Abstract
For all \(n\in {\mathbb {N}},\) we find sharp constants \(\mu _n^{(2)}\) and \(\mu _n^{(3)}\) such that \(\mu _n^{(2)}\sigma _n^{(2)}(z,f)\prec f(z)\) and \(\mu _n^{(3)}\sigma _n^{(3)}(z,f)\prec f(z)\) in the open unit disc \({\mathbb {D}}\) for all f- normalized convex univalent functions in \({\mathbb {D}}\). Here \(\sigma _n^{(\alpha )}(z,f)\) stands for nth Ces\(\grave{a}\)ro mean of order \(\alpha ,\; \alpha \ge 0,\) of \(f(z)=\sum _{k=1}^{\infty }a_k z^k\) defined by \(\sigma _n^{(\alpha )}(z,f):=\left( {\begin{array}{c}n+\alpha -1\\ n-1\end{array}}\right) ^{-1}\sum _{k=1}^n\left( {\begin{array}{c}n+\alpha -k\\ n-k\end{array}}\right) a_k z^k \) and the symbol \('\prec ' \) stands for subordination between two analytic functions. Among other things, a generalization of an earlier known result related to subordination is also presented.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let \({\mathbb {D}}=\{z\in {{\mathbb {C}}:|z|<1}\}\) denote the open unit disc in the complex plane and \({\mathscr {S}}\) the class of functions \(f(z)=z+a_2z^2+...\) which are analytic and univalent in \({\mathbb {D}}\). Denote by \({\mathscr {S}}^*\) and \({\mathscr {K}}\) the usual subclasses of \({\mathscr {S}}\) consisting of functions which map \({\mathbb {D}}\) onto starlike (w.r.t. the origin) and convex domains, respectively. Let \({\mathscr {S}}^{*}(1/2)\subset {\mathscr {S}}\) be the class of functions which are starlike of order 1/2. It is known that \({\mathscr {K}}\subset {\mathscr {S}}^{*}(1/2).\) Further, by \({\mathscr {C}}\) we denote the class of functions \(f\in {\mathscr {S}}\) for which there exist \(g\in {\mathscr {K}}\) such that \(\Re \left\{ f'(z)/g'(z)\right\} >0\), \(z\in {\mathbb {D}}\). Functions in the class \({\mathscr {C}}\) are called close-to-convex.
The convolution or Hadamard product of two power series \(f(z)=\sum _{n=0}^{\infty }a_nz^n\) and \(g(z)=\sum _{n=0}^{\infty }b_nz^n\) is denoted by \((f*g)(z)\) and is defined as
A function f is said to be subordinate to a function g (in symbols \(f(z)\prec g(z))\) in \(|z|<r ~(0<r<1)\) if g is univalent in \(|z|<r,f(0)=g(0)\) and \(f(|z|<r)\subset g(|z|<r)\).
For a given function \(f(z)=\sum _{n=1}^{\infty }a_nz^n\) and \(n\in {{\mathbb {N}}}\), \(\alpha \ge 0,\) let
and
denote, the nth partial sum, the nth Ces\(\grave{a}\)ro mean of order \(\alpha \) and the nth de la Vall\(\acute{e}\)e Poussin mean of f, respectively.
P\(\acute{o}\)lya and Schoenberg ( [7], Theorem 2, p.298) showed that the de la Vall\(\acute{e}\)e Poussin means \(V_n(z,f)\) are convex (starlike) if and only if f is convex (starlike) and also, established the following fascinating result : \(\square \)
Theorem 1.1
If \(f\in {{\mathscr {K}}}\), then \(V_n(z,f)\prec f(z)\) in \({\mathbb {D}}.\)
Robertson [8] further extended this result by proving that if f is univalent in \({\mathbb {D}}\), then converse of Theorem 1.1 is also true.
It has been a long tradition to study mapping properties of Ces\(\grave{a}\)ro means (for example, see [4, 2, 3, 6, 9, 11, 12]). In 1995, Singh and Singh [14] proved the following analogues of above theorem of P\(\acute{o}\)lya and Schoenberg for certain transformations of the nth partial sum, \(s_n(z,f),\) and the nth Ces\(\grave{a}\)ro mean of first order, \(\sigma _n^{(1)}(z,f),\) of \(f\in {\mathscr {K}}\).
Theorem 1.2
If \(f\in {{\mathscr {K}}}\), then
(i) \((1/z)\int _0^zs_n(t,f)dt\prec f(z)\) in \({\mathbb {D}}\) for every \(n\in {\mathbb {N}}\).
(ii) \( (n/(n+1))\sigma _n^{(1)}(z,f)\prec f(z)\) in \({\mathbb {D}}.\) This result is sharp for every \(n\in {\mathbb {N}}\).
\(\square \)
In the same paper, i.e., [14], Singh and Singh also proved the following result :
Theorem 1.3
For every \(f\in {\mathscr {S}}^{*}(1/2)\) and for every positive integer n, we have
\(\square \)
In the present note, we establish the analogues of Theorem 1.1 for certain transformations of the nth Ces\(\grave{a}\)ro means \(\sigma _n^{(2)}(z,f)\) and \(\sigma _n^{(3)}(z,f),\) of \(f\in {\mathscr {K}}\). We also prove an analogue of Theorem 1.3 by replacing \(\sigma _n^{(2)}(z,f)\) there with \(\sigma _n^{(3)}(z,f)\). A generalization of a result of Singh and Singh [13] related to subordination between \(V_2(z,f)\) and \(\sigma _2^{(1)}(z,f)\) of \(f\in {\mathscr {K}}\) is also presented.
2 Preliminaries
In this section, we collect following definition and results which shall be needed to prove our results in this paper.
Definition 2.1
A sequence \(\{b_n\}_1^\infty \) of complex numbers is said to be a subordinating factor sequence if, whenever \(f(z)= \sum _{n=1}^\infty a_n z^n\), \(a_1 =1\), is univalent and convex in \({\mathbb {D}}\), we have
\(\square \)
Lemma 2.2
[15] A sequence \(\{b_n\}_1^\infty \) of complex numbers is a subordinating factor sequence if and only if
\(\square \)
Lemma 2.3
( [5], p.3) If \(h(z)= nz+(n-1)z^2+ \ldots +z^n\), then
\(\square \)
Lemma 2.4
[10] Let f and g belong to \({\mathscr {S}}^{*}(1/2)\). Then for each function F analytic in \({\mathbb {D}}\) and satisfying \(\Re F(z)>0, z \in {\mathbb {D}}\), we have
\(\square \)
Lemma 2.5
[1] Suppose that \(b_0\), \(b_1\), \(b_2\) are complex numbers, \(b_2\ne 0,\) and let \(P(z)=b_0+b_1z+b_2z^2.\) Then the zeros of P(z) lie on \(\overline{{\mathbb {D}}}=\{z\in {\mathbb {C}}:|z|\le 1\}\) if, and only if
(i) \(|b_0|\le |b_2|\) and
(ii) \(|b_0 {\bar{b}}_1 -b_1 {\bar{b}}_2| \le |b_2|^2-|b_1|^2.\)
Lemma 2.6
[10] Let \(\phi \) and \(\psi \) be convex functions in \({\mathbb {D}}\) and suppose that f is subordinate to \(\phi \). Then \(f*\psi \) is subordinate to \(\phi *\psi \) in \({\mathbb {D}}.\)
\(\square \)
3 Main Results
Ces\(\grave{a}\)ro means not only play an important role in areas like approximation theory, summability, Fourier analysis etc., but also find many applications in geometric function theory. One of the striking properties of the polynomial approximations \(\sigma _n^{(\alpha )}(z,f)\) is that they converge to f in the sense of compact convergence as \(n\rightarrow \infty \). In the following two theorems we establish analogues of Theorem 1.1 for some transformations of \(\sigma _n^{(\alpha )}(z,f),\alpha = 2,3.\)
Theorem 3.1
For all elements \(f \in {\mathscr {K}}\) and for all positive integers n, we have
in \({\mathbb {D}}\). This result is sharp for every n.
\(\square \)
Proof
Let \(f(z)= z+\sum \limits _{n=2}^\infty a_n z^n\) be any member of the class \({\mathscr {K}}\). Then
Thus, in view of the Definition 2.1, the assertion (1) holds if and only if the sequence
is a subordinating factor sequence. By Lemma 2.2, this is equivalent to
where
Here
from which we get,
It compiles to
Setting \(z=e^{i\theta }, 0\le \theta <2\pi \) and making use of Lemma 2.3, we get
On substituting \(\frac{e^{i\theta }}{1-e^{i\theta }}=\left( \frac{-1}{2}+\frac{i}{2} \cot \theta /2\right) \), we get
From (2), (3) and (5), we obtain
or
We note that \({(n+1) \sin \theta -\sin (n+1) \theta }\) and \({\cot \theta /2}\) are both non-negative for \(\theta \in [0,\pi ]\) and are both negative for \(\theta \in (\pi , 2\pi ).\) Hence,
In order to prove the sharpness of our result, we consider the function \(h(z)=z/(1-z)\) which is a member of the class \({\mathscr {K}}\). We have
In view of (5), taking \(z=e^{i \theta }, 0\le \theta <2\pi \), we get
Taking \(\theta = 2 \pi /(n+1)\), for any positive real number \(\rho \), we have
because for all \(n\in {\mathbb {N}},\)
It follows that if \(\rho > n/(n+1)\), then \(\Re \rho \sigma _n^{(2)}(z,h)< -1/2\) and hence \(\rho \sigma _n^{(2)}(z,h)\) will not be subordinate to h in \({\mathbb {D}}\) as h maps \({\mathbb {D}}\) onto the right half plane \(\Re w > -1/2\).
This completes the proof.
For the sake of illustration, graphs of \(h(\partial {\mathbb {D}})\) and \((n/(n+1))\sigma _n^{(2)}(\partial {\mathbb {D}},h), n=2,3,4\) are plotted in Fig. 1, and Fig. 2 is an enlargement of the critical portion of Fig. 1. \(\square \)
Theorem 3.2
For all \(f \in {\mathscr {K}}\) and for all positive integers n, we have
in \({\mathbb {D}}\). This result is sharp for every n.
\(\square \)
Proof
For \(f(z)= z+\sum \limits _{n=2}^\infty a_n z^n\), we have
The assertion (6) will hold, if
is a subordinating factor sequence. By Lemma 2.2, it is equivalent to show that
for all z in \({\mathbb {D}},\) where
If we write
then
Thus, we obtain
where \(F_n\) is as in (3). Setting \(z=e^{i\theta },~ 0\le \theta <2\pi \) and using (4), we get
On substituting \(\frac{e^{i\theta }}{1-e^{i\theta }}=\left( -\frac{1}{2}+\frac{i}{2}\cot \theta /2\right) \) and \((\frac{e^{i\theta }}{1-e^{i\theta }})^{2}= \frac{-e^{i\theta }}{4\sin ^{2}\theta /2}\), we get
Writing the real part, we have
Writing \(\sin ^2(n+1) \theta /2=(1-\cos (n+1) \theta )/2\) and regrouping the terms, we have
i.e.,
or
which further simplifies to
All terms on the right-hand side of above expression are positive. So, we have
for all z in \({\mathbb {D}}.\)
Thus, the desired result holds.
To prove the sharpness of our result, we consider the function \(h(z)=z/(1-z)\) which is a member of the class \({\mathscr {K}}\). We have
Setting \(z=e^{i \theta }\), we get
For a positive real number \(\rho \), we have
Simplifying, we obtain
For \(\theta = \pi \), we get
It follows that if \(\rho > 2n(n+2)/((n+3)(2n+1))\), then \(\displaystyle \Re \rho \sigma _n^{(3)}(z,h)< -1/2\) and hence \(\displaystyle \rho \sigma _n^{(3)}(z,h)\) will not be subordinate to h in \(\displaystyle {\mathbb {D}}\) as h maps \(\displaystyle {\mathbb {D}}\) onto the right half plane \(\displaystyle \Re w > -1/2\).
This completes the proof.
For visual illustration, graphs of \(h(\partial {\mathbb {D}})\) and \(2n(n+2)/((2n+1)(n+3))\sigma _n^{(3)}(\partial {\mathbb {D}},h)\), \(n=2,3,4,\) are plotted in Fig. 3, and Fig. 4 is an enlargement of the critical portion of Fig. 3. \(\square \)
Ruscheweyh [11] proved that \(\displaystyle \sigma _n^{(3)}(z,z/(1-z))\in {\mathscr {K}}\subset {\mathscr {S}}^*(1/2)\). Since the class \(\displaystyle {\mathscr {S}}^*(1/2)\) is closed under convolution (see [10], Theorem 3.1), so \(\displaystyle \sigma _n^{(3)}(z,f)= \sigma _n^{(3)}(z,z/(1-z))*f\in {\mathscr {S}}^{*}(1/2)\) for every \(\displaystyle f\in {\mathscr {S}}^*(1/2).\) In the next theorem we establish that \(\sigma _n^{(2)}(z,f)\) in Theorem 1.3 can be replaced with \(\sigma _n^{(3)}(z,f)\).
Theorem 3.3
Let \(\displaystyle f(z)=z+ \sum _{n=2}^\infty a_n z^n\) be a member of the class \(\displaystyle {\mathscr {S}}^{*}(1/2)\). Then for every positive integer n and each \(\displaystyle z \in {\mathbb {D}}\), we have
\(\square \)
Proof
Consider the function
Then
Also, the function \(F_{n}\) defined above is regular (in fact, an entire function) in \({\mathbb {D}}\) and can be written in the form
In view of (10) and (11), it is clear that in \({{\mathbb {D}}}\), we have
Taking \(\displaystyle f(z)=\sigma ^{(3)}_{n}(z,f), \ g(z)=z/(1-z)\) and \( F(z)=F_{n}(z)\) in Lemma 2.4, we immediately get
as \(\Re F_n(z) >0\) in \({\mathbb {D}}\).
This completes the proof. \(\square \)
It is known (see [6]) that for \(\alpha \ge 1\) and \(n\in {\mathbb {N}}\), \(\sigma ^{(\alpha )}_{n}(z, z/(1-z))\in {\mathscr {C}}\). Then, using the fact that the class \({\mathscr {C}}\) is closed under convolution with convex functions (see [10], Theorem 2.2), we immediately get that for \(\alpha \ge 1\) and \(n\in {\mathbb {N}}\), \(\sigma ^{(\alpha )}_{n}(z,f)\in {\mathscr {C}}\) for all \(f\in {\mathscr {K}}.\) Singh and Singh [13] proved that if \(f\in {\mathscr {K}}\), then \(z/2\prec V_2(z,f)\prec \sigma _2^{(1)}(z,f)\) in \({\mathbb {D}}\). The theorem below generalizes this result of Singh and Singh [13] in the sense that the superordinate function \(\sigma _2^{(1)}(z,f)\) can be replaced with \(\sigma _2^{(\alpha )}(z,f)\), where \(\alpha ,\) \(\alpha \ge 1,\) is any real number.
Theorem 3.4
If \(f \in {\mathscr {K}}\), then for all real numbers \(\alpha \), \(\alpha \ge 1\),
in \( {\mathbb {D}}\).
\(\square \)
Proof
We note that for \(f(z)=z+\sum _{n=2}^{\infty }a_nz^n\),
and
For every \(f\in {\mathscr {K}}\), the relation, \( z/2\prec V_{2}(z,f)\), is well known. As \(\sigma _{2}^{(\alpha )}(z,f)\) is univalent in \({\mathbb {D}}\) for \(\alpha \ge 1\) and \(V_2(0,f)=\sigma _2^{(\alpha )}(0,f)\), we need to show only that for all \(f\in {\mathscr {K}}\),
But this is equivalent to showing that for each real \(\theta \), the polynomial
has a zero on \(\overline{{\mathbb {D}}}\). Suppose that for some \(\theta ,\) R(z) has no zero in \({\mathbb {D}}.\) Then the polynomial
has both zeros on \(\overline{{\mathbb {D}}}\). Hence by Lemma 2.5, condition (i), we must have
Writing \(a_2=\rho e^{i\phi }(\rho \le 1)\) and \(\phi +\theta =\psi \), this is equivalent to
From condition (ii) of Lemma 2.5, we must have
Again, writing \(a_2=\rho e^{i\phi }\), \(\rho \le 1,\) and \(\phi +\theta =\psi \), this is equivalent to
or,
As
therefore, if (14) holds, we must have
Obviously, minimum of left-hand side of (15) occurs at \(\psi =\pi \); so, we must have
Now, at \(\psi =\pi \), (13) gives: \(\rho \le 4(1+\alpha )/(7+\alpha ). \) But \(4(1+\alpha )/(7+\alpha )\ge 1\) for \(\alpha \ge 1\) and also for \(1\le \alpha \le 5\), \(2(1+\alpha )/(5-\alpha )\ge 1\). As \(\rho \le 1\), (16) gives a contradiction (and therefore, R(z) has a zero in \({\mathbb {D}})\) except if \(\rho = 1 ~(\alpha =1), \psi =\pi \) and \(\alpha >5.\) When \(\rho =1\) and \(\psi =\pi \), we easily verify that \(-e^{-i\phi }\) is a zero of R(z) on \(\overline{{\mathbb {D}}}\). For \(\alpha >5\), we proceed as follows. If \(I(z)=z/(1-z)\), then
and
as \(\alpha > 5.\) Thus, \(V_2({\mathbb {D}},I)\) is contained in the disc \(\{z:|z|\le 5/6\}\) and for \(\alpha >5\), \(\sigma _2^{(\alpha )}({\mathbb {D}},I)\) contains the disc \(\{z:|z|\le 5/6\}. \) Therefore, \(V_2(z,I)\prec \sigma _2^{(\alpha )}(z,I),\;\;\alpha >5.\) As, \(\sigma _2^{(\alpha )}(z,I)\) is convex for \(\alpha >5\) (infact, for \(\alpha \ge 3\), see [11]), using Lemma 2.6, we immediately conclude that \(V_2(z,f)\prec \sigma _2^{(\alpha )}(z,f)\) in \({\mathbb {D}},\) for all \(\alpha >5\) and \(f\in {\mathscr {K}}.\)
This completes the proof. \(\square \)
Data Availability Statement
Not Applicable.
References
Basg\(\ddot{o}\)ze, T., Frank, J. L., Keogh, F. R.: On Convex Univalent Functions. Canad. J. Math. 22(1): 123-127 (1970)
Bustoz, J.: Jacobi polynomial sums and univalent Cesáro means. Proc. Amer. Math. Soc. 50, 259–264 (1975)
Egerváry, E.: Abbildungseigenschaften der arithmetischen Mittel der geometrischen Reihe. Math. Z. 42, 221–230 (1937)
Fezér, L.: On new properties of arithmetical means of the partial sums of fourier series. J. Math. Phys. 13, 1–17 (1934)
Goodman, A.W.: Univalent Functions, vol. I. Mariner Publishing Co., Tampa, Florida (1983)
Lewis, J.: Applications of a convolution theorem to Jacobi polynomials. SIAM J. Math. Anal. 19, 1110–1120 (1979)
P\(\acute{o}\)lya, G., Schoenberg, I. J.: Remarks on de la Vall\(\acute{e}\)e Poussin means and convex conformal maps of the circle, Pacific J. Math., 8, 295–334 (1958)
Robertson, M.S.: Applications of subordination principle to univalent functions. Pacific J. Math. 11(1), 315–324 (1961)
Robertson, M.S.: Power series with multiply monotonic coefficients. Michigan Math. J. 16, 27–31 (1969)
Ruscheweyh, S., Sheil-Small, T.: Hadamard products of Schlicht functions and the P\(\acute{o}\)lya-Schoenberg conjecture, Comment. Math. Helv., 48, 119–135 (1973)
Ruscheweyh, S.: Geometric properties of the Cesáro means. Res. Math. 22, 739–748 (1992)
Ruscheweyh, S. and Salinas, L.: Subordination by Cesáro means, Complex Variables Theory Appl. 21, 279–285 (1993)
Singh, S., Singh, R.: Subordination by univalent functions. Proc. Amer. Math. Soc. 82, 39–47 (1981)
Singh, R., Singh, S.: On the partial sums, Cesáro and de la Vallée Poussin means of convex and starlike functions of order 1/2. Proc. Indian Acad. Sci. Math. Sci. 105(4), 399–404 (1995)
Wilf, H.S.: Subordinating factor sequence for convex maps of the unit circle. Proc. Amer. Math. Soc. 12, 689–693 (1961)
Funding
The first author, Manju Yadav, acknowledges financial support from CSIR-UGC, Govt. of India, in the form of JRF vide Award Letter No. 211610111581.
Author information
Authors and Affiliations
Contributions
Each author contributed equally to the research and preparation of manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest with respect to the research.
Additional information
Communicated by V. Ravichandran.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Yadav, M., Gupta, S. & Singh, S. Subordination of Ces\(\grave{a}\)ro Means of Convex Functions. Bull. Malays. Math. Sci. Soc. 46, 48 (2023). https://doi.org/10.1007/s40840-022-01423-9
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s40840-022-01423-9