.

CHAPTER 4

SOME APPLICATIONS OF
HYPERGEOMETRIC FUNCTIONS
AND FRACTIONAL CALCULUS
OPERATOR TO UNIVALENT AND
MULTIVALENT FUNCTIONS
 CHAPTER 4
Some Applications of Hypergeometric
Functions and Fractional Calculus
Operator to Univalent and
Multivalent Functions

4.1 Introduction

We breakup this Chapter into two main Sections. The first Section dealing
with the study of hypergeometric functions and Fox-Wright generalized hy-
pergeometric functions applied to both univalent and multivalent functions.

H. M. Srivastava [24], H. M. Srivastava and S. Owa [23], S. Ponnusamy [17],

J. M. Jahangiri [11] have put on remarkable contributions in the study of
hypergeometric functions and their applications to univalent starlike, con-
vex and other functions. The hypergeometric integral is any integral which

has hypergeometric function in its integral. Because of the complicated look

of hypergeometric functions, the study was overlooked. The functions in
these integrals range from simple Bessel and Whittaker functions upto gen-
eralized hypergeometric functions of several variables with many parameters.

There are many types of hypergeometric functions in literature, Gauss series

2 F1 (a, b; c; z) with one variable and three parameters. All of these quantities
may be any numbers, real or complex. This may be further extended to give
rise to the multiple hypergeometric functions of many variables. The term

“hypergeometric” was first used by Wallis in his work Arithmetica Infinito-

155
rum published in the year 1655, when he was a Professor at the University of
Oxford.
Here we obtain coefficient estimates, distortion and growth bounds, clo-

sure results, neighbourhoods and integral operators for various subclasses,

we also obtain some properties of subclass of holomorphic function for some
convolution operator on Hilbert space by using the Fox-Wright generalized
hypergeometric functions.

The Section 2 is interwoven with study of fractional calculus operator in-

cluding both fractional derivative and fractional integral. The eminent math-
ematicians like P. K. Banerji, L. Debnath and G. M. Shenan [5], Y. C. Kim
et. al. [14], H. M. Srivastava and S. Owa [23] have contributed a lot of re-

sults to study of univalent and multivalent functions using fractional calculus

technique.
Fractional calculus is the field of mathematical analysis which deals with
the investigation and application of integrals and derivatives of arbitrary or-

der. The fractional calculus is an old subject and yet novel topic, it is an old
topic starting from some speculations of G. W. Leibniz (1695) and L. Euler
(1730), it has been developed upto nowadays. N number of mathematicians
have provided important contributions up to middle of our century, includes

P. S. Laplace (1812), J. B. J. Fourier (1822), N. H. Abel (1823), J. Liouville

(1832), B. Riemann (1847), D. V. Widder (1941), M. Riesz (1949). In recent
years considerable interest in fractional calculus has been stimulated by the
applications that this calculus finds in numerical analysis and different areas

of physics and engineering.

156
 Our motivation stems from the results obtained earlier by these above
authors.
Firstly, we have tried to obtain the application of fractional calculus for

univalent functions with negative coefficients. Here we have generalized the

results obtained by R. Aghalary and S. R. Kulkarni [1], by introducing the
generalization, we have obtained the results which are routine, we have also
extended the study of fractional calculus in particular fractional derivative to

these subclasses of n-uniformly multivalent functions. We conclude this sec-

tion by obtaining the varieties of the results having geometrical importance.

SECTION 1

4.2 Some Properties on Hypergeometric Functions

Let A(p) denote the class of all holomorphic functions f (z) in the unit
disk U = {z : |z| < 1} defined by


∞
p
f (z) = z + ak z k , (ak ≥ 0, p ∈ IN) (4.1)
k=p+1

and let


∞
p
g(z) = z + bk z k , (bk ≥ 0)
k=p+1

be fixed multivalent holomorphic function in U. Define the class MNR(r, s, δ)

by
   
z(Hsr [α1 ](f ∗ g)(z))
1
M N R(r, s, δ) = f ∈ A(p) : Re < pδ, (1 < δ < 1 + , z∈U .
Hsr [α1 ](f ∗ g)(z) 2p
(4.2)

157
where


∞
(α1 )k−p · · · (αr )k−p
Hsr [α1 ](f ∗ g)(z) = z +p
ak bk z k , [8] (4.3)
k=p+1
(β1 )k−p · · · (βs )k−p (1)k−p

with Hsr [α1 ]f (z) defined in (0.16).

The generalized hypergeometric function has been studied by many math-
ematicians in univalent function theory like de Branges [8] Merkes-Scott [15],

Carlson-Shaffer [7] and Ruscheweyh - Sing [21]. Ahuja and Jahangiri [2]
studied generalized hypergeometric functions under the above operator on
univalent function on the certain classes.
Here we find the necessary and sufficient condition for function to be in

the class MNR(r, s, δ).

Theorem 4.2.1 : A function f (z) ∈ MNR(r, s, δ) if and only if


∞
(α1 )k−p · · · (αr )k−p
(k − pδ)ak bk < p(δ − 1), (4.4)
(β1 )k−p · · · (βs )k−p (1)k−p
k=p+1

1
where 1 < δ < 1 + 2p
.
Proof : Let f ∈ MNR(r, s, δ). Then by (4.2) we have
 ∞ 
(α1 )k−p ···(αr )k−p
pz p + ka b
k k z k
 k=p+1
(β1 )k−p ···(βs )k−p (1)k−p 

Re   < pδ.
∞
(α ) ···(αr ) 
zp + 1 k−p k−p
(β1 )k−p ···(βs )k−p (1)k−p
ak bk z k
k=p+1

By letting z → 1− through real values, we obtain

∞
(α1 )k−p ···(αr )k−p
p+ (β1 )k−p ···(βs )k−p (1)k−p
kak bk
k=p+1

∞ < pδ,
(α1 )k−p ···(αr )k−p
1+ a b
(β1 )k−p ···(βs )k−p (1)k−p k k
k=p+1

then


∞
(α1 )k−p · · · (αr )k−p
(k − pδ)ak bk < p(δ − 1).
(β1 )k−p · · · (βs )k−p (1)k−p
k=p+1

158
Conversely, assume that (4.4) holds true, then we have

z(Hsr [α1 ](f ∗ g)(z)) − pHsr [α1 ](f ∗ g)(z)

z(H r [α1 ](f ∗ g)(z)) − (2δ − 1)pH r [α1 ](f ∗ g)(z) ≤
s s

∞
(α1 )k−p ···(αr )k−p
(β1 )k−p ···(βs )k−p (1)k−p
(k − p)ak bk
k=p+1
∞ ≤ 1,
(α1 )k−p ···(αr )k−p
2(δ − 1)p − (β1 )k−p ···(βs )k−p (1)k−p
(k − (2δ − 1)p)ak bk
k=p+1

or equivalently f (z) ∈ MNR(r, s, δ).

Theorem 4.2.2 : Let f (z) ∈ MNR(r, s, δ). Then

p(δ − 1)
ak ≤ (α1 )k−p ···(αr )k−p
. (4.5)
(β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk

The result is sharp for function of the form

p(δ − 1)
fk (z) = z p + (α1 )k−p ···(αr )k−p
z k , k = p + 1, p + 2, · · · . (4.6)
(β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk

Proof : Since f (z) ∈ MNR(r, s, δ) by assumption, then we have by (4.4)

(α1 )k−p · · · (αr )k−p

(k − pδ)ak bk
(β1 )k−p · · · (βs )k−p (1)k−p

∞
(α1 )k−p · · · (αr )k−p
≤ (k − pδ)ak bk ≤ p(δ − 1),
k=p+1
(β1 )k−p · · · (βs )k−p (1)k−p

or
p(δ − 1)
ak ≤ (β1 )k−p · · · (βs )k−p (1)k−p .
(α1 )k−p · · · (αr )k−p (k − pδ)bk
It is clear that for

p(δ − 1)
fk (z) = z p + (β1 )k−p · · · (βs )k−p (1)k−p z k ∈ M N R(r, s, δ)
(α1 )k−p · · · (αr )k−p (k − pδ)bk

for k = p + 1, p + 2, · · ·, we have

p(δ − 1)
ak = (β1 )k−p · · · (βs )k−p (1)k−p.

(α1 )k−p · · · (αr )k−p (k − pδ)bk

159
 Now we obtain the bounds and growth theorems for f (z) ∈ MNR(r, s, δ).
Theorem 4.2.3 : Let f (z) ∈ MNR(r, s, δ). Then

p(δ − 1) p(δ − 1)
r p − α1 α2 ···αr r p+1 ≤ |f (z)| ≤ r p + α1 ···αr r p+1 ,
β1 β2 ···βs
(p(1 − δ) + 1)bp+1 β1 ···βs
(p(1 − δ) + 1)bp+1
(4.7)
for |z| = r < 1, provided bk ≥ bp+1 ≥ 1.

The result is sharp for

p(δ − 1)
f (z) = z p + α1 α2 ···αr z p+1 . (4.8)
β1 β2 ···βs
(p(1 − δ) + 1)bp+1

Proof : Since f (z) ∈ MNR(r, s, δ) by assumption, then by using (4.4) with

 
α1 α2 · · · αr (α1 )k−p (α2 )k−p · · · (αr )k−p
(p(1 − δ) + 1)bp+1 ≤ (k − pδ)bk ,
β1 β2 · · · βs (β1 )k−p · · · (βs )k−p (1)k−p

we obtain

α1 α2 · · · αr

∞
(p(1 − δ) + 1)bp+1 ak ≤ p(δ − 1),
β1 β2 · · · βs
k=p+1

or


∞
p(δ − 1)
ak ≤ α1 α2 ···αr . (4.9)
k=p+1 β1 β2 ···βs
(p(1 − δ) + 1)bp+1

For |z| = r and by using (4.9) for the function f (z) defined by (4.1), we have

∞

∞
p(δ − 1)
p k p p+1
|f (z)| ≤ r + ak r ≤ r +r ak ≤ r p + α1 α2 ···αr r p+1,
k=p+1 k=p+1 β1 β2 ···βs
(p(1 − δ) + 1)bp+1

and similarly,

p(δ − 1)
|f (z)| ≥ r p − α1 α2 ···αr r p+1 .
β1 β2 ···βs
− δ) + 1)bp+1
(p(1

Then the proof is complete.

160
Theorem 4.2.4 : Let f (z) ∈ MNR(r, s, δ). Then
p(p + 1)(δ − 1) p(p + 1)(δ − 1)
pr p−1 − α1 α2 ···αr r p ≤ |f (z)| ≤ pr p−1 + α1 α2 ···αr rp,
β1 β2 ···βs
(p(1 − δ) + 1)bp+1 β1 β2 ···βs
(p(1 − δ) + 1)bp+1
(4.10)
|z| = r < 1, provided bk ≥ bp+1 . The result is sharp for the function given by

(4.8).
Proof : The proof is simple and we avoid the details.

∞
Theorem 4.2.5 : Let Fj (z) = z p + ak,j z k (j = 1, · · · , n) be in
k=p+1

n
MNR(r, s, δ) and let γj ≥ 0 for j = 1, · · · , n and γj ≤ 1. Then
j=1

∞

n
f (z) = z p + ( γj ak,j )z k
k=p+1 j=1

also in MNR(r, s, δ).

Proof : By assumption we have Fj (z) ∈ MNR(r, s, δ), then from Theorem

4.2.1, we obtain


∞
(α1 )k−p · · · (αr )k−p
(k − δp)ak,j bk ≤ p(δ − 1)
k=p+1
(β1 )k−p · · · (βs )k−p (1)k−p

for every j = 1, · · · , n.

Hence
 n 


∞
(α1 )k−p · · · (αr )k−p

(k − pδ)bk γj ak,j
(β1 )k−p · · · (βs )k−p (1)k−p j=1
k=p+1
 ∞ 

n
(α1 )k−p · · · (αr )k−p
= γj (k − pδ)bk ak,j ≤ p(δ − 1).
j=1 k=p+1
(β1 )k−p · · · (βs )k−p (1)k−p

Therefore by Theorem 4.2.1, we conclude that f (z) ∈ MNR(r, s, δ).

Theorem 4.2.6 : Let Gp (z) = z p and
p(δ − 1)
Gk (z) = z p + (α1 )k−p ···(αr )k−p
z k , k = p + 1, p + 2, · · · .
(β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk

161
Then f (z) ∈ MNR(r, s, δ) if and only if f (z) can be expressed in the form

∞
∞
f (z) = γp z p + γk Gk (z) such that γk = 1 and γk ≥ 0.
k=p+1 k=p

∞
Proof : Let f (z) = γp z p + γk Gk (z). Then
k=p+1

 


∞
p(δ − 1)
f (z) = γp z p + γk z p + (α1 )k−p ···(αr )k−p
zk 
k=p+1 (β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk


∞
γk p(δ − 1)
= zp + (α1 )k−p ···(αr )k−p
zk ,
k=p+1 (β1 )k−p ···(βs )k−p (1)k−p (k − pδ)bk

therefore
(α1 )k−p ···(αr )k−p


∞
γk p(δ − 1) (β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk
×
(α1 )k−p ···(αr )k−p
k=p+1 (β1 )k−p ···(βs )k−p (1)k−p (k − pδ)bk p(δ − 1)

∞
= γk = 1 − γp ≤ 1.
k=p+1

So by Theorem 4.2.1, we have f (z) ∈ MNR(r, s, δ).

Conversely, let f (z) ∈ MNR(r, s, δ). Then

p(δ − 1)
ak ≤ (α1 )k−p ···(αr )k−p
(k = p + 1, p + 2, · · ·).
(β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk

Now, if we put
(α1 )k−p ···(αr )k−p
(β1 )k−p ···(βs )k−p (1)k−p
(k − pδ)bk
γk = for k = p + 1, p + 2, · · ·
p(δ − 1)

∞
and γp = 1 − γk , we obtain
k=p+1

∞
p
f (z) = γp z + γk Gk (z).
k=p+1

This completes the proof of theorem.

162

∞
Theorem 4.2.7 : Let q(z) = z p + ck z k be holomorphic in U and
k=p+1
0 ≤ ck ≤ 1. If f (z) ∈ MNR(r, s, δ), then (f ∗ q)(z) is also in MNR(r, s, δ).
Proof : By Theorem 4.2.1, we have


∞
(α1 )k−p · · · (αr )k−p
(k − pδ)ak bk < p(δ − 1).
k=p+1
(β1 )k−p · · · (βs )k−p (1)k−p

Thus from definition of convolution and since f ∈ MNR(r, s, δ) then, we get

∞
(α1 )k−p · · · (αr )k−p
(k − pδ)ak bk ck ≤ p(δ − 1),
(β1 )k−p · · · (βs )k−p (1)k−p
k=p+1

so by Theorem 4.2.1, we obtain the required result.

Definition 4.2.1 : A function f ∈ A(p) is said to be in the class MNR(ξ) (r, s, δ)

if there exists a function g(z) ∈ MNR(r, s, δ) such that

f (z)
− 1 < p − ξ, (z ∈ U, 0 ≤ ξ < 1). (4.11)
g(z)

Now we define the λ - neighbourhood of a function f ∈ A(p) by

 


∞
∞
Nλ (f ) = g ∈ A(p) : g(z) = z p + bk z k and k|ak − bk | ≤ λ .
k=p+1 k=p+1
(4.12)
Theorem 4.2.8 : If g ∈ MNR(r, s, δ) and
λ(α1 α2 · · · αr )(p(1 − δ) + 1)bp+1
ξ =p− , (4.13)
(p + 1)(α1 α2 · · · αr )(p(1 − δ) + 1)bp+1 − p(δ − 1)(β1 · · · βs )
then Nλ (g) ⊂ MNR(ξ) (r, s, δ).
Proof : Let f ∈ Nλ (g). Then from (4.12), we get


∞
k|ak − bk | < λ
k=p+1

which implies that

∞
λ
|ak − bk | ≤ .
p+1
k=p+1

163
Since g ∈ MNR(r, s, δ), then we have from (4.4)


∞
p(δ − 1)(β1 β2 · · · βs )
ak ≤ ,
k=p+1
(α1 α2 · · · αr )(p(1 − δ) + 1)bp+1

so that

∞ k
∞
(ak − bk )z |ak − bk |
f (z) k=p+1 k=p+1

g(z) − 1 < <
z p + bk z k
∞ ∞
1 − bk
k=p+1 k=p+1
λ (α1 α2 · · · αr )(p(1 − δ) + 1)bp+1
≤ · = p − ξ,
p + 1 (α1 α2 · · · αr )(p(1 − δ) + 1)bp+1 − p(δ − 1)(β1 β2 · · · βs )
provided ξ is given by (4.13).
Therefore, by Definition 4.2.1, we obtain f ∈ MNR(ξ) (r, s, δ). This com-
pletes the proof of theorem.

Theorem 4.2.9 : Let f (z) ∈ MNR(r, s, δ). Then the generalized Bernardi

integral operator defined by [6]


c + p z c−1
F (z) = t f (t)dt (c > −1, z ∈ U),
zc 0

is also in MNR(r, s, δ).

Proof : Since
 

∞
c+p k
∞
c+p
F (z) = f (z) ∗ zp + z = zp + ak z k
k=p+1
c + k k=p+1
c + k
c+p
and c+k
≤ 1, then in view of Theorem 4.2.1, we have the required result.

Next, we study some properties of subclasses of holomorphic functions for
some convolution operator on Hilbert space by using the Fox-Wright gener-

alized hypergeometric functions.

Let A1 denote the class of functions f of the form


∞
f (z) = an z n , (a1 > 0), (4.14)
n=1

164
which are holomorphic in the unit disk U = {z : |z| < 1}.
For complex parameters
 
αj
α1 , · · · , αq = 0, −1, −2, · · · ; j = 1, · · · , q
Aj

and
 
βj
β1 , · · · , βs = 0, −1, −2, · · · ; j = 1, · · · , s where q ≤ s + 1
Bj

we define the Fox-Wright generalized hypergeometric function [10] (see also

[25]),

q ψs [(α1 , A1 ), · · · , (αq , Aq ); (β1 , B1 ), · · · , (βs , Bs ); z]

= q ψs [(αj , Aj )1,q ; (βj , Bj )1,s ; z]

 q  s −1

∞   zn
= Γ(αj + Aj n) Γ(βj + Bj n) (4.15)
n=0 j=1 j=1
n!

s
q
(Aj > 0 (j = 1, · · · , q); Bj > 0 (j = 1, · · · , s); 1 + Bj − Aj ≥ 0).
j=1 j=1
If Aj = 1 (j = 1, · · · , q) and Bj = 1 (j = 1, · · · , s), we have the relationship

w q ψs [(αj , 1)1,q ; (βj , 1)1,s ; z] = q Fs (α1 , · · · , αq ; β1 , · · · , βs ; z) (4.16)

where q Fs is generalized hypergeometric function and

Γ(β1 ) · · · Γ(βs )
w= , (q ≤ s + 1; q, s ∈ IN0 = IN ∪ {0}; z ∈ U).
Γ(α1 ) · · · Γ(αq )

Now let q, s ∈ IN and suppose that α1 , · · · , αq and β1 , · · · , βs are also positive

real numbers. Then, we define a function

q φs [(αj , Aj )1,q ; (βj , Bj )1,s ; z] = wzq ψs [(αj , Aj )1,q ; (βj , Bj )1,s ; z]

and consider a linear operator [23]

165
 L[(αj , Aj )1,q ; (βj , Bj )1,s ] : A1 → A1 defined by

L[(αj , Aj )1,q ; (βj , Bj )1,s ]f (z) = q φs [(αj , Aj )1,q ; (βj , Bj )1,s ; z] ∗ f (z). (4.17)

For simplicity, we write

L[α1 ]f (z) = L[(α1 , A1 ), · · · , (αq , Aq ); (β1 , B1 ), · · · , (βs , Bs )]f (z) (4.18)

We note that special cases of this operator were investigated by Dziok and
Srivastava [9], by letting Aj = 1 (j = 1, · · · , q) and Bj = 1 (j = 1, · · · , s) in
(4.17) and includes the Noor Integral operator [16].
Let us denote by LM(q, s; A, B; IP, ξ) the class of functions f of the form


∞
f (z) = a1 z − an z n (a1 > 0; an ≥ 0; n ∈ IN \ {1}) (4.19)
n=2

and satisfying the following

 
1 IP L [α1 ]f (IP ) 1 + AIP
−ξ ≺ (−1 ≤ B < A ≤ 1) and 0 ≤ ξ < 1
1−ξ L[α1 ]f (IP ) 1 + BIP
(4.20)
for all operator IP such that IP = 00 and
IP
< 1, 00 being the null operator
on IH, where IP defined in Definition 0.1.23.

On the other hand, for real parameter E (0 < |E| < 1), we define the
following subclasses of the class LM(q, s; A, B; IP, ξ)

LME (q, s; A, B; IP, ξ) = {f : f ∈ LM(q, s; A, B; IP, ξ) and f (0) = f (E)−E = 0}

(4.21)
and

LME∗ (q, s; A, B; IP, ξ) = {f : f ∈ LM(q, s; A, B; IP, ξ) and f (0) = f  (E)−1 = 0}.

(4.22)

166
In particular, for q = s + 1 and αs+1 = As+1 = 1, we have

LM(s; A, B; IP, ξ) = LM(s + 1, s; A, B; IP, ξ), (4.23)

LME (s; A, B; IP, ξ) = LME (s + 1, s; A, B; IP, ξ) (4.24)

LME∗ (s; A, B; IP, ξ) = LME∗ (s + 1, s; A, B; IP, ξ). (4.25)

First we state a coefficient theorem.

Theorem 4.2.10 : A function f of the form (4.19) is in LM(q, s; A, B; IP, ξ)

if and only if


∞
[n(1 − B) + (A − 1) + ξ(B − A)]σn an < [(A − B) + ξ(B − A)]σ1 a1 (4.26)
n=2

where

Γ(α1 + A1 (n − 1)] · · · Γ(αq + Aq (n − 1)]

σn = , n ∈ IN. (4.27)
Γ(β1 + B1 (n − 1)] · · · Γ(βs + Bs (n − 1)](n − 1)!

Proof : Let f ∈ LM(q, s; A, B; IP, ξ), that is we have that

 
1 IP L [α1 ]f (IP ) 1 + AW (IP )
−ξ = (−1 ≤ B < A ≤ 1 and 0 ≤ ξ < 1)
1−ξ L[α1 ]f (IP ) 1 + BW (IP )

where W (00) = 00 and

W (IP )
< 1 for all operators IP = 00, thus we obtain
 
 IP L
[α ]f (I
P ) − L[α ]f (I
P ) 
 1 1 
 (A − B)(1 − ξ)L[α1 ]f (IP ) − B(IP L[α1 ]f (IP ) − L[α1 ]f (IP ))  < 1. (4.28)

By using (4.16), (4.17) and (4.18), we have

 

∞ 
 (n − 1)σn an IP
n−1 
 
 n=2  < 1.
 (A−B)(1−ξ) ∞ 
 a1 − ((A − B)(1 − ξ) − B(n − 1))σ a I
P n−1 
 w n n 
n=2

Setting IP = rII (0 < r < 1), we obtain

∞
(A − B)(1 − ξ)

∞
n−1
(n−1)σn an r < a1 − ((A−B)(1−ξ)−B(n−1))σn an r n−1 .
n=2
w n=2

167
By letting r → 1− , we get the assertion (4.26).
Conversely, it suffices to show that

IP L
[α1 ]f (IP )−L
[α1 ]f (IP )

−
(A−B)(1−ξ)L[α1 ]f (IP )−B(IP L
[α1 ]f (IP )−L[α1 ]f (IP ))
< 0.

Choosing IP = rII (0 < r < 1), we get the left hand side of the above normed

inequality equal to


∞
(A − B)(1 − ξ)

(n − 1)σn an IP n
−
a1 IP −
n=2
w


∞
((A − B)(1 − ξ) − B(n − 1))σn an IP n

n=2


∞ 
n (A − B)(1 − ξ)
≤ (n − 1)σn an r − a1 r−
n=2
w



∞
((A − B)(1 − ξ) − B(n − 1))σn an r n
n=2

∞
= [n(1 − B) + (A − 1) + ξ(B − A)]σn an r n − [(A − B) + ξ(B − A)]σ1 a1 r
n=2

∞
< [n(1 − B) + (A − 1) + ξ(B − A)]σn an − [(A − B) + ξ(B − A)]σ1 a1 ≤ 0
n=2

then we have f belongs to LM(q, s; A, B; IP, ξ).

Corollary 4.2.1 : A function f of the form (4.19) is in LME (q, s; A, B; IP, ξ)
if and only if

∞


[n(1−B)+(A−1)+ξ(B−A)]σn −[(A−B)+ξ(B−A)]σ1 E n−1 ]an ≤ [(A−B)+ξ(B−A)]σ1
n=2
(4.29)

where σn is defined by (4.27).

∞
Proof : Since f (E)
E
= 1 = a1 − an E n−1 , the result follows upon substituting
n=2

∞
a1 = 1 + an E n−1 in Theorem 4.2.10.

n=2

168
Corollary 4.2.2 : A function f of the form (4.19) is in LME∗ (q, s; A, B; IP, ξ)
if and only if

∞  
[n(1 − B) + (A − 1) + ξ(B − A)]σn n−1
− nE an ≤ 1, (4.30)
n=2
[(A − B) + ξ(B − A)]σ1

where σn is defined by (4.27). These results are entirely new.

Theorem 4.2.11 : Let a function f of the form (4.19) belong to the class
LM(q, s; A, B; IP, ξ). If the sequence ([n(1 − B) + (A − 1) + ξ(B − A)]σn )∞
n=2

is nondecreasing and positive, then

[(A − B) + ξ(B − A)]σ1 a1 2 [(A − B) + ξ(B − A)]σ1 a1 2

a1 r − r ≤
f (IP )
≤ a1 r + r .
[1 − 2B + A + ξ(B − A)]σ2 [1 − 2B + a + ξ(B − A)]σ2
(4.31)
 ∞
[n(1−B)+(A−1)+ξ(B−A)]σn
If the sequence n
is nondecreasing and positive,
n=2
then

2[(A − B) + ξ(B − A)]σ1 a1 2[(A − B) + ξ(B − A)]σ1 a1

a1 − r ≤
f
(IP )
≤ a1 + r, (4.32)
[1 − 2B + A + ξ(B − A)]σ2 [1 − 2B + A + ξ(B − A)]σ2

where σn is defined by (4.27). The result is sharp.

Proof : Since {[n(1 − B) + (A − 1) + ξ(B − A)]σn }∞
n=2 is nondecreasing and

positive, then


∞
[(A − B) + ξ(B − A)]σ1 a1
an ≤ . (4.33)
n=2
[1 − 2B + A + ξ(B − A)]σ2
 ∞
[n(1−B)+(A−1)+ξ(B−A)]σn
Also n
is nondecreasing and positive, then
n=2


∞
2[(A − B) + ξ(B − A)]σ1 a1
nan ≤ . (4.34)
n=2
[1 − 2B + A + ξ(B − A)]σ2
By using (4.19), we can write


∞

∞

∞
n n 2

f (IP )
=
a1 IP − an IP
≤ a1 r + an r = a1 r + r an r n−2
n=2 n=2 n=2


∞
[(A − B) + ξ(B − A)]σ1 a1 2
≤ a1 r + r 2 an ≤ a1 r + r
n=2
[1 − 2B + A + ξ(B − A)]σ2

169
and
[(A − B) + ξ(B − A)σ1 a1 2

f (IP )
≥ a1 r − r .
[1 − 2B + A + ξ(B − A)]σ2
Also, we have

∞

∞
 n−1

f (IP )
=
a1 − nan IP
≤ a1 + nan r n−1
n=2 n=2

∞

∞
= a1 + r nan r n−2 ≤ a1 + r nan
n=2 n=2
2[(A − B) + ξ(B − A)]σ1 a1
≤ a1 + r
[1 − 2B + A + ξ(B − A)]σ2

and
2[(A − B) + ξ(B − A)]σ1 a1

f  (IP )
≥ a1 − r.
[1 − 2B + A + ξ(B − A)]σ2
For sharpness we can take the extremal function f2 of the form

[(A − B) + ξ(B − A)]σ1 a1 2

f2 (z) = a1 z − z .
(4.35)
[1 − 2B + A + ξ(B − A)]σ2

Corollary 4.2.3 : Let a function f of the form (4.19) belong to the class
LM(s; A, B; IP, ξ). If βj ≤ αj , Bj ≤ Aj (j = 1, · · · , s), then the assertion
(4.31) and (4.32) hold true.

Proof : If q = s and βj ≤ αj , Bj ≤ Aj (j = 1, · · · , s), then

{[n(1 − B) + (A − 1) + ξ(B − A)]σn }∞

n=2

and
 ∞
[n(1 − B) + (A − 1) + ξ(B − A)]σn
n n=2

are nondecreasing. Thus, by Theorem 4.2.11, we have Corollary 4.2.3. This

result is entirely new.

170
Theorem 4.2.12 : Let a function f of the form (4.19) be in LME (q, s; A, B; IP, ξ).
If the sequence

 ∞
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1 n=2
(4.36)
is nondecreasing and positive, then

[1 − 2B + A + ξ(B − A)]σ2 r + [(A − B) + ξ(B − A)]σ1 r 2

φ(r) ≤
f (IP )
≤ ,
[1 − 2B + A + ξ(B − A)]σ2 − [(A − B) + ξ(B − A)]σ1 E
(4.37)
where

r r≤E
φ(r) = [1−2B+A+ξ(B−A)]σ2 r−[(A−B)+ξ(B−A)]σ1 r 2 .
[1−2B+A+ξ(B−A)]σ2 −[(A−B)+ξ(B−A)]σ1 E
r>E

If the sequence
 ∞
(n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1
n n=2
(4.38)
is nondecreasing and positive, then

2[(A − B) + ξ(B − A)]σ1 r

a1 − ≤
f (IP )

(1 − 2B + A + ξ(B − A)]σ2 − [(A − B) + ξ(B − A)]σ1 E

[1 − 2B + A + ξ(B − A)]σ2 r + 2[(A − B) + ξ(B − A)]σ1 r

≤ (4.39)
[1 − 2B + A + ξ(B − A)]σ2 r − [(A − B) + ξ(B − A)]σ1 r
(
IP
= r (0 < r < 1)), and σn is defined by (4.27).

The result is sharp.

Proof : Since f ∈ LME (q, s; A, B; IP, ξ), then we have

∞
[(A − B) + ξ(B − A)]σ1
an ≤
n=2
[1 − 2B + A + ξ(B − A)]σ2 − [(A − B) + ξ(B − A)]σ1 E

171
where the sequence (4.36) is positive and nondecreasing. Also, by (4.38), we
have by using (4.30)


∞
2[(A − B) + ξ(B − A)]σ1
nan ≤ .
n=2
[1 − 2B + A + ξ(B − A)]σ2 − 2[(A − B) + ξ(B − A)]σ1 E

For IP = rII (0 < r < 1), we have

 


∞

∞

f (IP )
=
a1 IP − an IP n
≤ r a1 + an r n−1
n=2 n=2
 

∞

∞
≤r 1+ an E n−1 + an r n−1
n=2 n=2
 


∞
≤ r 1 + (E + r) an
n=2
[1 − 2B + A + ξ(B − A)]σ2 r + [(A − B) + ξ(B − A)]σ1 r 2
≤
[1 − 2B + A + ξ(B − A)]σ2 − [(A − B) + ξ(B − A)]σ1 E

and
   

∞ 

∞
 

f (IP )
= a1 IP − an IP n  ≥ r a1 − an r n−1
 
n=2
  n=2

∞
= r 1+ an (E n−1 − r n−1 ) ,
n=2

then
f (IP )
≥ r if r ≤ E, while if r > E, we get the sequence {E n−1 −r n−1}∞
n=2

is negative and decreasing, therefore we have

 

∞

f (IP )
≥ r 1 + (E − r) an
n=2
[1 − 2B + A + ξ(B − A)]σ2 r − [(A − B) + ξ(B − A)]σ1 r 2
≥ .
[1 − 2B + A + ξ(B − A)]σ2 − [(A − B) + ξ(B − A)]σ1 E

By the same way we can obtain the assertion (4.39).

For sharpness we can take the extremal function f2 (z) and f (z) = z where

[1 − 2B + A + ξ(B − A)]σ2 z − [(A − B) + ξ(B − A)]σ1 z 2

f2 (z) = .

[(1 − 2B + A + ξ(B − A)]σ2 − [(A − B) + ξ(B − A)]σ1 E

172
Remark 4.2.1 : Let a function f of the form (4.19) belong to the class
LME (s; A, B; IP, ξ). Then if A1 ≤ α1 , βj ≤ αj (j = 2, · · · , s), and Bj =
Aj (j = 1, · · · , s), the assertion (4.37) holds true and if β1 ≤ α1 , the assertion

(4.39) holds true.

Next we introduce the region of univalency in particular to that of star-
likeness and convexity of LME (q, s; A, B, IP, ξ).
Theorem 4.2.13 : The radius of starlikeness for the class LME (q, s; A, B; IP, ξ)

is given by
  n−1
1
[n(1 − B) + (A − 1) + ξ(B − A)]σn
r1 = inf . (4.40)
n≥2 n[(A − B) + ξ(B − A)]σ1

The result is sharp.

Proof : We shall show that

 
 IP f  (IP ) 
 
 f (IP ) − 1 ≤ 1 , (IP = r1 II (0 < r1 < 1)).

It is easy to show that

 
 IP f (IP ) 
 
 f (IP ) − 1 < 1

if


∞

∞

∞
(n − 1)an r1n−1 ≤1+ an E n−1
− an r1n−1
n=2 n=2 n=2

or


∞
nan r1n−1 − an E n−1 ≤ 1. (4.41)
n=2

By (4.29),

∞  
[n(1 − B) + (A − 1) + ξ(B − A)]σn n−1
an −E ≤ 1.
n=2
[(A − B) + ξ(B − A)]σ1

173
Therefore, (4.41) will be true if

[n(1 − B) + (A − 1) + ξ(B − A)]σn

nr1n−1 − E n−1 ≤ − E n−1
[(A − B) + ξ(B − A)]σ1

or
  n−1
1
[n(1 − B) + (A − 1) + ξ(B − A)]σn
r1 < , n ≥ 2.
n[(A − B) + ξ(B − A)]σ1
The result is sharp for the function

[n(1 − B) + (A − 1) + ξ(B − A)]σn z − [(A − B) + ξ(B − A)]σ1 z n

f (z) = .
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1
(4.42)

Here we also have to mention that the radius of convexity for the class

LME (q, s; A, B; IP, ξ) is given by

  n−1
1
[n(1 − B) + (A − 1) + ξ(B − A)]σn
r2 = inf . (4.43)
n≥2 [(A − B) + ξ(B − A)]σ1 n2

The result is sharp for the function given by (4.42).

The proof of this theorem is similar to that of Theorem 4.2.13 and hence
omitted.

Here we introduce an integral operator due to Bernardi [6]

 z
c+1
Lc [f ] = tc−1 f (t)dt (4.44)
zc 0

where c is real number such that c > −1.

Theorem 4.2.14 : If f ∈ LM(q, s; A, B; IP, ξ), then so Lc [f ].

Proof : We can write

∞
Lc [f ] = a1 z − bn z n , bn ≥ 0,
n=2

174
 c+1
where bn = a .
c+n n
Therefore

∞
[n(1 − B) + (A − 1) + ξ(B − A)]σn bn
n=2


∞
c+1
= [n(1 − B) + (A − 1) + ξ(B − A)]σn an
n=2
c+n
≤ [n(1 − B) + (A − 1) + ξ(B − A)]σn an ≤ [(A − B) + ξ(B − A)]σ1 a1 .

By assumption f (z) ∈ LM(q, s; A, B; IP, ξ). Thus Lc [f ] ∈ LM(q, s; A, B; IP, ξ).

Theorem 4.2.15 : Let f1 (z) = z and

[n(1 − B) + (A − 1) + ξ(B − A)]σn z − [(A − B) + ξ(B − A)]σ1 z n

fn (z) = .
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1

Then f (z) ∈ LME (q, s; A, B; IP, ξ) if and only if it can be expressed in the

form

∞
∞
f (z) = λn fn (z), where λn ≥ 0 and λn = 1.
n=1 n=1

∞
∞
Proof : Let f (z) = λn fn (z) where λn ≥ 0, λn = 1. Therefore
n=1 n=1

f (z) =

∞
λn [n(1 − B) + (A − 1) + ξ(B − A)]σn z
λ1 z +
n=2
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1


∞
λn [(A − B) + ξ(B − A)]σ1 z n
− .
n=2
[((1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1

Now, by (4.29) we have

∞
λn [(A − B) + ξ(B − A)]σ1
×
n=2
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1
 
∞
[n(1 − B) + (A − 1) + ξ(B − A)]σn n−1
−E = λn = 1 − λ1 < 1.
[(A − B) + ξ(B − A)]σ1 n=2

This shows that f (z) ∈ LME (q, s; A, B; IP, ξ).

175
 Conversely, let f (z) ∈ LME (q, s; A, B; IP, ξ), then
[(A − B) + ξ(B − A)]σ1
an ≤ , n ≥ 2.
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1
Putting
[n(1 − B) + (A − 1) + ξ(B − A)]σn − [(A − B) + ξ(B − A)]σ1 E n−1
λn = , n≥2
[(A − B) + ξ(B − A)]σ1

∞
and λ1 = 1 − λn , we obtain
n=2

∞
f (z) = λn fn (z).

n=1

SECTION 2

4.3 Some Applications of Fractional Calculus to

Certain Subclasses of Univalent and Uniformly
Multivalent Holomorphic Functions
Let T (n) denote the class of functions of the form


∞
f (z) = z − ak z k (ak ≥ 0, n ∈ IN) (4.45)
k=n+1

which are holomorphic and univalent in the open disk

U = {z : z ∈ C, |z| < 1}.

In the beginning of this section we introduce the class Mη (A, B, α, λ, µ, β)

∞
consisting of all functions in the form f (z) = z − ak z k (ak ≥ 0, k ≥ 2)
k=2
satisfying the condition
µ,β,η
J0,z f (z) µ+1,β+1,η+1 1 + [B + (A − B)(1 − α)]z
(1 − λ) + λ(J0,z f (z)) ≺ ,z ∈ U
z 1 + Bz
(4.46)

176
(−1 ≤ A < B ≤ 1, 0 < B ≤ 1, 0 ≤ α < 1, λ ≥ 0, 0 ≤ β < 1, 0 ≤ µ < 1, η >
µ,β,η
max{µ, β} − 2), with J0,z f (z) is the Saigo type fractional calculus operator
defined in Definition 0.1.21.

Or equivalently,

µ,β,η
J0,z f (z) µ+1,β+1,η+1
(1 − λ) + λ(J0,z f (z)) − 1
z ≤ 1.
µ,β,η
B[(1 − λ) J0,z f (z) + λ(J0,zµ+1,β+1,η+1
f (z))] − [B + (A − B)(1 − α)]
z
(4.47)
µ,µ,η
If µ = β, we have Dzµ f (z) = J0,z f (z) (0 ≤ µ < 1) and let

Γ(β − µ + η + 2) µ,β,η β−1

J µ,β f (z) = zJ (z )f (z).
Γ(η + 2)Γ(β + 1) 0,z

By taking µ = β = ν, η = 1, we have

Γ(3) µ,µ,1 v−1

J µ,µ f (z) = zJ0,z (z )f (z). (4.48)
Γ(3)Γ(ν + 1)

So we have

z
∞
v v v−1
D f (z) = D (z f (z)) = z − ak Bk (λ)z k , (4.49)
Γ(v + 1) k=2

where
Γ(v + k)
Bk (λ) = , (see e.g. [4]).
Γ(v + 1)Γ(k)
The relation (4.49) is Ruscheweyh derivative of order v. Therefore, we get a
special case of class Mη (A, B, α, λ, v) which was studied by R. Aghalary and

S. R. Kulkarni [1].
Next we derive the coefficient inequality for the class Mη (A, B, α, λ, µ, β).
Theorem 4.3.1 : Let f ∈ T (n). Then a necessary and sufficient condition
for

177
f ∈ Mη (A, B, α, λ, µ, β) is

∞  
(1 + B)(1 − λ + λ(k − β)) 1 − λβ
ak ≤ (B−A)(1−α)+(1+B) −1
k=n+1
φk (µ, β, η) φ1 (µ, β, η)
(4.50)
where, −1 ≤ A < B ≤ 1, 0 < B ≤ 1, 0 ≤ α < 1, 1 − λβ ≥ φ1 (µ, β, η), λ ≥

0, 0 ≤ β < 1, 0 ≤ µ < 1, η > max{µ, β} − 2, and

Γ(1 + k − β)Γ(1 + k + η − µ)
φk (µ, β, η) = , k ∈ IN. (4.51)
Γ(1 + k)Γ(1 + k + η − β)

Proof : Let f ∈ Mη (A, B, α, λ, µ, β). Then in view of the fact Re(z) ≤ |z|

and by (0.14), we have


∞ 
1−λ+λ(k−β) k−β−1 1 −β λβ −β
ak z − φ1 (µ,β,η) z + φ1 (µ,β,η) z + 1
 φk (µ,β,η) 
Re   ≤ 1.
k=n+1
 1−λ+λ(k−β)
∞
z −β λβ

(B − A)(1 − α) − B φk (µ,β,η)
ak z k−β−1 + φ1 (µ,β,η)
B − φ1 (µ,β,η)
B − B
k=n+1

Choose the values of z on the real axis so that upon clearing the de-
nominator in the last expression and letting z → 1− through real values, we
have

∞
1 − λ + λ(k − β)
ak
k=n+1
φ k (µ, β, η)

∞  
1 − λ + λ(k − β) 1 − λβ
≤ (B − A)(1 − α) − B ak + (1 + B) −1 .
k=n+1
φ k (µ, β, η) φ 1 (µ, β, η)

Therefore, we have (4.50).

Conversely, if (4.50) holds true, then we obtain
 µ,β,η 
µ,β,η
J0,z f (z) J0,z f (z)
−
z
+ λz z
1

 µ,β,η  µ,β,η  

B J0,z f (z) + λz J0,z f (z) − [B + (A − B)(1 − α)]
z z

178
  ∞

1 − λ + λ(k − β)
1
≤ ak z k−β−1 − z −β +
φ k (µ, β, η) φ 1 (µ, β, η)
k=n+1
 
λβ
∞
−β
z + 1 / (B − A)(1 − α) − B
φ1 (µ, β, η)
k=n+1

1 − λ + λ(k − β) z −β
λβ
ak z k−β−1
+ B− z B − B < 1.
−β
φk (µ, β, η) φ1 (µ, β, η) φ1 (µ, β, η)

The function f (z) given by

 
1−λβ
[(B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 ]φk (µ, β, η)
f (z) = z− z k , (k = 2, 3, · · ·)
(1 + B)(1 − λ + λ(k − β))
(4.52)
is an extremal function for assertion of the Theorem 4.3.1.

Deduction 4.3.1 : Let f (z) ∈ T (n) be in the class Mη (A, B, α, λ, µ, β).
Then
 
1−λβ
[(B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 ]φk (µ, β, η)
ak ≤ (k = 2, 3, · · ·).
(1 + B)(1 − λ + λ(k − β))
(4.53)

∞
Theorem 4.3.2 : Let fi (z) = z − ak,i z k (ak,i ≥ 0, i = 1, · · · , m),
k=n+1
n, m ∈ IN be in the class Mη (A, B, α, λ, µ, β). Then the function
 m 

m

h(z) = ci fi , ci = 1, ci ≥ 0, i = 1, · · · , m (4.54)
i=1 i=1

is in the class Mη (A, B, α, λ, µ, β).

Proof : From Definition of h(z), we have
 m 


∞

h(z) = z − ci ak,i z k ,
k=n+1 i=1

and from Theorem 4.3.1

∞
(B + 1)(1 − λ + λ(k − β))

m
[ ci ak,i ]
φk (µ, β, η) i=1
k=n+1

179
  
m
1 − λβ
≤ [(B − A)(1 − α) + (B + 1) −1 ci
φ1 (µ, β, η) i=1
 
m
1 − λβ
= (B − A)(1 − α) + (B + 1) − 1 , since ci = 1.
φ1 (µ, β, η) i=1

This proves that h(z) ∈ Mη (A, B, α, λ, µ, β).

We note that Mη (A, B, α, λ, µ, β) is a convex set.
Theorem 4.3.3 : Let f (z) ∈ Mη (A, B, α, λ, µ, β). Then for |z| = r < 1
 
[(B − A)(1 − α) + (B + 1) φ11−λβ
(µ,β,η)
− 1 ]φn+1 (µ, β, η)
r− r n+1 ≤ |f (z)|
(B + 1)(1 − λ + λ((n + 1) − β))
 
1−λβ
[(B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 φn+1 (µ, β, η)
≤r+ r n+1 (4.55)
(B + 1)(1 − λ + λ((n + 1) − β))
Furthermore

µ,β,η
r 1−β − An r n+1−β ≤ J0,z f (z) ≤ r 1−β + An r n+1−β , (4.56)

where  
1−λβ
(B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
−1
An = (4.57)
(B + 1)(1 − λ + λ((n + 1) − β))
(−1 ≤ A < B ≤ 1, 0 < B ≤ 1, 0 ≤ α < 1, 1 − λβ ≥ φ1 (µ, β, η), λ ≥ 0,
0 ≤ β ≤ µ < 1 and η > µ − 2).
Proof : Let f (z) ∈ Mη (A, B, α, λ, µ, β), (z ∈ U). Then by Theorem 4.3.1,
we obtain
(B + 1)(1 + λ(n − β))


∞ ∞
(B + 1)(1 − λ + λ(k − β))
ak ≤ ak
φn+1 (µ, β, η) k=n+1 k=n+1
φ k (µ, β, η)
 
1 − λβ
≤ (B − A)(1 − α) + (B + 1) −1
φ1 (µ, β, η)
which immediately yields
  
1−λβ

∞ (B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 φn+1 (µ, β, η)
ak ≤ .
(B + 1)(1 + λ(n − β))
k=n+1

180
Now

∞

∞
k n+1
|f (z)| ≥ |z| − ak |z| ≥ |z| − |z| ak
k=n+1 k=n+1
  
(B − A)(1 − α) + (B + 1) φ11−λβ
(µ,β,η)
− 1 φn+1 (µ, β, η)
≥ |z| − |z|n+1 .
(B + 1)(1 + λ(n − β))

On the other hand, we note that

∞
n+1
|f (z)| ≤ |z| + |z| ak
k=n+1
  
1−λβ
(B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 φn+1 (µ, β, η)
≤ |z| + |z|n+1 .
(B + 1)(1 + λ(n − β))

Next,

µ,β,η 1
∞
1
1−β n+1−β
|J0,z ≥
f (z)| |z| − |z| ak
φ1 (µ, β, η) k=n+1
φk (µ, β, η)
  
1−λβ
1 (B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 |z|n+1−β
1−β
≥ |z| −
φ1 (µ, β, η) (B + 1)(1 + λ(n − β))

and
  
1 (B − A)(1 − α) + (B + 1) φ11−λβ
(µ,β,η)
− 1 |z|n+1−β
µ,β,η
|J0,z f (z)| ≤ |z|1−β +
φ1 (µ, β, η) (B + 1)(1 + λ(n − β))

where φk (µ, β, η) is given by (4.51).

Corollary 4.3.2 : Under the hypothesis of Theorem 4.3.3, f (z) is included

in a disc with center at the origin and radius ρ given by
 
[(B − A)(1 − α) + (B + 1) φ11−λβ
(µ,β,η)
− 1 ]φn+1 (µ, β, η)
ρ=1+ .
(B + 1)(1 + λ(n − β))

We claim that this result is entirely new and not found in the literature.

181
Theorem 4.3.4 : Let ψn (z) = z
 
1−λβ
[(B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 ]φk (µ, β, η)
ψk (z) = z− z k , k ≥ n+1.
(1 + B)(1 − λ + λ(k − β))

Then f (z) ∈ Mη (A, B, α, λ, µ, β) if and only if it can be expressed in the

form


∞
f (z) = ck ψk (z),
k=n
where


∞
ck ≥ 0 and ck = 1. (4.58)
k=n

Proof : Let f (z) ∈ Mη (A, B, α, λ, µ, β). Then by (4.53) and taking

(1 + B)(1 − λ + λ(k − β))

ck =   , (k ≥ n + 1)
[(B − A)(1 − α) + (B + 1) φ11−λβ
(µ,β,η)
− 1 ]φ k (µ, η)

∞
we have ck ≥ 0 and if we put cn = 1 − ck we obtain
k=n+1

∞
f (z) = z − ck z k
k=n+1
 
1−λβ


∞ [(B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 ]φk (µ, βη)
= z− ck z k
k=n+1
(1 + B)(1 − λ + λ(k − β))

∞
= z− ck (z − ψk (z))
k=n+1


∞

∞

∞
= (1 − ck )z + ck ψk (z) = ck ψk (z).
k=n+1 k=n+1 k=n

Conversely, let


∞
f (z) = ck ψk (z)
k=n
 
1−λβ


∞ [(B − A)(1 − α) + (B + 1) φ1 (µ,β,η)
− 1 ]φk (µ, βη)
= z− ck z k .
(1 + B)(1 − λ + λ(k − β))
k=n+1

182
Then we have (in view of (4.58))

   
1−λβ

∞
(1 + B)(1 − λ + λ(k − β))  [(B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 ]φk (µ, β, η)
ck 
φk (µ, β, η) (1 + B)(1 − λ + λ(k − β))
k=n+1
 
1 − λβ
≤ (B − A)(1 − α) + (B + 1) −1 .
φ1 (µ, β, η)

It follows therefore from Theorem 4.3.1 that f (z) ∈ Mη (A, B, α, λ, µ, β).

We do mention that the extreme points of the class Mη (A, B, α, λ, µ, β)
are the functions ψn and ψk , (k ≥ n + 1, n ∈ IN) given by Theorem 4.3.4.
In the next discussion, we concentrate upon getting the radii of close-to-

convexity, starlikeness and convexity.

Theorem 4.3.5 : Let f (z) ∈ T (n) and let f (z) ∈ Mη (A, B, α, λ, µ, β). Then
f (z) is close-to-convex of order δ(0 ≤ δ < 1) in
  k−1
1

(1 + B)(1 − λ + λ(k − β))(1 − δ)

|z| < r = inf      ,
k
φk (µ, β, η)k (B − A)(1 − α) + (B + 1) φ11−λβ
(µ,β,η)
(4.59)
where −1 ≤ A < B ≤ 1, 0 ≤ B < 1, 0 ≤ α < 1, 1 − λβ ≥ φ1 (µ, β, η),

λ ≥ 0, 0 ≤ β ≤ µ < 1, η > µ − 2, k ≥ n + 1, n ∈ IN and z ∈ U.

The result is sharp, the extremal function f (z) is given by (4.52).
Proof : It is enough to show that |f  (z) − 1| ≤ 1 − δ for |z| < r where r is
given by (4.59)


∞
kak
|f  (z) − 1| < 1 − δ if |z|k−1 ≤ 1.
k=n+1
1−δ

By Theorem 4.3.1, the last expression holds true if

 
k (1 + B)(1 − λ + λ(k − β))
|z|k−1 ≤    ,
1−δ (B − A)(1 − α) + (1 + B) φ11−λβ − 1 φ (µ, β, η)
(µ,β,η) k

183
then
  k−1
1

(1 + B)(1 − λ + λ(k − β))(1 − δ)

|z| ≤      .

k (B − A)(1 − α) + (1 + B) φ11−λβ
(µ,β,η)
− 1 φ k (µ, β, η)

Theorem 4.3.6 : Let f (z) ∈ T (n) and f (z) ∈ Mη (A, B, α, λ, µ, β). Then

f (z) is starlike of order δ(0 ≤ δ < 1) in |z| < r1 , where

  k−1
1

(1 − δ)(1 + B)(1 − λ + λ(k − β))

r1 = inf      .
k 1−λβ
φk (µ, β, η) (B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 (k − δ)
(4.60)
The result is sharp.
Proof : We want to show that

zf (z)

f (z) − 1 < 1 − δ.

The rest of the details involved are obvious and may be omitted.

By Theorem 0.2.3 [Alexander’s theorem], we have the following result.
Corollary 4.3.3 : Let f (z) ∈ T (n) and f (z) ∈ Mη (A, B, α, λ, µ, β). Then
f (z) is convex of order δ(0 ≤ δ < 1) in |z| < r2 , where
  k−1
1

(1 − δ)(1 + B)(1 − λ + λ(k − β))

r2 = inf      .
k 1−λβ
φk (µ, β, η) (B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 k(k − δ)
(4.61)

This result is entirely new.

Theorem 4.3.7 : Let f (z) ∈ T (n) be in the class Mη (A, B, α, λ, µ, β) and
let σ > 0. Then the function I σ f (z) (Jung -Kim-Srivastava operator) defined
by
 z ∞ 

σ
σ 2σ z σ−1 2
I f (z) = (log ) f (t)dt = z − ak z k (4.62)
zΓ(σ) 0 t k+1
k=n+1

184
belongs to the class Mη (A, B, α, λ, µ, β).
Proof : Note that

∞  σ
σ 2
I f (z) = z − ak z k .
k=n+1
k + 1

∞ % 2
&σ
Therefore if I σ f (z) = z − bk z k , then bk = k+1
ak and so
k=n+1

∞
(1 + B)(1 − λ + λ(k − β))
bk
k=n+1
φ k (µ, β, η)

∞  σ
(1 + B)(1 − λ + λ(k − β)) 2
= ak
k=n+1
φ k (µ, β, η) k + 1

∞
(1 + B)(1 − λ + λ(k − β))
≤ ak
k=n+1
φ k (µ, β, η)
 
1 − λβ
≤ (B − A)(1 − α) + (B + 1) −1 .
φ1 (µ, β, η)
Then by Theorem 4.3.1 we get the required result.

Theorem 4.3.8 : Let f (z) ∈ T (n) be in the class Mη (A, B, α, λ, µ, β). Let
σ > 0. Then the function I σ f (z) given by (4.62) is starlike in |z| < R where

R = inf
k
  k−1
1

σ
  (1 + B)(1 − λ + λ(k − β))(k + 1) 
  ,
1−λβ
k (B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 φk (µ, β, η)(2)σ

n ∈ IN. (4.63)

Proof : By Definition of I σ f (z), we have

∞  σ
σ 2
I f (z) = z − ak z k
k=n+1
k+1

and it is starlike if


∞  σ
2
k ak |z|k−1 ≤ 1.
k+1
k=n+1

185
The last expression is obtained by the same technique of Theorem 4.3.6 and
it is true if
 σ
2 (1 + B)(1 − λ + λ(k − β))
k |z|k−1 ≤   
k+1 1−λβ
φk (µ, β, η) (B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1

that is, if
  k−1
1

σ
(1 + B)(1 − λ + λ(k − β))(k + 1)
|z| ≤      .
1−λβ
kφk (µ, β, η) (B − A)(1 − α) + (B + 1) φ1 (µ,β,η) − 1 2 σ

This completes the proof of theorem.

Denote by A∗p the subclass of functions f (z) of the form

2p−1


p
f (z) = z + qk−p+1 z k−p+1 − 2 F1 (a, b; c; z), |z| < 1 (4.64)
k=p−1

(a,k−p+1)(b,k−p+1)
where qk−p+1 = (c,k−p+1)(k−p+1)!
. Therefore f (z) can be expressed in the form

∞
p
f (z) = z − ak z k , |z| < 1, (4.65)
k=p+1

where
Γ(a + k)Γ(b + k)Γ(c)
ak = , k ≥ p + 1. (4.66)
Γ(a)Γ(b)Γ(c + k)Γ(k + 1)
We denote by n − UCVδp (λ, β) the class of uniformly convex multivalent func-
tions in U and n − STδp (λ, β) the class of n-starlike multivalent functions in

U which are defined in Definition 0.1.4 and Definition 0.1.5 respectively.

The class n−STδp (λ, β) is a naturally way emerged as the class of functions
with the property that g ∈ n − UCVδp (λ, β) if and only if zg  ∈ n − STδp (λ, β).
This class is introduced by Kanas and Wiśniowska [13].

186
Definition 4.3.1 : For f ∈ A∗p , the fractional derivative of f of order δ is
defined by

1
∞
Γ(k + p)
Dzδ f (z) = z p−δ
− ak z k−δ , 0 ≤ δ < 1. (4.67)
Γ(2 − δ) k=2
Γ(k + p − δ)

By making use of (4.67), Srivastava and Owa [23] introduced the operator

Lδz f (z) = Γ(2 − δ)z δ Dzδ f (z), 0 ≤ δ < 1

and for δ = 0 we have L0z f (z) = f (z). Now let f (z) be in the form (4.65),
then

∞
Γ(k + p)
Lδz f (z) = Γ(2 − δ)z δ
Dzδ f (z) =z −p
ak z k (4.68)
k=p+1
Γ(k + p − δ)

∞
p
=z − ζ p (k, δ)ak z k (4.69)
k=p+1

where
Γ(2 − δ)Γ(k + p)
ζ p (k, δ) = .
Γ(k + p − δ)
Definition 4.3.2 : A function f (z) ∈ A∗p is said to be in the class n −
STδp (λ, β) if satisfies the inequality
 
 z(Lδz f (z))
 z(Lδz f (z))

Lδz f (z) Lδz f (z)
Re ≥ n − p + β, (4.70)
 λ z(Lδδz f (z))
+ 1 − pλ  z(Lδz f (z))

Lz f (z) λ Lδz f (z) + 1 − pλ

where 0 ≤ β < p, 0 ≤ λ < 1p , n ≥ 0, z ∈ U, 0 ≤ δ < 1 and p ∈ IN.

Theorem 4.3.9 : The function f (z) defined by (4.65) is in the class n −

STδp (λ, β) if and only if

∞
ζ p (k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]ak ≤ p − β. (4.71)
k=p+1

187
Proof : Let f ∈ n − STδp (λ, β). Then by Theorem 0.2.1, we have
 
 z(Lδz f (z))

Lδz f (z) iθ iθ
Re (1 + ne ) − (npe + β) ≥ 0,
 λ z(Lδδz f (z))
+ 1 − pλ 
L f (z) z

where −π ≤ θ ≤ π. By using (4.69), we have


∞ 

 (pz p − kζ p (k, δ)ak z k )(1 + neiθ ) 

 
k=p+1 iθ
Re
∞
∞ − (npe + β) ≥ 0.

 

 λpz p − λkζ p (k, δ)ak z k + (1 − pλ)z p − (1 − pλ)ζ p (k, δ)ak z k 
k=p+1 k=p+1

The last inequality holds for all z ∈ U. Choosing values of z on the real axis,
and letting z → 1− through real values, then we have


∞
∞ 

 p(1 + ne iθ
) − kζ p
(k, δ)(1 + ne iθ
)a − (npe iθ
+ β)(1 − (λk + 1 − pλ)ζ p
(k, δ)a ) 

 k k 
k=p+1 k=p+1
Re
∞ ≥ 0.

 

 1− (λk + 1 − pλ)ζ p (k, δ)a k 
k=p+1

By mean value theorem we obtain

∞
ζ p(k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]ak ≤ p − β.
k=p+1

Conversely, suppose (4.71) holds true, then by using the Theorem 0.2.2 it is

enough to show that

 
z(Lδz f (z))
z(Lδz f (z))

δ f (z)) δ f (z)
A = z(Lδ f (z))z
− p + n z(Lδ f (z))z
− p + β 
L L

λ Lδzzf (z) + (1 − pλ) λ Lδzzf (z) + 1 − pλ

 
z(Lδz f (z))
z(Lδz f (z))

δ f (z)) δ f (z)
< z(Lδ f (z))
+ p − n z(Lδ f (z))
− p − β  = B
L z L z

λ Lδzzf (z) + (1 − pλ) λ Lδzzf (z) + 1 − pλ

M
For letting eiθ = |M |
, where

z(Lδz f (z))
M =λ + 1 − pλ,
Lδz f (z)

188
we may write
 
1 z(Lδz f (z))
z(Lδz f (z))

B = + (p − β) λ + 1 − pλ −
|M | Lδz f (z) Lδz f (z)
 
z(Lδz f (z))
z(Lδz f (z))

neiθ δ
− p λ δ
+ 1 − pλ

Lz f (z) Lz f (z)

∞  ∞ 
p
pz − kζ p (k, δ)ak z k zp − (λk + 1 − pλ)ζ p (k, δ)ak z k
1 k=p+1  k=p+1 
= + (p − β)  −
|M | p
∞
p k

p
∞
p k

z − ζ (k, δ)ak z z − ζ (k, δ)ak z
k=p+1 k=p+1
∞  ∞ 
p
pz − kζ p (k, δ)ak z k zp − (λk + 1 − pλ)ζ p (k, δ)ak z k
k=p+1  k=p+1 
neiθ ∞ −p
 ∞


zp − ζ p (k, δ)ak z k zp − ζ p (k, δ)ak z k

k=p+1 k=p+1

∞ 
2p − β − [k(1 + n) + (λk + 1 − pλ)(p − β − np)]ζ p (k, δ)ak
1   k=p+1 

>
|M | 
∞ 
1− ζ p (k, δ)ak
k=p+1

 
1 z(Lδz f (z)) z(Lδz f (z))
A = − (p + β) λ + 1 − pλ − neiθ
|M| Lδz f (z) Lδz f (z)
 
z(Lδz f (z)) z(L δ
f (z)) 
− p λ z
+ 1 − pλ
Lδ f (z) L δ f (z)
z z
 ∞ 
p
β+ [k(1 + n) − (λk + 1 − pλ)(p + β + np)]ζ (k, δ)ak
1   k=p+1 

<  ∞ 
|M|
1− ζ p (k, δ)ak
k=p+1

It is easy to verify that B − A > 0, if (4.71) holds, therefore the proof is

complete.

Deduction 4.3.4 : If f (z) ∈ n − STδp (λ, β), then
(p − β)
ak ≤ .
ζ p (k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
Corollary 4.3.5 : f (z) ∈ 0 − ST01 (α, β) if and only if


∞
(k − 1)(1 − λβ)ak < 1 − β (0 ≤ β < 1, 0 ≤ λ < 1)
k=2

189
this is a class studied by Altintas and Owa [3].
Corollary 4.3.6 : f (z) ∈ 0 − ST01 (0, β) if and only if

∞
(k − 1)ak ≤ 1 − β (0 ≤ β < 1)
k=2

this is a class studied by Silverman [22].

In the next Theorem we obtain distortion theorem for the class n −

STδp (λ, β).

Theorem 4.3.10 : Let f (z) ∈ n − STδp (λ, β), then for |z| ≤ r < 1, we have

(p − β)
r p − r p+1 ≤ |Lδz f (z)|
ζ p (p + 1, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]

(p − β)
≤ r p + r p+1
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
and

(p − β)
pr p−1 − (p + 1)r p
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]

(p − β)
≤ |(Lδz f (z)) | ≤ pr p−1 −(p+1)r p .
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
Proof : In view of (4.71), we have

ζ p (p + 1, δ)[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]ap+1

∞
≤ ζ p (k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + b)]ak
k=p+1
≤ p − β,

then

∞
p−β
ak ≤ .
ζ p (p + 1, δ)[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
k=p+1

190
Hence


∞
|Lδz f (z)| p
≤ |z| + |z| p+1 p
ζ (p + 1, δ) ak
k=p+1


∞
≤ r p + r p+1 ζ p (p + 1, δ) ak
k=p+1
p−β
≤ r p + r p+1
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
and
p−β
|Lδz f (z)| ≥ r p − r p+1 .
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
Also, we have


∞
|(Lδz f (z)) | ≤ pr p−1 p p
+ (p + 1)r ζ (p + 1, δ) ak
k=p+1
p−β
≤ pr p−1 + r p
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
and
p−β
|Lδz f (z) | ≥ pr p − (p + 1)r p .
[(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]

Corollary 4.3.7 : Let the function f (z) given by (4.65) be in the class
n − STδp (λ, β). Then
r p−δ r p−δ+1 p−β
|Dzδ f (z)| ≥ −
Γ(2 − δ) Γ(2 − δ) [(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
r p−δ r p−δ+1 p−β
|Dzδ f (z)| ≤ + .
Γ(2 − δ) Γ(2 − δ) [(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
Corollary 4.3.8 : Let the function f (z) given by (4.65) be in the class
n − STδp (λ, β). Then
 z  
r p+1 r p+2 p−β
| f (t)dt| ≥ −
0 p + 1 p + 2 [(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]

191
  z  
r p+1 r p+2 p−β
| f (t)dt| ≤ + .
0 p+1 p+2 [(p + 1)(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
All the above results are entirely new.
Remark 4.3.1 : Under the hypothesis of Corollary 4.3.7, Dzδ f (z) is included
in a disc with its center at the origin and radius r given by
 
1 p−β
r= 1+ .
Γ(2 − δ) [(p + 1)(1 + n − λ(np + β) − (1 − pλ)(np + β)]

∞
Theorem 4.3.11 : Let the functions fi (z) = z p − ak,i z k , (ak,i ≥ 0,
k=p+1
k ∈ IN, k ≥ p + 1, i = 1, · · · , ) be in the class n − STδp (λ, β). Then the
function


∞
p
g(z) = z − bk z k , (bk ≥ 0, k ∈ IN, k ≥ p + 1)
k=p+1

also in the class n − STδp (λ, β) where bn = 1

ak,i.
i=1
Proof : Since fi (z) ∈ n − STδp (λ, β), then by Theorem 4.3.9, we have

∞
ζ p(k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]ak ≤ p − β.
k=p+1

∞
Now consider g(z) = z p − bk z k , so that
k=p+1

 


∞
1

ζ p (k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]ak ] ak,i

k=p+1
 i=1

1

p

∞
= ζ (k, δ)[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]ak,i
 i=1 k=p+1

1

≤ (p − β) ≤ p − β.
 i=1

This completes the proof of theorem.

192

∞
Theorem 4.3.12 : Let fi (z) = z p − ak,iz k , (ak,i ≥ 0, k ∈ IN,
k=p+1
k ≥ p + 1, i = 1, · · · , ) be in the class ni − STδp (λi , βi ). Then the function



∞
1
g(z) = z p −
ak,i z k is in the class n − STδp (λ, β), where
k=p+1 i=1

β = min {βi }, n = min {ni } and λi = min {λi }.

1≤i≤
1≤i≤
1≤i≤

Proof : By assumption fi (z) ∈ ni − STδp (λi , βi ) for every i = 1, · · · , ,

so by Theorem 4.3.9, we have

 


∞
1

ζ p (k, δ)[k(1 + ni − λi (ni p + βi )) − (1 − pλi )(ni p − βi )] ak,i

k=p+1
 i=1

1

p

∞
= ζ (k, δ)[k(1 + ni − λi (ni p + βi )) − (1 − pλi )(ni p − βi )]
 i=1 k=p+1

1

≤ (p − βi ) ≤ (p − βi ),
 i=1

then g(z) ∈ n − STδp (λ, β).

Theorem 4.3.13 : Let f ∈ n − STδp (λ, β). Then f is closed under convex
combination.
Proof : Let f and g be in the class n − STδp (λ, β),

∞

∞
p k p
f (z) = z − ak z , g(z) = z − ak z k ,
k=p+1 k=p+1

consider the function h(z) = λf (z) + (1 − λ)g(z), 0 ≤ λ ≤ 1, then

∞
p
h(z) = z − [λak + (1 − λ)ak ]z k .
k=p+1

By Theorem 4.3.9, we have

∞
[k(1 + n − λ(np + β)) − (1 − pλ)(np + β)]
ζ p (k, δ) [λak + (1 − λ)ak ]
p−β
k=p+1

193


∞
k(1 + n − λ(np + β)) − (1 − pλ)(np + β)
=λ ζ p (k, δ) ak +
k=p+1
(p − β)

∞
k(1 + n − λ(np + β)) − (1 − pλ)(np + β) 
(1 − λ) ζ p (k, δ) ak ≤ 1,
k=p+1
p−β

then h(z) ∈ n − STδp (λ, β).

Theorem 4.3.14 : Let f ∈ n − STδp (λ, β). Then
 z 
n − E(t)β)
Lδz f (z) = exp dt , |E(z)| < 1, z ∈ U.
0 (n − E(t))t

Proof : For n = 0 obvious.

z(Lδz f (z))

Let n = 0, for f ∈ n − STδp (λ, β) and w = Lδz f (z)
we have

Re w > n|w − 1| + β.

Thus

w − 1 1

w − β < n,

or
w−1 E(z)
= , where |E(z)| < 1, z ∈ U.
w−β n
This yields
z(Lδz f (z))

Lδz f (z)
−1 E(z)
z(Lδz f (z))

=
Lδz f (z)
−β n

or equivalently
(Lδz f (z)) n − βE(z)
= .
δ
Lz f (z) z(n − E(z)
Then we obtain the desired result by integration. This completes the proof

of theorem.

194
 REFERENCES

[1] R. Aghalary and S. R. Kulkarni, On a certain family of univa-

lent functions with negative coefficient defined by Ruscheweyh
derivative, Hacettepe Bulletin of Natural Sciences and Engi-

neering, 30 (2001), 45-54.

[2] O. P. Ahuja and J. M. Jahangiri, Inclusion relations between

classes of hypergeometric functions, J. Appl. Math. and Sto-

chastic Anal., 7(1) (1994), 79-89.

[3] O. Altintas and S. Owa, On subclasses of univalent functions

with negative coefficients, Pusan Kyongnam Math. ., 4 (1988),
41-46.

[4] O. Altintas and Y. Ertekin, A new subclass of analytic func-

tion with negative coefficients, Current topics on analytic func-
tion theory, World Scientific Publishing Company, New Jersey,

London and HongKong, (1992), 36-47.

[5] P. K. Banerji, L.Debnath and G. M. Shenan, Application of

fractional derivative operators to the mapping properties of
analytic functions, Fractional Calculus and Applied Analysis,
5(2) (2002), 169-180.

[6] S. D. Bernardi, Convex and starlike univalent functions, Trans.

Amer. Math. Soc., 135 (1969), 429-446.

[7] B. Carlson and D. B. Shaffer, Starlike and pre starlike hyperge-

ometric functions, SIAM J. Math. Anal., 15 (1984), 737-745.

195
[8] L. de Branges, A proof of the Bieberbach conjecture, Preprint
E-5-84, Steklov Math. Institute, Lomi, Leningrad (1984), 1-
12.

[9] J. Dziok and H. M. Srivastava, Classes of analytic functions as-

sociated with the generalized hypergeometric function, Appl.

Math. Comput., 103 (1999), 1-13.

[10] C. Fox, The asymptotic expansion of generalized hyperge-

ometric functions, Proc. London Math. Soc., (Ser. 2), 27

(1928), 389-400.

[11] J. M. Jahangiri, A new criterion for close-to-convexity of par-

tial sums of certain hypergeometric function, J. of Applied
Math. and Stochastic Analysis, 10(2) (1997), 197-202.

[12] I. B. Jung, Y. C. Kim and H. M. Srivastava, The Hardy space

of analytic functions associated with certain one-parameter
families of Integral Operators, J. Math. Appl., 176 (1993),

138-147.

[13] S. Kanas and A. Wiśniowska, Conic regions and k-uniform

convexity, J. Comp. Appl. Math., 105 (1999), 327-336.

[14] Y. C. Kim and F. Rønning, Integral transforms of certain

subclasses of analytic functions, J. Math. Anal. and Appl.,

256(2001), 466-489.

[15] E. P. Merkes and W. T. Scott, Starlike functions, Proc. Amer.

Math. Soc., 12 (1961), 885-888.

196
[16] K. I. Noore and M. A. Noore, On integral operators, J. Math.
Anal. Appl., 238 (1999), 341-352.

[17] S. Ponnusamy, Geometric properties for convolutions of hy-

pergeometric functions and functions with derivative in a half-
plane, Integral Transforms and Special Functions, 8 (102) (1990),

121-238.

[18] S. Owa, On the distortion theorems I, Kyungpook Math. J.,

18 (1978), 53-59.

[19] R. K. Raina and H. M. Srivastava, A certain subclass of ana-

lytic functions associated with operators of fractional calculus,

Comput. Math. Appl., 32 (1996), 13-19.

[20] M. Rosihan, M. H. Khan, V. Ravichandran and K. G. Sub-

ramanian, A class of multivalent functions with positive coef-
ficients defined by convolution, JIPAM, 6 (1) (2005).

[21] St. Ruscheweyh and V. Singh, On the order of starlikeness of

hypergeometric functions, J. Math. Anal. Appl., 113 (1986),
1-11.

[22] H. Silverman, Univalent functions with negative coefficients,

Proc. Amer. Math. Soc., 51 (1975), 109-116.

[23] H.M. Srivastava and S. Owa, An application of the fractional

derivative, Math. Japan, 29 (1984), 383-389.

[24] H. M. Srivastava, Some Fox-Wright generalized hypergeomet-

ric functions and associated families of convolution operators,

197
 Appl. Anal. and Discrete Math., 1 (2007), 56-71.

[25] E. M. Wright, The asymptotic expansion of the generalized

hypergeometric function, J. London Math. Soc., 10 (1935),
286-293.

198