0% found this document useful (0 votes)
17 views10 pages

Calon 2

fungsi univelen 3

Uploaded by

Muhammad Jakfar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
17 views10 pages

Calon 2

fungsi univelen 3

Uploaded by

Muhammad Jakfar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 10

DIFFERENTIAL INEQUALITIES AND UNIVALENT FUNCTIONS

ROSIHAN M. ALI, MILUTIN OBRADOVIĆ, AND SAMINATHAN PONNUSAMY

Abstract. Let M be the class of analytic functions in the unit disk D with the
normalization f (0) = f ′ (0) − 1 = 0, and satisfying the condition
arXiv:1905.01694v1 [math.CV] 5 May 2019

 ′′  2
z z
z2 + f ′ (z) − 1 ≤ 1, z ∈ D.
f (z) f (z)
Functions in M are known to be univalent in D. In this paper, it is shown that
the harmonic mean of two functions in M are closed, that is, it belongs again
to M. This result also holds for other related classes of normalized univalent
functions. A number of new examples of functions in M are shown to be starlike
in D. However we conjecture that functions in M are not necessarily starlike, as
apparently supported by other examples.

1. Introduction
Let H denote the family of analytic functions in the open unit disk D := {z ∈ C :
|z| < 1}, and A its subclass of normalized functions f (z) = z + a2 z 2 + a3 z 3 + · · · .
Further, let S denote the subclass of A consisting of functions f univalent in D.
Denote by S ∗ and C respectively the subclasses of S consisting of starlike and convex
functions. Functions f ∈ S ∗ map D onto starlike domains with respect to the
origin, while f ∈ C whenever f (D) is a convex domain. Analytically, f ∈ S ∗ if
Re (zf ′ (z)/f (z)) > 0, while f ∈ C if Re (1 + zf ′′ (z)/f ′ (z)) > 0.
Investigations into particular subclasses of A continued to be of recent interest.
These include the class U consisting of functions f ∈ A satisfying
 2
′ z
f (z) − 1 ≤ 1, z ∈ D,
f (z)
as well as the class P of functions f ∈ A with
 ′′
z
≤ 2, z ∈ D.
f (z)
The strict inclusion P ( U ( S holds within these classes (see [2, 5, 14] for a proof).
There are several generalizations [7] of this result. For recent investigations on U
and its generalization, we refer to [11–13] and the references therein.
In this paper, the phrase f ∈ U (respectively, f ∈ P) in |z| < r means that the
defining inequality holds in |z| < r instead of the full disk |z| < 1. We also follow
2010 Mathematics Subject Classification. 30C45.
Key words and phrases. Differential inequalities, harmonic mean, subclasses of analytic univa-
lent functions, starlike functions.

1
2 R. M. Ali, M. Obradović, and S.Ponnusamy

this standard convention for other classes. In [8] and [9]), the authors discussed
the classes M and N of functions from A satisfying respectively the differential
inequality
|Mf (z)| ≤ 1, and |Nf (z)| ≤ 1, z ∈ D,
where
 ′′  2
2 z ′ z
Mf (z) = z + f (z) −1
f (z) f (z)
and
 ′′′  2
3 z ′ z
Nf (z) = −z + f (z) − 1.
f (z) f (z)
These classes are also closely related to the class U in the sense of the strict inclusions
N ( M ( P ( U.
A slightly general version of this result is given in [1].
In [10], Obradović, and Ponnusamy discussed “harmonic mean” of two univalent
analytic functions. These are functions F of the form
2f (z)g(z)
(1) F (z) = ,
f (z) + g(z)
or equivalently,
   
1 1 1 1 1 1 1
(2) − = − + − ,
F (z) z 2 f (z) z g(z) z
where f, g ∈ S. In particular, the authors in [10] determined the radius of univalency
of F , and proposed the following two conjectures.
Conjecture 1. (a) The function F defined by (1) is not necessarily univalent
in D whenever f, g ∈ S such that ((f (z) + g(z))/z) 6= 0 in D.
(b) The function F defined by (1) is univalent in D whenever f, g ∈ C such that
((f (z) + g(z))/z) 6= 0 in D.
The authors in [10] showed that whenever
q√ f, g ∈ U, then the function F defined
5−1
by (1) belongs to U in the disk |z| < 2
≈ 0.78615.
While Conjecture 1 remains open, the aim of this paper is to show that Conjecture
1 (a) does not hold when the class S is replaced by U. Indeed, it does not hold true
even for the classes M, N , and P. The second objective of the paper is to consider
several examples in examining starlikeness of functions in the classes M, N , and P.
We conclude with a conjecture that functions in the class M are not necessarily
starlike in D.

2. On the harmonic mean of univalent functions


Theorem 1. Let f, g ∈ U satisfy f (z)+g(z)
z
6= 0 for z ∈ D. Then the function F given
by (1) also belongs to the class U.
Differential inequalities and univalent functions 3

Proof. From (2), it readily follows from the triangle inequality that the function F
satisfies
 2  ′
′ z 2 1 1
F (z) − 1 = −z −
F (z) F (z) z
 ′  ′
1 2 1 1 1 2 1 1
≤ −z − + −z −
2 f (z) z 2 g(z) z
 2  2
1 ′ z 1 ′ z
= f (z) − 1 + g (z) − 1 < 1.
2 f (z) 2 g(z)
Thus F ∈ U. 

Moreover, we see that Theorem 1 holds true if the class U is replaced by the class
M.
Theorem 2. Suppose f, g ∈ M satisfy f (z)+g(z)
z
6= 0 for z ∈ D. Then the function
F given by (1) also belongs to the class M.

Proof. Now  2  ′
′ z 2 1 1
f (z) − 1 = −z − .
f (z) f (z) z
Using this equality, it follows that
 ′′  ′ 
2 z 1 1
Mf (z) = z − −
f (z) f (z) z
" ′ ′ #
z 1 1
= z2 − +
f (z) f (z) z
"  ′ ′ #
1 1
= z2 z +
f (z) z
"  ′ ′ #
1 1
= z2 z −
f (z) z
 ′′  ′
3 1 1 2 1 1
= z − +z − .
f (z) z f (z) z
In view of (2), this means that
1
MF (z) = (Mf (z) + Mg (z)) ,
2
and use of the triangle inequality yields the desired result. 

Theorem 3. Let f, g ∈ N satisfy f (z)+g(z)


z
6= 0 for z ∈ D. Then the function F
given by (1) also belongs to the class N .
4 R. M. Ali, M. Obradović, and S.Ponnusamy

Proof. As in the proof of Theorem 2, we see that


"  ′ ′′  ′ #
z 1 1
Nf (z) = −z 2 z + −
f (z) f (z) z
"  2 !′′  ′ #
2 1 1 ′ z 1 1
= −z z − f (z) + −
f (z) z f (z) f (z) z
"   ′ ′′  ′ #
1 1 1 1 1 1
= −z 2 z z − + − + −
f (z) z f (z) z f (z) z
 ′′′  ′′  ′
4 1 1 3 1 1 2 1 1
= −z − − 3z − −z − .
f (z) z f (z) z f (z) z
Thus relation (2) gives
1
(Nf (z) + Ng (z)) ,
NF (z) =
2
and the proof of theorem readily follows. 

Finally, it is also readily shown that the above theorem holds true for the class P.

3. Examples and a Conjecture


It is known that functions in the class U are not necessarily starlike. There are
a number of examples displaying functions in U that are not starlike in D, see for
instance [6]. However, is M ⊂ S ∗ ? This section discusses the latter problem.
Example 1. To present a one-parameter family of functions in M that are also
starlike, consider the function f given by
z
= 1 + (1 − α)z + αz m ,
f (z)
where α ∈ (0, 1) and m ∈ N\{1} = {2, 3, . . .} are such that α(m − 1)2 = 1. Then
z/f (z) 6= 0 in D and
X∞
(k − 1)2 |bk | = (m − 1)2 α = 1,
k=2

and therefore, f ∈ M.
Next, we show that f is starlike whenever m > 1 is an odd integer. Now, a simple
calculation shows
zf ′ (z) 1 − α(m − 1)z m
= .
f (z) 1 + (1 − α)z + αz m
With z = eiθ , then
eiθ f ′ (eiθ ) A(θ) + iB(θ)
= ,
f (eiθ ) |1 + (1 − α)eiθ + αeimθ |2
Differential inequalities and univalent functions 5

where
A(θ) = 1 + (1 − α) cos θ − α(m − 2) cos(mθ)
− α(1 − α)(m − 1) cos(m − 1)θ − α2 (m − 1).
Note that A(θ) = A(−θ). As α = 1/(m − 1)2 , the expression for A(θ) reduces to
1 m(m − 2)
A(θ) = 1 − 3
− D(θ),
(m − 1) (m − 1)2
where
1 cos(m − 1)θ
D(θ) = − cos θ + cos(mθ) + .
m m−1
To show starlikeness, that is, f ∈ S ∗ , it suffices to show that A(θ) ≥ 0 for
0 ≤ θ ≤ π. First we prove the assertion for the case m = 3, while the general case
is obtained separately. Setting m = 3, A(θ) reduces to
 
7 3 1 1
A(θ) = − − cos θ + cos 3θ + cos 2θ ,
8 4 3 2
and from the identities cos 2θ = 2 cos2 θ − 1 and cos 3θ = 4 cos3 θ − 3 cos θ,
1
A(θ) = (5 + 6 cos θ − 4 cos3 θ − 3 cos2 θ)
4
1
= (1 + cos θ)2 (5 − 4 cos θ),
4
which shows that A(θ) ≥ 0. Thus, the function f3 (z) given by
z 4z
f3 (z) = 3 1 3 = ,
1 + 4z + 4z (1 + z)(4 − z + z 2 )
is starlike in D.
Next, we proceed to prove starlikeness for the general case. This requires more
computations. First,
D ′ (θ) = sin θ − sin(mθ) − sin(m − 1)θ
(2m − 1)θ θ
= sin θ − 2 sin cos
2 2 
(2m − 1)θ

θ θ
= 2 cos sin − sin
2 2 2
θ mθ (m − 1)θ
= 4 cos cos sin .
2 2 2
We need to show that A(θ) ≥ 0 for 0 ≤ θ ≤ π. It is convenient to set m = 2n + 1,
n ≥ 2 so that
θ (2n + 1)θ
D ′ (θ) = 4 cos cos sin nθ, n ≥ 2,
2 2
where D(θ) takes the form
1 1
D(θ) = − cos θ + cos(2n + 1)θ + cos(2nθ).
2n + 1 2n
6 R. M. Ali, M. Obradović, and S.Ponnusamy

Clearly, D ′ (θ) = 0 for θ = 0, π, and the critical points of D(θ) in the open interval
(0, π) are given by
(2j − 1)π

 θj =
 for j = 1, 2, . . . , n,
2n + 1
 θ′ = jπ

for j = 1, 2, . . . , n − 1,
j
n
n ≥ 2. Moreover, for each n ≥ 2,

(2n + 1)θ
cos > 0 for 0 < θ < θ1 ,


2



(2n + 1)θ
(−1)j cos > 0 for θj < θ < θj+1 and for j = 1, 2, . . . , n,


 2
(−1)j−1 sin nθ > 0 for θj−1
 ′
< θ < θj′ and for j = 1, 2, . . . , n.

In view of the above inequalities and after a careful scrutiny, it follows that

 =0 for θ = 0, θj , θj′ for j = 1, 2, . . . , n,

D (θ) > 0 for θ ∈ (0, θ1 ) ∪ (θj′ , θj+1 ) for j = 1, 2, . . . , n − 1,
<0 for θ ∈ (θj , θj′ ) for j = 1, 2, . . . , n,

where 0 < θ1 < θ1′ < θ2 < · · · < θj < θj′ < θj+1 < · · · < θn < θn′ = π. Therefore,
 
D(θ) ≤ max D(0), D (θj ) , D θj′ : j = 1, 2, . . . , n .
Since
1 1 2n 1
D(0) = −1 + + =− + ,
2n + 1 2n 2n + 1 2n
1 1 2n 1
D(π) = 1 − + = + > 0,
2n + 1 2n 2n + 1 2n
then D(0) ≤ D(π). Moreover,
1 1
D (θj ) = − cos θj + cos(2j − 1)π + cos(2n + 1 − 1)θj
2n + 1 2n
1 1
= − cos θj − − cos θj
 2n
 + 1 2n
2n + 1 1
=− cos θj − ,
2n 2n + 1
and
 1 j 1
D θj′ = − cos θj′ + cos(2n + 1) π + cos(2jπ)
 2n + 1 n 2n
1 1
=− 1− cos θj′ +
2n + 1 2n
2n 1
=− cos θj′ + .
2n + 1 2n
Differential inequalities and univalent functions 7

-1

-2
-3 -2 -1 0 1 2 3

Figure 1


We deduce that D (θj ) ≤ D(π) and D θj′ ≤ D(π) holds for each j = 1, 2, . . . , n.
Thus, D(θ) ≤ D(π) for θ ∈ [0, π]. This observation shows that
(2n + 1)(2n − 1)
 
1 2n 1
A(θ) ≥ A(π) = 1 − 3 − + = 0 for θ ∈ [0, π].
8n 4n2 2n + 1 2n
Hence Re (eiθ f ′ (eiθ )/f (eiθ )) ≥ 0, which implies that f is starlike in D. Summarizing,
for each n ≥ 1, the function fn given by
 
z 1 1
= 1 + 1 − 2 z + 2 z 2n+1 ,
fn (z) 4n 4n
belongs M, and fn is starlike in D.
Example 2. Consider

zn
 
z ζ(5) 1 X
f (z) = , φ(z) = 1 + 1 − z+ .
φ(z) ζ(3) ζ(3) n=2 (n − 1)5

We may rewrite φ as
1 z 2 1 (log(1/t))4 dt
 
ζ(5)
Z
φ(z) = 1 + 1 − z+ .
ζ(3) ζ(3) 4! 0 1 − tz
It is a simple exercise to see that φ(z) 6= 0 in D and f ∈ M. The Mathematica
software is used to display the image of the unit disk under f as shown in Figure 1.
It apparently displays that f (D) is a starlike domain.
8 R. M. Ali, M. Obradović, and S.Ponnusamy

Example 3. It is illustrative to present a general example showing that functions


in U do not necessarily belong to S ∗ . For n ≥ 3, consider the function
z
fn (z) = .
1 + ibz + (1/(n − 1))e2iβ z n
For |b| ≤ (n − 2)/(n − 1) and β a real number, then
 
z 1
Re > 1 − |b| − ≥ 0,
fn (z) n−1
and  2
z
fn′ (z) − 1 = −e2iβ z n < 1 for z ∈ D,
fn (z)
so that fn ∈ U for each n ≥ 3. On the other hand, fn is not in S ∗ when 0 < b ≤
(n − 2)/(n − 1) and 0 < β < arctan(b(n − 1)/(n − 2)). This follows on account that
 ′ 
zfn (z) [(2(n − 2)/(n − 1)) sin β − 2b cos β] sin β
Re = < 0.
fn (z) z=1 |1 + ib + (e2iβ /(n − 1))|2
Example 4. Consider the function f defined by
z
= 1 + (1 − α)z + αz m ,
f (z)
where α ∈ (0, 1) and m ≥ 3 is an odd integer such that αm(m − 1) = 2. Then
z/f (z) 6= 0 in D and
 ′′
z
= |αm(m − 1)z m−2 | < αm(m − 1) = 2,
f (z)
and therefore, f ∈ P.
As in Example 1,
 iθ ′ iθ 
e f (e ) A(θ)
Re = ,

f (e ) |1 + (1 − α)eiθ + αeimθ |2
where
A(θ) = 1 + (1 − α) cos θ − α(m − 2) cos(mθ)
− α(1 − α)(m − 1) cos(m − 1)θ − α2 (m − 1).
Substituting α = 2/(m(m − 1)) and m = 2n + 1 (n ≥ 1), the last expression for
A(θ) reduces to
2 2n − 1
(3) A(θ) = 1 − 2
+ D(θ),
n(2n + 1) n(2n + 1)
where
2(n + 1)
D(θ) = (n + 1) cos θ − cos(2n + 1)θ − cos 2nθ.
2n + 1
To prove that f is not starlike in D, it suffices to show that A(θ) < 0 for some
θ ∈ (−π, π). In the case of m = 3 (i.e. n = 1), it is a simple exercise to see that
1
A(θ) = (1 + cos θ)(11 + 4 cos θ − 12 cos2 θ),
9
Differential inequalities and univalent functions 9

n value of A(θ) n value of A(θ)


1 −0.0258011 8 −0.000243709
2 −0.0103986 9 −0.000154718
3 −0.00437311 10 −0.0000989276
4 −0.00211511 11 −0.0000628326
5 −0.00113174 12 −0.0000388937
6 −0.00064961 13 −0.000022708
7 −0.00039145 14 −0.0000116051

Table 1. Values of A(θ) for certain choices of θ

which is clearly negative for θ near π. Indeed, substituting cos θ = −8/9 or θ0 =


6π/7, it can be verified that A(θ) ≈ −55/2187 < 0, and A(θ0 ) ≈ −0.25811 < 0.
Thus, the function
z 3z
f3 (z) = 2 1 3 =
1 + 3z + 3z (1 + z)(3 − z + z 2 )
belongs to P\S ∗ .
To do away the problem for some other values of n, we proceed as follows. Set
2(2n + 1)π π
θ= and φ =
4n + 3 2(4n + 3)
so that φ = (π − θ)/2. Then
cos θ = − cos 2φ = 2 sin2 φ − 1,
cos(2n + 1)θ = − cos 2(2n + 1)φ = − sin φ,
and
cos 2nθ = cos 4nφ = sin 3φ = 3 sin φ − 4 sin3 φ.
Thus, A(θ) given by (3) can be simplified leading to
2 2(2n − 1)(n + 1)
A(θ) = 1 − 2

n(2n + 1) 2n(2n + 1)
2n − 1
 
2 4n + 5 8(n + 1) 3
+ 2(n + 1) sin φ − sin φ + sin φ .
n(2n + 1) 2n + 1 2n + 1
It is seen from the computer algebra system Mathematica that A(θ) < 0 for n =
1, 2, . . . , 15. For easy reference, Table 1 lists the values of A(θ) for n = 1, 2, . . . , 14.
Thus, we conclude that the above procedure helps us to show that for each n ∈
{1, 2, . . . , 14}, the function fn given by
 
z 1 1
=1+ 1− z+ z 2n+1
fn (z) n(2n + 1) n(2n + 1)
is not starlike in D. By a minor modification in the choice of θ, one can show that
fn is not starlike for some n ≥ 15 although it is not clear whether fn is starlike for
larger values of n.
10 R. M. Ali, M. Obradović, and S.Ponnusamy

The ideas and the motivations behind the above examples lead to the following
conjecture:
Conjecture 2. The class M is not contained in S ∗ .
Acknowledgments. The first author gratefully acknowledged support from a Uni-
versiti Sains Malaysia research university grant 1001/PMATHS/8011101. The work
of the third author is supported by Mathematical Research Impact Centric Support
of DST, India (MTR/2017/000367).
References
1. A. Baricz, and S. Ponnusamy, Differential Inequalities and Bessel Functions, J. Math. Anal.
and Appl. 400(2) (2013), 558–567.
2. L.A. Aksentiev, Sufficient conditions for univalence of regular functions. (Russian) Izv. Vysš.
Učebn. Zaved. Matematika 3(4) (1958), 3–7.
3. R. Fournier and S. Ponnusamy, A class of locally univalent functions defined by a differential
inequality, Complex Var. Elliptic Equ. 52(1) (2007), 1–8.
4. B. Friedman, Two theorems on schlicht functions, Duke Math. J. 13 (1946), 171–177.
5. M. Nunokawa, M. Obradović, and S. Owa, One criterion for univalency, Proc. Amer. Math.
Soc. 106 (1989), 1035–1037.
6. M. Obradović and S. Ponnusamy, New criteria and distortion theorems for univalent func-
tions, Complex Variables Theory Appl. 44 (2001), 173–191.
7. M. Obradović and S. Ponnusamy, On certain subclasses of univalent functions and radius
properties, Rev. Roumaine Math. Pures Appl., 54(4) (2009), 317–329.
8. M. Obradović and S. Ponnusamy, A class of univalent functions defined by a differential
inequality, Kodai Math. J. 34 (2011), 169-178.
9. M. Obradović, S. Ponnusamy, On a class of univalent functions, Appl. Math. Lett. 25(10)
(2012) 1373–1378.
10. M. Obradović and S. Ponnusamy, On harmonic combination of univalent functions, Bull.
Belg. Math. Soc. (Simon Stevin) 19(3) (2012), 461–472.
11. M. Obradović, S. Ponnusamy, and K.-J. Wirths, Geometric studies on the class U(λ), Bull.
Malaysian Math. Sci. Soc. 39(3) (2016), 1259–1284.
12. S. Ponnusamy and K.-J. Wirths, Elementary considerations for classes of meromorphic uni-
valent functions, Lobachevskii J. Math. 39(5) (2018), 713–716. Preprint.
13. S. Ponnusamy and K.-J. Wirths, Coefficient problems on the class U (λ), Probl. Anal. Issues
Anal. 7(25), No. 1, (2018), 87–103.
14. S. Ozaki and M. Nunokawa, The Schwarzian derivative and univalent functions, Proc. Amer.
Math. Soc. 33 (1972), 392–394.

R. M. Ali, School of Mathematical Sciences, Universiti Sains Malaysia, 11800


USM Penang, Malaysia
E-mail address: [email protected]

M. Obradović, Department of Mathematics, Faculty of Civil Engineering, Uni-


versity of Belgrade, Bulevar Kralja Aleksandra 73, 11000 Belgrade, Serbia.
E-mail address: [email protected]

S. Ponnusamy, Department of Mathematics, Indian Institute of Technology Madras,


Chennai-600 036, India.
E-mail address: [email protected]

You might also like