Journal of Computational and Applied Mathematics 105 (1999) 245–255
Classes of functions deÿned by certain dierential–integral operators J. Dziok ∗ Institute of Mathematics, Pedagogical University of RzeszÃow, ul. Rejtana 16A, PL-35-310 RzeszÃow, Poland Received 1 October 1997; received in revised form 25 February 1998 Dedicated to Haakon Waadeland on the occasion of his 70th birthday
Abstract In this paper we investigate a class of p-valent analytic functions with ÿxed argument of coecient, which is deÿned in terms of certain dierential–integral operators. Coecient estimates, distortion theorems, extreme points, the radii of c 1999 Elsevier Science Ltd. All rights reserved. convexity and starlikeness in this class are given. MSC: 30C45; 26A33
1. Introduction Let A(p; k), where p; k ∈ N = {1; 2; : : :}; p ¡ k, denote the class of functions f of the form p
f(z) = z +
∞ X
an z n ;
(1.1)
n=k
which are analytic in U = U(1), where U(r) = {z : |z| ¡ r}. Let A = A(1; 2). We say that a function f ∈ A(p; p + 1) is convex in U(r); r ∈ (0; 1]; if
Re 1 +
zf00 (z) ¿ 0; z ∈ U(r) f0 (z)
(1.2)
and we say that a function f ∈ A(p; p + 1) is starlike in U(r); r ∈ (0; 1]; if
zf0 (z) Re ¿ 0; f(z) ∗
z ∈ U(r):
E-mail:
[email protected].
c 1999 Elsevier Science B.V. All rights reserved. 0377-0427/99/$ - see front matter PII: S 0 3 7 7 - 0 4 2 7 ( 9 9 ) 0 0 0 1 4 - X
(1.3)
246
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
By Cp we denote the class of all functions convex in U and by Sp∗ we denote the class of all starlike functions in U. We say that a function f ∈ A is subordinate to a function F ∈ A, and write f ≺ F, if and only if there exists a function ! ∈ A; !(0) = 0; |!(z)| ¡ 1; z ∈ U, such that f(z) = F(!(z)); z ∈ U: Let p; k ∈ N; A; B; ; ; ∈ R; − ¡ p ¡ k; + ¿ − p; 06B61; −B6A ¡ B; (B 6= 1 or cos ¡ 0) and let denote the gamma function. In [3, 4] Kim and Srivastava et al. studied the following integral operator: Deÿnition 1.1. Let f ∈ A(p; k) and let log(z − ) be real when z − ¿ 0; z; ∈ U: The integral operator Q f(z) is deÿned for ¿ 0 and the function f by Q f(z) =
1 ()
Z z 0
1−
z
−1
−1 f() d:
Now, we deÿne the dierential–integral operator f(z) for 60: Then there exists a nonnegative integer m and a real number ; −1 ¡ 60; such that = − m: Deÿnition 1.2. Let f ∈ A(p; k) and let log(z − ) be real when z − ¿ 0; z; ∈ U: The integral operator f(z) is deÿned for 60 ( = − m; m ∈ N ∪ {0}; −1 ¡ 60) and for the function f by f(z)
=
d m+1 1 (1 + ) d z m+1
Z
0
z
(z − ) −1 f() d:
The multiplicities of (z − )−1 ; (z − ) in Deÿnitions 1.1 and 1.2 are removed by requiring log(z − ) ∈ R when z − ¿ 0: By using these deÿnitions we deÿne the linear operator
: A(p; k) → A(p; k) by
f(z) =
(p + + ) − z Q f(z) (p + )
f(z) =
(p + + ) 1−− z f(z) (p + )
for ¿ 0
and for 60:
Putting = 1 in Deÿnitions 1.1 and 1.2 we obtain the integral operator deÿned by Owa [5] and investigated by Srivastava and Owa [7], Dziok [1, 2] and others. Deÿnition 1.3. Let T (; ) denote the class of functions f ∈ A(p; k) satisfying the following condition:
f(z) 1 + Az : (1.4) ≺ p z 1 + Bz
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
247
Deÿnition 1.4. Let T (; ) denote the subclass of the class T (; ) of functions f of the form (1.1), such that arg an = for an 6= 0; n = k; k + 1; k + 2; : : : . We can write every function f from the class T (; ) in the form p
f(z) = z + e
i
∞ X
|an |z n ; z ∈ U:
(1.5)
n=k
In the present paper we obtain coecient estimates, distortion theorems, extreme points, the radii of convexity and starlikeness for the class T (; ) and coecient estimates for the class T (; ). 2. Coecient estimates By Deÿnition 1.4 we obtain: ˜ Lemma 2.1. If A6A˜ and B6B; then the class T (; ) for parameters A; B is included in the class ˜ ˜ T (; ) for parameters A; B: Lemma 2.2 (Ratti [6]). Let f be a function of the form (1.1). If f ≺ g and g is convex function, then |an |61; n = k; k + 1; k + 2; : : : . After some calculations we obtain: Lemma 2.3. If a function f of the form (1.1) belongs to the class A(p; k); then
f(z) = z p +
∞ (n + ) (p + + ) X an z n ; z ∈ U: (p + ) n=k (n + + )
Theorem 2.1. If a function f of the form (1.5) belongs to the class T (; ); then ∞ X n=k
−1 n |an |6(; A; B);
(2.1)
where B−A (; A; B) = p 2 1 − B sin2 − B cos
(2.2)
(n + + ) (p + ) : (n + ) (p + + )
(2.3)
and n
=
Proof. Let f ∈ T (; ). By Deÿnition 1.4 we obtain
f(z) 1 + A!(z) = ; zp 1 + B!(z)
248
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
where !(z) is an analytic function in U, such that !(0) = 0 and |!(z)| ¡ 1 for z ∈ U. Thus, we have f(z) − z p = |!(z)| ¡ 1: B f(z) − Az p Using Lemma 2.3 we obtain ∞ X n=k
where
n
−1 n−p ¡ B n |an |z
− A + Be
i
∞ X n=k
−1 n−p ; n |an |z
is deÿned by (2.3). Thus, putting z = r; 0 ¡ r ¡ 1; we have
|w| ¡ |B − A + Bwei |;
(2.4)
where w=
∞ X n=k
−1 n−p : n |an |r
Since w ∈ R, by (2.4) we have (1 − B2 )w2 − (2B(B − A)cos )w − (B − A)2 ¡ 0: Solving this inequality with respect to w we obtain ∞ X n=k
−1 n−p n |an |r
where (; A; B);
n
¡ (; A; B);
are deÿned by (2.2) and (2.3). Thus letting r → 1− we obtain (2.1).
Theorem 2.2. If a function f of the form (1.1) belongs to the class T (; ); then |an |6(B − A) where
n
n;
n = k; k + 1; k + 2; : : : ;
(2.5)
is deÿned by (2.3). The result is sharp.
Proof. Let a function f of the form (1.1) belong to the class T (; ) and let us put z : g(z) = (z −p f(z) − 1)(A − B)−1 ; h(z) = 1 + Bz By (1.4) we have g ≺ h. Since the function g is the function of the form g(z) =
∞ X
[(A − B)
n]
−1
an z n−p
n=k
and the function h is convex in U; by Lemma 1.2 we obtain [(A − B)
n]
−1
|an |61;
n = k; k + 1; k + 2; : : : :
Thus we have (2.5). Equality in (2.6) holds for the functions gn of the form gn (z) = h(z n−p ) = z n−p +
∞ X i=n−p+1
(−B) i−n+p z i ; n = k; k + 1; k + 2; : : : :
(2.6)
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
249
Thus equality in (2.5) holds for the functions fn of the form p
fn (z) = z + (A − B)
nz
n
+
∞ X
(A − B)
i
(−B) i−n z i ; n = k; k + 1; : : : :
i=n+1
By Theorem 2.1 we obtain: Corollary 2.1. If a function f of the form (1.5) belongs to the class T (; ); then |an |6(; A; B) where (; A; B);
n;
n = k; k + 1; k + 2; : : : ;
are deÿned by (2.2) and (2.3).
n
Putting = in Corollary 2.1 we obtain: Corollary 2.2. If a function f of the form (1.5) belongs to the class T (; ); then |an |6 where form
n
B−A 1+B
n;
n = k; k + 1; k + 2; : : : ;
is deÿned by (2.3). The result is sharp. The extremal functions are functions fn of the
fn (z) = z p −
B−A 1+B
nz
n
; n = k; k + 1; k + 2; : : : :
(2.7)
3. Distortion theorems and extreme points Theorem 3.1. If f ∈ T (; ); |z| = r ¡ 1; then for 60 r p − (; A; B)
k
r k 6|f(z)|6r p + (; A; B)
k
rk
(3.1)
and for 6 − 1 pr p−1 − k(; A; B) where (; A; B);
n
kr
k−1
6|f0 (z)|6pr p−1 + k(; A; B)
k−1
;
(3.2)
are deÿned by (2.2) and (2.3).
Proof. Let f ∈ T (; ); |z| = r ¡ 1: Since the sequences { decreasing and positive, by Theorem 2.1 we obtain ∞ X
kr
|an |6(; A; B)
for 60;
k
n}
for 60 and {
n =n}
for 6 − 1 are (3.3)
n=k ∞ X n=k
n|an |6k(; A; B)
k
for 6 − 1:
(3.4)
250
Since
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
∞ ∞ X X p i n |f(z)| = z + e |an |z 6r p + |an |r n n=k
= rp + rk
∞ X
n=k
|an |r n−k 6r p + r k
n=k
and
∞ X
|an |
n=k
∞ ∞ X X p i n p |f(z)| = z + e |an |z ¿r − |an |r n n=k
p
=r −r
k
∞ X
n=k
|an |r
n−k
p
¿r − r
k
n=k
∞ X
|an |;
n=k
by (3.3) we obtain (3.1). Using (3.4) the estimations (3.2) we prove analogously. Putting = in Theorem 3.1 we obtain: Corollary 3.1. If f ∈ T (; ); |z| = r ¡ 1; then for 60 B−A B−A k p k rp − k r 6|f(z)|6r + kr 1+B 1+B and for 6 − 1 B−A B−A k−1 k−1 pr p−1 − k 6|f0 (z)|6pr p−1 + k ; kr kr 1+B 1+B where n is deÿned by (2.3). The result is sharp. The extremal function is function fk of the form (2.7). Theorem 3.2. Let fk−1 (z) = z p and let fn ; n = k; k + 1; k + 2; : : : ; be deÿned by (2.7). A function f belongs to the class T (; ) if and only if it is of the form f(z) =
∞ X
n fn (z);
z ∈ U;
(3.5)
n=k−1
where
∞ X
n = 1 and n ¿0; n = k − 1; k; k + 1; : : : :
n=k−1
Proof. (⇒) Let a function f of the form f(z) = z p −
∞ X
|an |z n
n=k
belong to the class T (; ): Let us put 1 + B −1
n = |an |; n = k; k + 1; k + 2; : : : B−A n
(3.6)
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
251
and ∞ X
k−1 = 1 −
n :
n=k
We have n ¿0; n = k; k + 1; k + 2; : : : and by (1:7) we have k−1 ¿0: Thus, ∞ X
n fn (z) = k−1 fk−1 (z) +
n=k−1
∞ X
n fn (z)
n=k
= 1−
∞ X
! p
n z +
∞ X 1+B
n=k
=z p −
∞ X 1+B
B−A
n=k
−
n=k
∞ X
−1 n |an |
B−A
−1 p n |an |z
+
∞ X 1+B n=k
B−A
B−A z − 1+B p
nz
n
−1 p n |an |z
|an |z n = f(z)
n=k
and the condition (3.5) follows. (⇐=) Let a function f satisfy (3.5). Thus, ∞ X
f(z) =
n fn (z) = k−1 fk−1 +
n=k−1
n fn (z)
n=k
= 1−
!
∞ X
n z p +
n=k
= zp −
∞ X
n=k
∞ X B−A n=k
∞ X
1+B
n n z
zp −
B−A 1+B
n
n nz
n
and we can write the function f in the form (3.6), where |an | = n
B−A 1+B
n:
Since ∞ X n=k
−1 n |an |
=
∞ X n=k
n
B−A B−A B−A = (1 − 1−k )6 ; 1+B 1+B 1+B
we have f ∈ T (; ), which ends the proof. By Theorem 3.1 we have: Corollary 3.2. The class T (; ) is convex. The extremal points are functions fn of the form (2.7) and the function fk−1 (z) = z p .
252
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
4. The radii of convexity and starlikeness Theorem 4.1. If f ∈ T (; ); then a function f is convex in the disc U(rc ); where rc = inf
n¿k
n2 (; A; B) 2 p
and (; A; B);
n
!1=(p−n)
(4.1)
n
are deÿned by (2.2) and (2.3).
Proof. Let a function f of the form (1.5) belong to the class T (; ). By (1.2) the function f is convex in U(r); 0 ¡ r61; if 00 1 + zf (z) − p ¡ p; z ∈ U(r): (4.2) 0 f (z) Since i P∞ 2 00 n−1 e (z) (n − np)|a |z zf n n=k 1 + P − p = p−1 n−1 pz f0 (z) + ei ∞ n|a |z n n=k P ∞ (n2 − np)|a k z|n−p n n=k P 6 ; n−p p− ∞ n=k n|an k z|
putting |z| = r, condition (4.2) is true if ∞ X n2 n=k
p2
|an |r n−p 61:
(4.3)
From Theorem 2.1 we have ∞ X
[(; A; B)
−1 n ] |an |61;
(4.4)
n=k
where (; A; B);
n
are deÿned by (2.2) and (2.3). Thus it is sucient to show that
2
n n−p r 6[(; A; B) p2
n]
−1
for n = k; k + 1; k + 2; : : : ;
that is r6(bn )1=(p−n)
for n = k; k + 1; k + 2; : : : ;
(4.5)
where
n2 n: p2 Condition (4.5) is true for r = rc , where rc is deÿned by (4.1). Since bn = (; A; B)
(4.6)
bn ¿ 0; n = k; k + 1; k + 2; : : : and lim (bn )1=(p−n) = 1;
n→∞
we have 0 ¡ rc 61; which ends the proof.
(4.7)
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
253
By Theorem 4.1 we have: Corollary 4.1. Let 6 − 2 and let (; A; B); 2
k p2
n
be deÿned by (2.2) and (2.3). If
k (; A; B)61;
(4.8)
then T (; ) ⊂ Cp :
(4.9)
In the other case a function f ∈ T (; ) is convex in U(rc ); where k2 rc = (; A; B) 2 p
!1=(p−k)
:
k
(4.10)
Proof. Let f ∈ T (; ). By Theorem 4.1 the function f is convex in U(rc ), where rc is deÿned by (4.1). For 6 − 2 the sequence {bn }, deÿned by (4.6), is decreasing. Hence, if bk 61; that is condition (4.8) is true, then (bn )1=(p−n) ¿1 for = k; k + 1; k + 2; : : : : This, by (4.7) we have rc = 1; that is condition (4.9) is true. In the other case we have (bk )1=(p−k) 6(bn )1=(p−n) for n = k; k + 1; k + 2; : : : ; which gives (4.10). Since for 6 − 2 k2 (; A; B) 2 k ¡ 1; p by Corollary 4.1 we have: Corollary 4.2. If 6 − 2; then T (; ) ⊂ Cp : We now determine the radius of starlikeness for the class T (; ). Theorem 4.2. If f ∈ T (; ); then the function f is starlike in U(r ∗ ); where
n¿k
and (; A; B);
n
1=(p−n) n : n p are deÿned by (2.2) and (2.3).
r ∗ = inf (; A; B)
(4.11)
Proof. Let a function f of the form (1.5) belong to the class T (; ). By (1.3) the function f is starlike in U(r); 0 ¡ r61; if 0 zf (z) ¡ p; z ∈ U(r): (4.12) − p f(z)
254
Since
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
0 i P∞ n zf (z) e (n − p)|a |z n n=k f(z) − p = z p + ei P∞ |a |z n−1 n n=k P∞ n=k (n − p)|an ||z|n−p ; P 6 1 − ∞ |a ||z|n−p n
n=k
putting |z| = r, the condition (4.12) is true if ∞ X n n=k
p
|an |r n−p 61:
(4.13)
By (4.4) it is sucient to prove n n−p r 6[(; A; B) n ]−1 for n = k; k + 1; k + 2; : : : ; p that is r6(hn )1=(p−n) ; where
n n: p The last inequality is true for r = r ∗ , where r ∗ is deÿned by (4.11). Since hn = (; A; B)
hn ¿ 0 for n = k; k + 1; k + 2; : : : and lim (hn )1=(p−n) = 1;
n→∞
we have 0 ¡ r ∗ 61; which completes the proof. From Theorem 4.2 we obtain the following corollaries analogously as Corollaries 4.1, 4.2 from Theorem 4.1. Corollary 4.3. Let 6 − 1. If k (; A; B) k 61; p where (; A; B); n are deÿned by (2.2) and (2.3), then T (; ) ⊂ Sp∗ : In the other case a function f ∈ T (; ) is starlike in U(r ∗ ); where
k r = (; A; B) p ∗
1=(k−p)
k
Corollary 4.4. If 6 − 1; then T (; ) ⊂ Sp∗ :
:
J. Dziok / Journal of Computational and Applied Mathematics 105 (1999) 245–255
255
References [1] J. Dziok, Classes of p-valent analytic functions with ÿxed argument of coecients, I, to appear. [2] J. Dziok, Classes of p-valent analytic functions with ÿxed argument of coecients, II, Proceedings of the Fifth Finish-Polish–Ukrainian Sommer School in Complex Analysis, Poland, Lublin, 1996, to appear. [3] I.B. Jung, Y.C. Kim, H.M. Srivastava, The Hardy space of analytic functions associated with certain one-parameter families of integral operators, J. Math. Anal. Appl. 176 (1993) 138–147. [4] Y.C. Kim, K.S. Lee, H.M. Srivastava, Certain classes of integral operators associated with the Hardy space of analytic functions, Complex Variables Theory Appl. 20 (1992) 1–12. [5] S. Owa, On the distortion theorems, Kyungpook Math. J. 18 (1978) 53–59. [6] J.S. Ratti, The radius of univalence of certain analytic functions, Math. Z. 107 (1968) 241–248. [7] H.M. Srivastava, S. Owa, An application of the fractional derivative, Math. Japon. 29 (1984) 383–389.