbius transformation M_{f}(z,\zeta) (in z ) that agrees with f(z) to second order at \zeta . Explicitly, (1) M_{f}(z,\zeta)=f(\zeta)+\frac{(z-\zeta)f^{\prime}(\zeta)}{1-\frac{1}{2}(z-\zeta)\dfrac{f^{\prime\prime}}{f^{\prime}}(\zeta)}. Thus, the pole P_{f}(\zeta) of the BMA of f at \zeta is given by: P_{f}(\zeta)=\zeta+2\frac{f^{\prime}}{f^{\prime\prime}}(\zeta), which includes the case P_{f}=\infty when f^{\prime\prime}(\zeta)=0 . In [ 8 ] , the authors gave a characterization of the localization of the poles of the BMAs as follows: Theorem A . Let f be locally injective in a domain D , and let \zeta\in D . Then the pole P_{f}(\zeta) of the BMA of f at \zeta satisfies the following: (i) The modulus |P_{f}(\zeta)| is bigger than, equal to, or smaller than |\zeta| according to whether Re \{1+\zeta(f^{\prime\prime}/f^{\prime})(\zeta)\} is positive, zero, or negative. (ii) The points 0,\,\zeta,\,P_{f}(\zeta) are collinear if and only if Im \{\zeta(f^{\prime\prime}/f^{\prime})(\zeta)\}=0 . (iii) P_{f}(\zeta)=-\zeta if and only if 1+\zeta(f^{\prime\prime}/f^{\prime})(\zeta)=0 . Moreover, an analytic proof was given that f is convex if and only if the poles are outside \mathbb{D} , while f is concave if and only if the poles lie inside \mathbb{D} . We have found the following geometric proof more intuitive. Let \mathbb{D}_{r} denote the subdisk |\zeta|<r , C_{r} be the circle |\zeta|=r<1 and z\in C_{r} be fixed. Because of the order of contact at z , the circles M_{f}(z,C_{r}) and f(C_{r}) are tangent at f(z) and have the same curvature there, with a sign equal to that of Re \{1+z(f^{\prime\prime}/f^{\prime})(z)\} . Since conformal mappings preserve orientation, we see that the set M_{f}(z,\mathbb{D}_{r}) will correspond to the bounded disk in the complement of M_{f}(z,C_{r}) when the curvature is positive, while it will be the unbounded disk in that complement when the curvature is negative. This means that M_{f}(z,\zeta) has no poles in \overline{\mathbb{D}_{r}} for positive curvature, and otherwise for negative curvature. This is depicted in the figure below. In the intermediate stages, the set M_{f}(z,C_{r}) must be a straight line. M_{f}(z,\mathbb{D}_{r}) f(z) f(z) M_{f}(z,\mathbb{D}_{r}) In the development of this article, the operator A_{f} introduced by Pommerenke in [ 18 ] will play a fundamental role. If f is a locally univalent analytic mapping defined in \mathbb{D} , we define A_{f}(z)=\frac{(1-|z|^{2})}{2}\frac{f^{\prime\prime}}{f^{\prime}}(z)-\overline{z}, for all z\in\mathbb{D} . From this operator, Pommerenke defined the (upper) order [ 18 ] and the lower order [ 11 ] by \alpha_{f}=\sup_{z\in\mathbb{D}}|A_{f}(z)|\qquad\text{and}\qquad\mu_{f}=\inf_{z\in\mathbb{D}}|A_{f}(z)|. One of the most important properties is that \alpha_{f}=1 if and only if f is convex, and \mu_{f}=1 if and only if f is concave (see [ 18 ] and [ 11 ] , respectively). 2. Properties of P_{f} function A straightforward computation shows the following result. Proposition 2.1 . Let f be a locally univalent analytic mapping defined in \mathbb{D} , then (i) P_{T\circ f}=P_{f} , where T(z)=az+b , a,b\in\mathbb{C} , a\neq 0 . (ii) If \sigma_{a}(z)=(z+a)/(1+\overline{a}z) , a\in\mathbb{D} , then (2) P_{f\circ\sigma_{a}}=\sigma_{-a}\circ P_{f}\circ\sigma_{a}. (iii) Let f be a locally univalent analytic mapping defined in \mathbb{D} and r\in(0,1) . If f_{r}(z):=f(rz) , then (3) P_{f_{r}}(z)=\frac{1}{r}\cdot P_{f}(rz), for all z\in\mathbb{D} . Remark 2.2 . If f is a concave function, it follows from Theorem A and ( 3 ), that |P_{f_{r}}(z)|\leq\frac{1}{r}|rz|\leq 1. So, for r\in(0,1) , f_{r} is a concave function. This tells us that concavity, in some sense, is a hereditary property. In the following lemma, we show that the operators A_{f} and P_{f} are closely related; therefore, A_{f} gives us information about the localizations of the poles of the BMA of f . Lemma 2.3 . Let f be a locally univalent analytic mapping defined in \mathbb{D} , then for any z\in\mathbb{D} P_{f}(z)=\frac{w+z}{1+\overline{z}w},\quad\quad w=\frac{1}{A_{f}(z)}. Proof. Let F=f\circ\sigma_{a} , where \sigma_{a}(z)=(z+a)/(1+\overline{a}z) , a\in\mathbb{D} , is an automorphism of the unit disk. From Proposition 2.1 P_{F}=\sigma_{-a}\circ P_{f}\circ\sigma_{a}. On the other hand P_{F}(z)=z+\displaystyle\frac{2(1+\overline{a}z)^{2}}{\sigma_{a}^{\prime}(z)(f^{\prime\prime}/f^{\prime})(\sigma_{a}(z))-2\overline{a}(1+\overline{a}z)}. Therefore, (4) \displaystyle w=\frac{1}{A_{f}(a)}=\frac{2}{(1-|a|^{2})(f^{\prime\prime}/f^{\prime})(a)-2\overline{a}}=P_{F}(0)=\frac{P_{f}(a)-a}{1-\overline{a}P_{f}(a)}. ∎ Remark 2.4 . An immediate consequence of the previous lemma is |A_{f}(z)|>1 if and only if |P_{f}(z)|<1 , for all z\in\mathbb{D} . Thus, f is a convex mapping if and only if A_{f}(z) lies inside \mathbb{D} , which is equivalent to w lies outside \mathbb{D} and therefore also P_{f} . Analogously, f is a concave mapping if and only if A_{f} lies outside \overline{\mathbb{D}} and therefore P_{f} lies in \mathbb{D} . 3. Location of P_{f} in terms of \mu_{f} and \alpha_{f} Theorem 3.1 . Let f be a locally univalent analytic mapping defined in \mathbb{D} . If \mu_{f}>0 , then (5) |P_{f}(z)|\leq\frac{1+\mu_{f}|z|}{|z|+\mu_{f}}, for all z\in\mathbb{D} . The inequality ( 5 ) is sharp. Proof. Let w=A_{f}^{-1}(z) , for z\in\mathbb{D} . We have, from Lemma 2.3 , that \displaystyle|P_{f}(z)|^{2} \displaystyle=\left|\frac{w+z}{1+\overline{z}w}\right|^{2}=1-\frac{(1-|w|^{2})(1-|z|^{2})}{|1+\overline{z}w|^{2}} \displaystyle\leq 1-\frac{\left(1-\dfrac{1}{\mu_{f}^{2}}\right)(1-|z|^{2})}{\left(1+|z|\dfrac{1}{\mu_{f}}\right)^{2}}=\frac{(1+\mu_{f}|z|)^{2}}{(|z|+\mu_{f})^{2}}. The example below shows the last statement. ∎ We observe that Theorem 3.1 also implies that if f is a concave function, then |P_{f}(z)|\leq 1 for all z\in\mathbb{D} . Example 3.2 . Let a\in\mathbb{R} , a\neq