Abstract

We define and study some generalized classes of Bazilevic functions associated with convex domains. These convex domains are formed by conic regions which are included in the right half plane. Such results as inclusion relationships and integral-preserving properties are proved. Some interesting special cases of the main results are also pointed out.

1. Introduction

Let 𝐴 denote the class of analytic functions 𝑓(𝑧) defined in the unit disc 𝐸={π‘§βˆΆ|𝑧|<1} and satisfying the conditions 𝑓(0)=0,π‘“ξ…ž(0)=1. Let 𝑆 denote the subclass of 𝐴 consisting of univalent functions in 𝐸, and let π‘†βˆ— and 𝐢 be the subclasses of 𝑆 which contains, respectively, star-like and convex in Bazilevič [1] introduced the class 𝐡(𝛼,𝛽,β„Ž,𝑔) as follows.

Let π‘“βˆˆπ΄. Then, π‘“βˆˆπ΅(𝛼,𝛽,β„Ž,𝑔),𝛼,𝛽 real and 𝛼>0 if ξ‚Έξ€œπ‘“(𝑧)=(𝛼+𝑖𝛽)𝑧0β„Ž(𝑧)𝑔𝛼(𝑑)π‘‘π‘–π›½βˆ’1𝑑𝑑1/(𝛼+𝑖𝛽),(1.1) for some π‘”βˆˆπ‘†βˆ— and Reβ„Ž(𝑧)>0,π‘§βˆˆπΈ.

The powers appearing in (1.1) are meant as principle values. The functions 𝑓 in the class 𝐡(𝛼,𝛽,β„Ž,𝑔) are shown to be analytic and univalent, see [1]. 𝐡(𝛼,𝛽,β„Ž,𝑔) is the largest known subclass of univalent functions defined by an explicit formula and contains many of the heavily researched subclasses of 𝑆. We note the following:(i)𝐡(1,0,1,𝑔)=𝐢, (ii)𝐡(1,0,π‘§π‘”ξ…ž/𝑔,𝑔)=π‘†βˆ—, (iii)𝐡(1,0,β„Ž,𝑔)=𝐾, where 𝐾 is the class of close-to-convex functions introduced by Kaplan [2],(iv)𝐡(cos𝛾,sin𝛾,cos(π‘§π‘”ξ…ž/(𝑔+𝑖sin𝛾)),𝑔) is the class of 𝛾-spiral like functions which are univalent for |𝛾|<πœ‹/2.

For analytic functions βˆ‘π‘“(𝑧)=βˆžπ‘›=0π‘Žπ‘›π‘§π‘› and βˆ‘π‘”(𝑧)=βˆžπ‘›=0𝑏𝑛𝑧𝑛, by π‘“βˆ—π‘” we denote the Hadamard product (convolution) of 𝑓 and 𝑔, defined by (π‘“βˆ—π‘”)(𝑧)=βˆžξ“π‘›=0π‘Žπ‘›π‘π‘›π‘§π‘›.(1.2) For π‘˜βˆˆ[0,∞), the conic domain Ξ©π‘˜ is defined in [3] as follows: Ξ©π‘˜=𝑒+π‘–π‘£βˆΆπ‘’>π‘˜(π‘’βˆ’1)2+𝑣2ξ‚Ό.(1.3) For fixed π‘˜,Ξ©π‘˜ represents the conic region bounded successively by the imaginary axis (π‘˜=0), the right branch of hyperbola (0<π‘˜<1), a parabola (π‘˜=1) and an ellipse (π‘˜>1).

The following univalent functions, defined by π‘π‘˜(𝑧) with π‘π‘˜(0)=1 and π‘ξ…žπ‘˜(0)>0, map the unit disc 𝐸 onto Ξ©π‘˜βˆΆπ‘π‘˜βŽ§βŽͺβŽͺβŽͺβŽͺ⎨βŽͺβŽͺβŽͺβŽͺ⎩(𝑧)=1+𝑧21βˆ’π‘§,(π‘˜=0),1+πœ‹2ξƒ©βˆšlog1+π‘§βˆš1βˆ’π‘§ξƒͺ22,(π‘˜=1),1+1βˆ’π‘˜2sinh2ξ‚ƒβˆšπ΄(π‘˜)arctanh𝑧2,(0<π‘˜<1),1+π‘˜2βˆ’1sin2ξ‚΅πœ‹πΉξ‚΅ξ‚™2𝐾(𝑑)𝑧𝑑,𝑑,(π‘˜>1),(1.4) where 𝐴(π‘˜)=(2/πœ‹)arccosπ‘˜,𝐹(𝑀,𝑑) is the Jacobi elliptic integral of the first kind:ξ€œπΉ(𝑀,𝑑)=𝑀0𝑑π‘₯√1βˆ’π‘₯2√1βˆ’π‘‘2π‘₯2,(1.5)and π‘‘βˆˆ(0,1) is chosen such that π‘˜=cosh(πœ‹πΎξ…ž(𝑑)/2𝐾(𝑑)), where 𝐾(𝑑) is the complete elliptic integral of the first kind, √𝐾(𝑑)=𝐹(1,𝑑),𝐾′(𝑑)=𝐾(1βˆ’π‘‘2).

It is known that π‘π‘˜(𝑧) are continuous as regards to π‘˜ and have real coefficients for π‘˜βˆˆ[0,∞).

Let 𝑃(π‘π‘˜) be the subclass of the class 𝑃 of Caratheodory functions 𝑝(𝑧), analytic in 𝐸 with 𝑝(0)=1 and such that 𝑝(𝑧) is subordinate to π‘π‘˜(𝑧), written as 𝑝(𝑧)β‰Ίπ‘π‘˜(𝑧) in 𝐸.

We define the following.

Definition 1.1. Let β„Ž(𝑧) be analytic in 𝐸 with β„Ž(0)=1. Then, β„Žβˆˆπ‘ƒπ‘š(π‘π‘˜) if and only if, for π‘šβ‰₯2,π‘˜βˆˆ[0,∞),β„Ž1,β„Ž2βˆˆπ‘ƒ(π‘π‘˜) we can write ξ‚€π‘šβ„Ž(𝑧)=4+12ξ‚β„Ž1ξ‚€π‘š(𝑧)βˆ’4βˆ’12ξ‚β„Ž2(𝑧),π‘§βˆˆπΈ.(1.6) We note that 𝑃2(π‘π‘˜)=𝑃(π‘π‘˜), and π‘ƒπ‘š(𝑝0)=π‘ƒπ‘š, see [4].

Definition 1.2. Let π‘“βˆˆπ΄. Then, 𝑓(𝑧) is said to belong to the class π‘˜βˆ’βˆͺπ‘…π‘š if and only if π‘§π‘“ξ…ž/π‘“βˆˆπ‘ƒπ‘š(π‘π‘˜) for π‘˜βˆˆ[0,∞),π‘šβ‰₯2, and π‘§βˆˆπΈ.
For π‘š=2,π‘˜=0, the class 0βˆ’βˆͺ𝑅2=𝑅2 coincides with the class π‘†βˆ— of starlike functions, and 0βˆ’βˆͺπ‘…π‘š=π‘…π‘š consists of analytic functions with bounded radius rotation, see [5, 6]. Also π‘˜βˆ’βˆͺ𝑅2 is the class βˆͺ𝑆𝑇 studied by several authors, see [7, 8].

Definition 1.3. Let π‘“βˆˆπ΄. Then, π‘“βˆˆπ‘˜βˆ’βˆͺπ΅π‘š(𝛼,𝛽,β„Ž,𝑔) if and only if 𝑓(𝑧) is as given by (1.1) for some π‘”βˆˆπ‘˜βˆ’βˆͺ𝑅2,β„Žβˆˆπ‘ƒπ‘š(π‘π‘˜) in 𝐸 with π‘˜βˆˆ[0,∞),π‘šβ‰₯2,𝛼>0 and 𝛽 real.
When π‘š=2 and π‘˜=0, we obtain the class 𝐡(𝛼,𝛽,β„Ž,𝑔) of Bazilevic functions.

We shall assume throughout, unless otherwise stated, that π‘˜βˆˆ[0,∞),π‘šβ‰₯2,𝛼>0,𝛽 real and π‘§βˆˆπΈ.

2. Preliminary Results

Lemma 2.1 (see [3]). Let 0β‰€π‘˜<∞, and let 𝛽0,𝛿 be any complex numbers with 𝛽0β‰ 0 and Re(𝛽0π‘˜/(π‘˜+1)+𝛿)>0. If β„Ž(𝑧) is analytic in 𝐸,β„Ž(0)=1 and satisfies ξ‚»β„Ž(𝑧)+π‘§β„Žβ€²(𝑧)𝛽0ξ‚Όβ„Ž(𝑧)+π›Ώβ‰Ίπ‘π‘˜(𝑧)(2.1) and π‘žπ‘˜(𝑧) is analytic solution of ξ‚»π‘žπ‘˜(𝑧)+π‘§π‘žξ…žπ‘˜(𝑧)𝛽0π‘žπ‘˜ξ‚Ό(𝑧)+𝛿=π‘π‘˜(𝑧),(2.2) then π‘žπ‘˜(𝑧) is univalent, β„Ž(𝑧)β‰Ίπ‘žπ‘˜(𝑧)β‰Ίπ‘π‘˜(𝑧),(2.3) and π‘žπ‘˜(𝑧) is the best dominant of (2.1).

Lemma 2.2 (see [9]). Let π‘ž(𝑧) be convex in 𝐸 and π‘—βˆΆπΈβ†’β„‚ with Re𝑗(𝑧)>0,π‘§βˆˆπΈ. If 𝑝(𝑧) is analytic in 𝐸 with 𝑝(0)=1 and satisfies {𝑝(𝑧)+𝑗(𝑧)β‹…π‘§π‘ξ…ž(𝑧)}β‰Ίπ‘ž(𝑧), then 𝑝(𝑧)β‰Ίπ‘ž(𝑧).

Lemma 2.3 (see [9]). Let 𝑒=𝑒1+𝑖𝑒2,𝑣=𝑣1+𝑖𝑣2, and let πœ“(𝑒,𝑣) be a complex-valued function satisfying the conditions(i)πœ“(𝑒,𝑣) is continuous in a domain π·βŠ‚β„‚2,(ii)(0,1)∈𝐷 and Reπœ“(1,0)>0,(iii)Reπœ“(𝑖𝑒2,𝑣1)≀0, whenever (𝑖𝑒2,𝑣1)∈𝐷 and 𝑣1β‰€βˆ’(1/2)(1+𝑒22).If β„Ž(𝑧)=1+𝑐1𝑧+𝑐2𝑧2+β‹― is a function analytic in 𝐸 such that (β„Ž(𝑧),π‘§β„Žξ…ž(𝑧))∈𝐷 and Reπœ“{β„Ž(𝑧),π‘§β„Žβ€²(𝑧)}>0 for π‘§βˆˆπΈ, then Reβ„Ž(𝑧)>0 in 𝐸.

3. Main Results

Theorem 3.1. Let (1/(1βˆ’π›Ύ)){π‘§π‘“ξ…ž(𝑧)/𝑓(𝑧)βˆ’π›Ύ}βˆˆπ‘ƒπ‘š(π‘π‘˜), for π‘§βˆˆπΈ and π›Ύβˆˆ[0,1]. Define 𝑔(𝑧)=(𝑐+1)π‘§βˆ’π‘ξ€œπ‘§0π‘‘π‘βˆ’1𝑓𝛼(𝑑)𝑑𝑑1/𝛼,𝛼>0,π‘βˆˆβ„‚,Re𝑐β‰₯0.(3.1) Then, (1/(1βˆ’π›Ύ)){π‘§π‘”ξ…ž(𝑧)/𝑔(𝑧)βˆ’π›Ύ}βˆˆπ‘ƒπ‘š(π‘π‘˜) in 𝐸. In particular π‘”βˆˆπ‘˜βˆ’βˆͺπ‘…π‘š in 𝐸.

Proof. Let 𝑧𝑔′(𝑧)𝑔(𝑧)=(1βˆ’π›Ύ)𝑝(𝑧)+𝛾,(3.2) where 𝑝(𝑧) is analytic in 𝐸 with 𝑝(0)=1, and let ξ‚€π‘šπ‘(𝑧)=4+12𝑝1ξ‚€π‘š(𝑧)βˆ’4βˆ’12𝑝2(𝑧).(3.3) From (3.1) and (3.2), we have 𝑔𝛼[](𝑧)=𝛼(1βˆ’π›Ύ)𝑝(𝑧)+𝑐+𝛼𝛾=𝑓𝛼(𝑧).(3.4) Logarithmic differentiation of (3.4) and some computation yield 𝑝(𝑧)+π‘§π‘ξ…ž(𝑧)=1𝛼(1βˆ’π›Ύ)𝑝(𝑧)+(𝑐+𝛼𝛾)ξ‚»1βˆ’π›Ύπ‘§π‘“ξ…ž(𝑧)𝑓(𝑧)βˆ’π›Ύ.(3.5) That is 𝑝(𝑧)+π‘§π‘ξ…ž(𝑧)𝛼(1βˆ’π›Ύ)𝑝(𝑧)+(𝑐+𝛼𝛾)βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έin𝐸.(3.6) Let πœ™π‘Ž,π‘βˆ‘(𝑧)=𝑧+βˆžπ‘›=2𝑧𝑛/((π‘›βˆ’1)π‘Ž+𝑏). Then, ξ‚΅πœ™π‘(𝑧)βˆ—π‘Ž,𝑏(𝑧)π‘§ξ‚Άπ‘Žξ€·=𝑝(𝑧)+π‘§π‘ξ…žξ€Έ(𝑧)𝑝(𝑧)+𝑏.(3.7) Using convolution technique (3.7) with π‘Ž=1/𝛼(1βˆ’π›Ύ),𝑏=(𝑐+𝛼𝛾)/𝛼(1βˆ’π›Ύ), we obtain, from (3.3) and (3.6), 𝑝𝑖(𝑧)+π‘§π‘ξ…žπ‘–(𝑧)𝛼(1βˆ’π›Ύ)𝑝𝑖(𝑧)+(𝑐+𝛼𝛾)β‰Ίπ‘π‘˜(𝑧)in𝐸,𝑖=1,2.(3.8) Since Re{(𝛼(1βˆ’π›Ύ)π‘˜/(π‘˜+1))+𝑐+𝛼𝛾}β‰₯0, we apply Lemma 2.1 with 𝛽0=𝛼(1βˆ’π›Ύ),𝛿=𝑐+𝛼𝛾 to obtain 𝑝𝑖(𝑧)β‰Ίπ‘žπ‘˜(𝑧)β‰Ίπ‘π‘˜(𝑧), where π‘žπ‘˜(𝑧) is the best dominant and is given as π‘žπ‘˜ξƒ¬π›½(𝑧)=0ξ€œ10𝑑𝛽0+π›Ώβˆ’1ξ€œexpπ‘§π‘‘π‘§π‘π‘˜(𝑒)βˆ’1𝑒𝑑𝑒𝛽0ξƒ­π‘‘π‘‘βˆ’1βˆ’π›Ώπ›½0.(3.9)
Consequently, π‘βˆˆπ‘ƒπ‘š(π‘π‘˜) in 𝐸, and this completes the result.
As a special case, we prove the following.

Corollary 3.2. Let π‘˜=0 and let (1/(1βˆ’π›Ύ1)){π‘§π‘“ξ…ž(𝑧)/𝑓(𝑧)βˆ’π›Ύ1}βˆˆπ‘ƒπ‘š in 𝐸. Then, for 𝑔 defined by (3.1), 1/(1βˆ’π›Ύ){π‘§π‘”ξ…ž(𝑧)/𝑔(𝑧)βˆ’π›Ύ}βˆˆπ‘ƒπ‘š in 𝐸 where

2𝛾=ξ‚»ξ€·2π‘βˆ’2𝛼𝛾1ξ€Έ++1ξ€·2π‘βˆ’2𝛼𝛾1ξ€Έ+12ξ‚Ό+8𝛼.(3.10)

Proof. We can write π‘§π‘“ξ…ž(𝑧)=𝑓(𝑧)1βˆ’π›Ύ1ξ€Έβ„Ž(𝑧)+𝛾1,(3.11) where β„Žβˆˆπ‘ƒπ‘š in 𝐸.
Now proceeding as before, we have, withπ‘§π‘”ξ…ž(𝑧)ξ‚€π‘šπ‘”(𝑧)=(1βˆ’π›Ύ)𝑝(𝑧)+𝛾=4+12(1βˆ’π›Ύ)𝑝1ξ€Ύβˆ’ξ‚€π‘š(𝑧)+𝛾4βˆ’12(1βˆ’π›Ύ)𝑝2ξ€Ύ(𝑧)+𝛾(3.12)(1βˆ’π›Ύ)𝑝(𝑧)+𝛾+(1βˆ’π›Ύ)π‘§π‘ξ…ž(𝑧)=𝛼(1βˆ’π›Ύ)𝑝(𝑧)+(𝑐+𝛼𝛾)π‘§π‘“ξ…ž(𝑧).𝑓(𝑧)(3.13) Using convolution technique together with (3.11), we obtain ξ‚»Re(1βˆ’π›Ύ)𝑝𝑖(𝑧)+π›Ύβˆ’π›Ύ1ξ€Έ+(1βˆ’π›Ύ)π‘§π‘ξ…žπ‘–(𝑧)𝛼(1βˆ’π›Ύ)𝑝𝑖(𝑧)+(𝑐+𝛼𝛾)>0,(3.14) for 𝑖=1,2.
We construct the functional πœ“(𝑒,𝑣) by taking 𝑒=𝑝𝑖(𝑧),𝑣=π‘§π‘ξ…žπ‘–(𝑧) asξ€·πœ“(𝑒,𝑣)=(1βˆ’π›Ύ)𝑒+π›Ύβˆ’π›Ύ1ξ€Έ+(1βˆ’π›Ύ)𝑣𝛼(1βˆ’π›Ύ)𝑒+(𝑐+𝛼𝛾).(3.15) The first two conditions of Lemma 2.3 are clearly satisfied. We verify condition (iii) as follows. ξ€·Reπœ“π‘–π‘’2,𝑣1ξ€Έ=ξ€·π›Ύβˆ’π›Ύ1ξ€Έξ‚»+Re(1βˆ’π›Ύ)𝑣1𝑖𝛼(1βˆ’π›Ύ)𝑒2ξ‚Ό,=ξ€·+(𝑐+𝛼𝛾)π›Ύβˆ’π›Ύ1ξ€Έ+(𝑐+𝛼𝛾)(1βˆ’π›Ύ)𝑣1(𝑐+𝛼𝛾)2+𝛼2(1βˆ’π›Ύ)2𝑒22,β‰€ξ€·π›Ύβˆ’π›Ύ1ξ€Έ+ξ€·(𝑐+𝛼𝛾)(1βˆ’π›Ύ)1+𝑒22ξ€Έ2ξ€Ί(𝑐+𝛼𝛾)2+𝛼2(1βˆ’π›Ύ)2𝑒22ξ€»,𝑣1β‰€βˆ’1+𝑒222ξƒͺ,=𝐴+𝐡𝑒22,2𝐢(3.16) where
𝐴=2(π›Ύβˆ’π›Ύ1)(𝑐+𝛼𝛾)2βˆ’(1βˆ’π›Ύ)(𝑐+𝛼𝛾),𝐡=2𝛼2(π›Ύβˆ’π›Ύ1)(1βˆ’π›Ύ)2βˆ’(1βˆ’π›Ύ)(𝑐+𝛼𝛾), 𝐢=(𝑐+𝛼𝛾)2+𝛼2(1βˆ’π›Ύ)2𝑒22>0.
Reπœ“(𝑖𝑒2,𝑣1)≀0 if and only if 𝐴≀0,𝐡≀0. From 𝐴≀0, we obtain 𝛾 as given by (3.10) and 𝐡≀0 ensures that π›Ύβˆˆ[0,1).
Now proceeding as before, it follows from (3.12) that π‘βˆˆπ‘ƒπ‘š, and this proves our result.

By assigning certain permissible values to different parameters, we obtain several new and some known result.

Corollary 3.3. Let π‘“βˆˆπ‘˜βˆ’βˆͺ𝑅2=π‘˜βˆ’βˆͺ𝑆𝑇. Then, it is known that π‘“βˆˆπ‘†βˆ—(𝛾1),𝛾1=π‘˜/(π‘˜+1) and, form Corollary 3.2, it follows that π‘”βˆˆπ‘†βˆ—(𝛾) where 𝛾 is given by (3.10). Also a starlike function is π‘˜-uniformly convex for |𝑧|<π‘Ÿπ‘˜, π‘Ÿπ‘˜=12√(π‘˜+1)+4π‘˜2[8]+6π‘˜+3,see.(3.17) Therefore, for π‘“βˆˆπ‘˜βˆ’βˆͺ𝑅2, it follows that (1/(1βˆ’π›Ύ)){(π‘§π‘”ξ…ž(𝑧))ξ…ž/π‘”ξ…ž(𝑧)βˆ’π›Ύ}β‰Ίπ‘π‘˜ for |𝑧|<π‘Ÿπ‘˜, where 𝛾 is given by (3.10).
As special cases we note the following.
(i)For π‘˜=0, we have π‘Ÿ0√=1/(2+3) and π‘“βˆˆπ‘†βˆ—(0) implies that π‘”βˆˆπΆ(π›Ύβˆ—), with π›Ύβˆ—=22√(𝑐+1)+2(𝑐+1)2+8𝛼.(3.18)(ii)When π‘˜=1, we have 𝛾1√=1/2,𝛾=2/((2π‘βˆ’π›Ό+1)+(2π‘βˆ’π›Ό+1)2+8𝛼) and π‘Ÿ1√=1/(4+13).

Theorem 3.4. Let πΉβˆˆπ‘˜βˆ’βˆͺπ΅π‘š(𝛼,𝛽,𝑝,𝑓),π‘“βˆˆπ‘˜βˆ’βˆͺ𝑅2,π‘βˆˆπ‘ƒπ‘š(π‘π‘˜). Define, for Re[π›Όπ‘˜/(π‘˜+1)+(𝑐+𝑖𝛽)]>0, 𝐺(𝑧)=(𝑐+1)π‘§βˆ’π‘ξ€œπ‘§0π‘‘π‘βˆ’1𝐹𝛼+𝑖𝛽(𝑑)𝑑𝑑1/(𝛼+𝑖𝛽).(3.19) Then, πΊβˆˆπ‘˜βˆ’βˆͺπ΅π‘š(𝛼,𝛽,β„Ž,𝑔) in 𝐸, where 𝑔(𝑧) is given by (3.1), and β„Ž(𝑧) is analytic in 𝐸 with β„Ž(0)=1.

Proof. Set π‘§πΊξ…ž(𝑧)𝐺𝛼+π‘–π›½βˆ’1(𝑧)π‘§π‘–π›½π‘”π›Όξ‚€π‘š(𝑧)=β„Ž(𝑧)=4+12ξ‚β„Ž1ξ‚€π‘š(𝑧)βˆ’4βˆ’12ξ‚β„Ž2(𝑧).(3.20) We note that β„Ž(𝑧) is analytic in 𝐸 with β„Ž(0)=1. From (3.20), we have 𝑧𝑖𝛽𝑔𝛼(𝑧)π‘§β„Žξ…žξ‚Έπ›Ό(𝑧)+β„Ž(𝑧)π‘§π‘”ξ…ž(𝑧)𝑔(𝑧)+𝑐+𝑖𝛽=π‘§πΉξ…ž(𝑧)𝐹𝛼+π‘–π›½βˆ’1(𝑧),(3.21) using (3.1), we note that 𝑓(𝑧)𝑔(𝑧)𝛼=π›Όπ‘§π‘”ξ…ž(𝑧)𝑔(𝑧)+𝑐+𝑖𝛽.(3.22) From (3.21) and (3.22), it follows that ξ‚»β„Ž(𝑧)+π‘§β„Žξ…ž(𝑧)π›Όβ„Ž0ξ‚Ό(𝑧)+(𝑐+𝑖𝛽)βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έ,(3.23) where β„Ž0(𝑧)=π‘§π‘”ξ…ž(𝑧)/𝑔(𝑧)βˆˆπ‘ƒ(π‘π‘˜) since π‘”βˆˆπ‘˜βˆ’βˆͺ𝑅2 by Theorem 3.1.
It can easily be seen that π‘”βˆˆπ‘†βˆ—(π‘˜/(π‘˜+1)) and Re{π›Όπ‘§π‘”ξ…ž(𝑧)/𝑔(𝑧)+𝑐+𝑖𝛽}>0.
Now, using (3.8), we can easily derive ξ€½β„Žπ‘–ξ€·(𝑧)+𝑗(𝑧)π‘§β„Žξ…žπ‘–(𝑧)ξ€Έξ€Ύβ‰Ίπ‘π‘˜(𝑧)in𝐸,𝑖=1,2,(3.24) where 1/𝑗(𝑧)={π›Όπ‘§π‘”ξ…ž(𝑧)/𝑔(𝑧)+𝑐+𝑖𝛽} and Re𝑗(𝑧)>0.
Applying Lemma 2.2, it follows from (3.24) β„Žπ‘–(𝑧)β‰Ίπ‘π‘˜(𝑧) in 𝐸 and therefore β„Žβˆˆπ‘ƒπ‘š(π‘π‘˜) in 𝐸. This completes the proof.

Theorem 3.5. Let 𝑓(𝑧) be given by (1.1) with β„Ž(𝑧)=1,{(𝛼2+𝛽2)1/2𝑒𝑖𝛾(π‘§π‘”ξ…ž/𝑔)}βˆˆπ‘ƒπ‘š(π‘π‘˜)(𝛼2+𝛽2)1/2𝑒𝑖𝛾=𝛼+𝑖𝛽,|𝛾|<πœ‹/2. Then, for π‘§βˆˆπΈ(i)𝑒𝑖𝛾(π‘§π‘“ξ…ž(𝑧)/𝑓(𝑧))=cos𝛾(𝑝(𝑧))+𝑖sin𝛾,π‘βˆˆπ‘ƒπ‘š(π‘π‘˜), (ii)For π›Όξ…ž+π‘–π›½ξ…ž=𝑑(𝛼+𝑖𝛽),𝑑β‰₯1,π‘˜βˆ’βˆͺπ΅π‘š(𝛼,𝛽,1,𝑔)βŠ‚π‘˜π‘‘βˆ’βˆͺπ΅π‘šξ€·π›Όξ…ž,π›½ξ…žξ€Έ,1,𝑔.(3.25)

Proof. (i) From (1.1), we have 1+π‘§π‘“ξ…žξ…ž(𝑧)π‘“ξ…ž(𝑧)+(π›Όβˆ’1+𝑖𝛽)π‘§π‘“ξ…ž(𝑧)=𝛼𝑓(𝑧)π‘§π‘”ξ…ž(𝑧)𝑔(𝑧)+𝑖𝛽=𝐻2(𝑧),𝐻2βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έin𝐸.(3.26) Define a function 𝑝(𝑧) analytic in 𝐸 by π‘’π‘–π›Ύπ‘§π‘“ξ…ž(𝑧)𝑓(𝑧)=cos𝛾(𝑝(𝑧))+𝑖sin𝛾,𝛾=tanβˆ’1𝛽𝛼.(3.27) We can easily check that 𝑝(0)=1.
Now, from (3.26) and (3.27), we have ξ‚Έπ‘§π‘ξ…ž(𝑧)𝑝(𝑧)+𝑖tan𝛾+𝛼𝑝(𝑧)+π‘–π›½βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έin𝐸.(3.28) That is ξ‚Έπ›Όπ‘§π‘ξ…ž(𝑧)𝛼𝑝(𝑧)+𝑖𝛽+𝛼𝑝(𝑧)+π‘–π›½βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έ,(3.29) and, with β„Ž(𝑧)=𝛼𝑝(𝑧)+𝑖𝛽=(π‘š/4+1/2)β„Ž1(𝑧)βˆ’(π‘š/4βˆ’1/2)β„Ž2(𝑧), we apply convolution technique used before to have ξ‚»β„Žπ‘–(𝑧)+π‘§β„Žπ‘–ξ…ž(𝑧)β„Žπ‘–(𝑧)β‰Ίπ‘π‘˜(𝑧)in𝐸.(3.30) Applying Lemma, it follows that β„Žπ‘–(𝑧)β‰Ίπ‘žπ‘˜(𝑧)β‰Ίπ‘π‘˜(𝑧),π‘§βˆˆπΈ,(3.31) where π‘žπ‘˜(𝑧) is the best dominant and is given by π‘žπ‘˜ξ‚Έξ€œ(𝑧)=10ξ‚΅ξ€œexp0π‘‘π‘§π‘π‘˜(𝑒)βˆ’1π‘’π‘‘π‘’ξ‚Άξ‚Ήβˆ’1.(3.32) From (3.31), we have β„Ž(𝑧)=(𝛼𝑝(𝑧)+𝑖𝛽)βˆˆπ‘ƒπ‘š(π‘π‘˜) in 𝐸, and this proves part (i).
(ii) From part (i), we have 𝛼2+𝛽2ξ€Έ1/2π‘’π‘–π›Ύπ‘§π‘“ξ…ž(𝑧)𝑓(𝑧)=𝐻1(𝑧),𝐻1βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έin𝐸.(3.33) Now, 1+π‘§π‘“ξ…žξ…ž(𝑧)π‘“ξ…ž+𝛼(𝑧)ξ…žβˆ’1+π‘–π›½ξ…žξ€Έπ‘§π‘“ξ…ž(𝑧)=𝑓(𝑧)1+π‘§π‘“ξ…žξ…ž(𝑧)π‘“ξ…ž(𝑧)+(π›Όβˆ’1+𝑖𝛽)π‘§π‘“ξ…ž(𝑧)𝑓𝛼(𝑧)+(π‘‘βˆ’1)2+𝛽2ξ€Έ1/2π‘’π‘–π›Ύπ‘§π‘“ξ…ž(𝑧),𝑓(𝑧)=𝐻2(𝑧)+(π‘‘βˆ’1)𝐻1(𝑧),π»π‘–βˆˆπ‘ƒπ‘šξ€·π‘π‘˜ξ€Έ1,𝑖=1,2,=𝑑1βˆ’π‘‘ξ‚π»11(𝑧)+𝑑𝐻2ξ‚„,(𝑧)=𝑑𝐻,𝑑β‰₯1,(3.34)π»βˆˆπ‘ƒπ‘š(π‘π‘˜), since π‘ƒπ‘š(π‘π‘˜) is convex set, see [8].
Therefore, π‘“βˆˆπ‘˜π‘‘βˆ’βˆͺπ΅π‘š(𝛼′,𝛽′,1,𝑔) for π‘§βˆˆπΈ. This completes the proof.

As a special case, with π‘š=2,π‘˜=0, we obtain a result proved in [10].

By assigning certain permissible values to the parameters 𝛼,𝛽 and π‘š, we have several other new results.

Acknowledgment

The authors are grateful to Dr. S. M. Junaid Zaidi, Rector, COMSATS Institute of Information Technology, Islamabad, Pakistan, for providing excellent research facilities and environment