A new subclass of analytic functions is introduced. For this class, firstly the Fekete-Szegö type coefficient inequalities are derived. Various known or new special cases of our results are also pointed out. Secondly some applications of our main results involving the Owa-Srivastava fractional operator are considered. Thus, as one of these applications of our result, we obtain the Fekete-Szegö type inequality for a class of normalized analytic functions, which is defined here by means of the Hadamard product (or convolution) and the Owa-Srivastava fractional operator.
1. Introduction and Definitions
Let 𝒜 denote the class of functions of the form
(1)f(z)=z+∑n=2∞anzn
which are analytic in the unit disk
(2)𝕌={z∈ℂ:|z|<1}.
Also let 𝒮 denote the subclass of 𝒜 consisting of univalent functions in 𝕌.
Fekete and Szegö [1] proved a noticeable result that the estimate
(3)|a3-λa22|≤{-4λ+3,λ≤01+2exp(-2λ1-λ),0≤λ≤14λ-3,λ≥1
holds for f∈𝒮. The result is sharp in the sense that for each λ there is a function in the class under consideration for which equality holds.
The coefficient functional
(4)ϕλ(f)=a3-λa22=16(f′′′(0)-3λ2(f′′(0))2)
on f∈𝒜 represents various geometric quantities as well as in the sense that this behaves well with respect to the rotation; namely,
(5)ϕλ(e-iθf(eiθz))=e2iθϕλ(f)(θ∈ℝ).
In fact, rather than the simplest case when
(6)ϕ0(f)=a3,
we have several important ones. For example,
(7)ϕ1(f)=a3-a22
represents Sf(0)/6, where Sf denotes the Schwarzian derivative
(8)Sf(z)=(f′′(z)f′(z))′-12(f′′(z)f′(z))2.
Moreover, the first two nontrivial coefficients of the nth root transform
(9)(f(zn))1/n=z+cn+1zn+1+c2n+1z2n+1+⋯
of f with the power series (1) are written by
(10)cn+1=a2n,c2n+1=a3n+(n-1)a222n2,
so that
(11)a3-λa22=n(c2n+1-μcn+12),
where
(12)μ=λn+n-12.
Thus it is quite natural to ask about inequalities for ϕλ corresponding to subclasses of 𝒮. This is called Fekete-Szegö problem. Actually, many authors have considered this problem for typical classes of univalent functions (see, e.g., [1–12]).
For two functions f and g, analytic in 𝕌, we say that the function f(z) is subordinate to g(z) in 𝕌, and we write
(13)f(z)≺g(z)(z∈𝕌),
if there exists a Schwarz function w(z), analytic in 𝕌, with
(14)w(0)=0,|w(z)|<1(z∈𝕌),
such that
(15)f(z)=g(w(z))(z∈𝕌).
In particular, if the function g is univalent in 𝕌, the above subordination is equivalent to
(16)f(0)=g(0),f(𝕌)⊂g(𝕌).
Let φ(z) be an analytic function with
(17)φ(0)=1,φ′(0)>0,Re{φ(z)}>0,(z∈𝕌),
which maps the open unit disk 𝕌 onto a star-like region with respect to 1 and is symmetric with respect to the real axis.
This paper contains analogues of (3) for the following classes of analytic functions.
Definition 1.
Let
(18)α>0,0≤γ≤1,b∈ℂ∖{0}.
A function f∈𝒜 is said to be in the class ℛb(α,γ;φ) if it satisfies the following subordination condition:
(19)1+1b((1-α+2γ)f(z)z+(α-2γ)f′(z)+γzf′′(z)-1f(z)z)≺φ(z),
where φ(z) is defined to be the same as above for z∈𝕌.
Remark 2.
(i) If we set
(20)α=2γ+1
in Definition 1, then we have the class
(21)ℛb(2γ+1,γ;φ)=ℛγb(φ)
which consists of functions satisfying
(22)1+1b(f′(z)+γzf′′(z)-1)≺φ(z)(z∈𝕌).
This class was introduced by Bansal [13].
If we set
(23)φ(z)=1+Az1+Bz(-1≤B<A≤1;z∈𝕌)
in Definition 1, then we have a new class
(24)ℛb(α,γ;1+Az1+Bz)=ℛb(α,γ;A,B)
which consists of functions satisfying
(25)1+1b((1-α+2γ)f(z)z+(α-2γ)f′(z)+γzf′′(z)-1f(z)z)≺1+Az1+Bz.
Taking
(26)α=2γ+1
in (25), we have the class
(27)ℛb(2γ+1,γ;A,B)=ℛγb(A,B)
which consists of functions satisfying
(28)1+1b(f′(z)+γzf′′(z)-1)≺1+Az1+Bzmmmmmm(-1≤B<A≤1;z∈𝕌).
This class was introduced by Bansal [14].
If we set
(29)γ=0,b=1,A=1,B=-1
in (25), then we have the class
(30)ℛ1(α,0;1,-1)=ℛ(α)
which consists of functions satisfying
(31)Re{(1-α)f(z)z+αf′(z)}>0.
This class was introduced by Murugusundaramoorthy and Magesh [15]. The subclass ℛ1(1,0;1,-1)=ℛ(1)=ℛ was studied by MacGregor [16].
We denote by 𝒫 the class of the analytic functions in 𝕌 with
(32)p(0)=1,Re{p(z)}>0.
We will need the following lemmas.
Lemma 3 (see [12]).
If p∈𝒫 with p(z)=1+c1z+c2z2+⋯, then
(33)|c2-νc12|≤{-4ν+2,ν≤02,0≤ν≤14ν-2,ν≥1.
When ν<0 or ν>1, equality holds true if and only if p(z) is (1+z)/(1-z) or one of its rotations. If 0<ν<1, then equality holds true if and only if p(z) is (1+z2)/(1-z2) or one of its rotations. If ν=0, then the equality holds true if and only if
(34)p(z)=(12+12η)1+z1-z+(12-12η)1-z1+z(0≤η≤1)
or one of its rotations. If ν=1, then the equality holds true if and only if p(z) is the reciprocal of one of the functions such that the equality holds true in the case when ν=0.
Although the above upper bound is sharp, in the case when 0<ν<1, it can be further improved as follows:
(35)|c2-νc12|+ν|c1|2≤2(0<ν≤12),|c2-νc12|+(1-ν)|c1|2≤2(12<ν≤1).
Lemma 4 (see [17]).
Let p∈𝒫 with p(z)=1+c1z+c2z2+⋯. Then for any complex number ν(36)|c2-νc12|≤2max{1,|2ν-1|},
and the result is sharp for the functions given by
(37)p(z)=1+z21-z2,p(z)=1+z1-z.
2. Fekete-Szegö Problem for the Function Class ℛb(α,γ;φ)
By making use of Lemma 3, we first prove the Fekete-Szegö type inequalities asserted by Theorem 5 below.
Theorem 5.
Let
(38)α>0,0≤γ≤1,b>0.
Also let
(39)φ(z)=1+B1z+B2z2+B3z3+⋯,
where
(40)B1>0,B2≥0.
If f(z) given by (1) belongs to the function class ℛb(α,γ;φ), then
(41)|a3-μa22|≤{B1b(1+2α+2γ)(B2B1-μB1b(1+2α+2γ)(1+α)2),μ≤σ1B1b(1+2α+2γ),σ1≤μ≤σ2B1b(1+2α+2γ)(-B2B1+μB1b(1+2α+2γ)(1+α)2),μ≥σ2,
where
(42)σ1=(B2-B1)(1+α)2B12b(1+2α+2γ),σ2=(B2+B1)(1+α)2B12b(1+2α+2γ),σ3=B2(1+α)2B12b(1+2α+2γ).
If σ1≤μ≤σ3, then
(43)|a3-μa22|+(1+α)2B1b(1+2α+2γ)×[1-B2B1+μB1b(1+2α+2γ)(1+α)2]|a2|2≤B1b(1+2α+2γ).
Furthermore, if σ3≤μ≤σ2, then
(44)|a3-μa22|+(1+α)2B1b(1+2α+2γ)×[1+B2B1-μB1b(1+2α+2γ)(1+α)2]|a2|2≤B1b(1+2α+2γ).
Each of these results is sharp.
Proof.
Since f∈ℛb(α,γ;φ), we have
(45)h(z)≺φ(z),
where
(46)h(z)=1+1b((1-α+2γ)f(z)z+(α-2γ)f′(z)+γzf′′(z)-1f(z)z)=1+h1z+h2z2+⋯.
From (46), we have
(47)h1=1b(1+α)a2,h2=1b(1+2α+2γ)a3.
Since φ(z) is univalent and h(z)≺φ(z), the function
(48)p1(z)=1+φ-1(h(z))1-φ-1(h(z))=1+c1z+c2z2+c3z3+⋯
is analytic and has a positive real part in 𝕌. Also we have
(49)h(z)=φ(p1(z)-1p1(z)+1)=1+B1c12z+[B12(c2-c122)+B2c124]z2+⋯.
Thus by (47) and (49) we get
(50)a2=B1c1b2(1+α),a3=b(1+2α+2γ)[B12(c2-c122)+B2c124].
Taking into account (50), we obtain
(51)a3-μa22=b(1+2α+2γ)×[B12(c2-c122)+B2c124]-μB12c12b24(1+α)2=B1b2(1+2α+2γ)(c2-δc12),
where
(52)δ=12(1-B2B1+μB1b(1+2α+2γ)(1+α)2).
The assertion of Theorem 5 now follows by an application of Lemma 3. On the other hand, using (51) for the values of σ1≤μ≤σ3, we have
(53)|a3-μa22|+(μ-σ1)|a2|2=B1b2(1+2α+2γ)|c2-δc12|+(μ-σ1)B12b2|c1|24(1+α)2=B1b2(1+2α+2γ)|c2-δc12|+(μ-(B2-B1)(1+α)2B12b(1+2α+2γ))B12b2|c1|24(1+α)2=B1b2(1+2α+2γ){|c2-δc12|+δ|c1|2}≤B1b(1+2α+2γ).
Similarly, for the values of σ3≤μ≤σ2, we get
(54)|a3-μa22|+(σ2-μ)|a2|2=B1b2(1+2α+2γ)|c2-δc12|+(σ2-μ)B12b2|c1|24(1+α)2=B1b2(1+2α+2γ)|c2-δc12|+((B2+B1)(1+α)2B12b(1+2α+2γ)-μ)B12b2|c1|24(1+α)2=B1b2(1+2α+2γ){|c2-δc12|+(1-δ)|c1|2}≤B1b(1+2α+2γ).
To show that the bounds asserted by Theorem 5 are sharp, we define the following functions:
(55)Kφn(z)(n=2,3,…),
with
(56)Kφn(0)=0=Kφn′(0)-1,
by
(57)1+1b((1-α+2γ)Kφn(z)z+(α-2γ)Kφn′(z)+γzKφn′′(z)-1Kφn(z)z)=φ(zn-1),
and the functions Fη(z) and Gη(z)(0≤η≤1), with
(58)Fη(0)=0=Fη′(0)-1,Gη(0)=0=Gη′(0)-1,
by
(59)1+1b((1-α+2γ)Fη(z)z+(α-2γ)Fη′(z)+γzFη′′(z)-1Fη(z)z)=φ(z(z+η)1+ηz),1+1b((1-α+2γ)Gη(z)z+(α-2γ)Gη′(z)+γzGη′′(z)-1Gη(z)z)=φ(-z(z+η)1+ηz),
respectively. Then, clearly, the functions Kφn, Fη, and Gη∈ℛb(α,γ;φ). We also write
(60)Kφ=Kφ2.
If μ<σ1 or μ>σ2, then the equality in Theorem 5 holds true if and only if f is Kφ or one of its rotations. When σ1<μ<σ2, then the equality holds true if and only if f is Kφ3 or one of its rotations. If μ=σ1, then the equality holds true if and only if f is Fη or one of its rotations. If μ=σ2, then the equality holds true if and only if f is Gη or one of its rotations.
By making use of Lemma 4, we immediately obtain the following Fekete-Szegö type inequality.
Theorem 6.
Let
(61)α>0,0≤γ≤1,b∈ℂ∖{0}.
Also let
(62)φ(z)=1+B1z+B2z2+B3z3+⋯,
where
(63)B1>0,B2≥0.
If f(z) given by (1) belongs to the function class ℛb(α,γ;φ), then for any complex number μ(64)|a3-μa22|≤B1|b|(1+2α+2γ)max{1,|B2B1-μB1b(1+2α+2γ)(1+α)2|}.
The result is sharp.
Remark 7.
The coefficient bounds for |a2| and |a3| are special cases of those asserted by Theorem 5.
Taking α=2γ+1 in Theorem 6, we have the following corollary.
Corollary 8 (see [13]).
Let
(65)0≤γ≤1,b∈ℂ∖{0}.
Also let
(66)φ(z)=1+B1z+B2z2+B3z3+⋯,
where
(67)B1>0,B2≥0.
If f(z) given by (1) belongs to the function class ℛγb(φ), then for any complex number μ(68)|a3-μa22|≤B1|b|3(1+2γ)max{1,|B2B1-3μB1b(1+2γ)4(1+γ)2|}.
The result is sharp.
If we set
(69)φ(z)=1+Az1+Bz(-1≤B<A≤1;z∈𝕌)
in Theorem 6, then we have
(70)B1=A-B,B2=-B(A-B).
So we get the following corollary.
Corollary 9.
Let
(71)α>0,0≤γ≤1,b∈ℂ∖{0}.
Also let
(72)φ(z)=1+Az1+Bz(-1≤B<A≤1;z∈𝕌).
If f(z) given by (1) belongs to the function class ℛb(α,γ;A,B), then for any complex number μ(73)|a3-μa22|≤(A-B)|b|(1+2α+2γ)×max{1,|B+μ(A-B)b(1+2α+2γ)(1+α)2|}.
The result is sharp.
Putting α=2γ+1 in Corollary 9, we obtain the following corollary.
Corollary 10 (see [13]).
Let
(74)0≤γ≤1,b∈ℂ∖{0}.
Also let
(75)φ(z)=1+Az1+Bz(-1≤B<A≤1;z∈𝕌).
If f(z) given by (1) belongs to the function class ℛγb(A,B), then for any complex number μ(76)|a3-μa22|≤(A-B)|b|3(1+2γ)×max{1,|B+3μ(A-B)b(1+2γ)4(1+γ)2|}.
The result is sharp.
Also putting γ=0, b=1, A=1, and B=-1 in Corollary 9, we obtain the following corollary.
Corollary 11.
Let α>0. If f(z) given by (1) belongs to the function class ℛ(α), then for any complex number μ(77)|a3-μa22|≤2(1+2α)max{1,|1-2μ(1+2α)(1+α)2|}.
3. Applications to Analytic Functions Defined by Using Fractional Calculus Operators and Convolution
The subject of fractional calculus (i.e., calculus of integrals and derivatives of any arbitrary real or complex order) has gained considerable popularity and importance during the past three decades or so. For the applications of the results given in the preceding sections, we first introduce the class ℛbρ(α,γ;φ), which is defined by means of the Hadamard product (or convolution) and a certain operator of fractional calculus, known as the Owa-Srivastava operator (see, e.g., [18, 19]).
Definition 12.
The fractional integral of order ρ is defined, for a function f(z), by
(78)Dz-ρf(z)=1Γ(ρ)ddz∫0zf(ζ)(z-ζ)1-ρdζ(ρ>0),
where the function f(z) is analytic in a simply connected domain of the complex z-plane containing the origin, and the multiplicity of (z-ζ)ρ-1 is removed by requiring log(z-ζ) to be real when z-ζ>0.
Definition 13.
The fractional derivative of order ρ is defined, for a function f(z), by
(79)Dzρf(z)=1Γ(1-ρ)ddz∫0zf(ζ)(z-ζ)ρdζ(0≤ρ<1),
where f(z) is constrained, and the multiplicity of (z-ζ)-ρ is removed, as in Definition 12.
Definition 14.
Under the hypotheses of Definition 13, the fractional derivative of order n+ρ is defined, for a function f(z), by
(80)hhDzn+ρf(z)=dndzn(Dzρf(z)),(0≤ρ<1,n∈ℕ0=ℕ∪{0}).
Using Definitions 12, 13, and 14, fractional derivatives, and fractional integrals, Owa and Srivastava [20] introduced the operator Ωρ:𝒜→𝒜 defined by
(81)(Ωρf)(z)=Γ(2-ρ)zρDzρf(z),ρ≠2,3,4,…(82)=z+∑n=2∞Γ(n+1)Γ(2-ρ)Γ(n-ρ+1)anzn.
This operator is known as the Owa-Srivastava operator. In terms of the Owa-Srivastava operator Ωρ defined by (81), we now introduce the function class ℛbρ(α,γ;φ) in the following way:
(83)ℛbρ(α,γ;φ)={f∈𝒜:Ωρf∈ℛb(α,γ;φ)}.
Note that the function class ℛbρ(α,γ;φ) is a special case of the function class ℛbg(α,γ;φ) when
(84)g(z)=z+∑n=2∞Γ(n+1)Γ(2-ρ)Γ(n-ρ+1)zn.
Now suppose that
(85)g(z)=z+∑n=2∞gnzn(gn>0).
Since
(86)f(z)=z+∑n=2∞anzn∈ℛbg(α,γ;φ)⟺(f*g)(z)=z+∑n=2∞gnanzn∈ℛb(α,γ;φ),
we can obtain the coefficient estimates for functions in the class ℛbg(α,γ;φ) from the corresponding estimates for functions in the class ℛb(α,γ;φ). By applying Theorem 5 to the following Hadamard product (or convolution):
(87)(f*g)(z)=z+g2a2z2+g3a3z3+⋯,
we get the following theorem after an obvious change of the parameter μ.
Theorem 15.
Let
(88)α>0,0≤γ≤1,b>0.
Also let
(89)φ(z)=1+B1z+B2z2+B3z3+⋯,
where
(90)B1>0,B2≥0,gn>0(n=3,4,…).
If f(z) given by (1) belongs to the function class ℛbg(α,γ;φ), then
(91)|a3-μa22|≤{B1b(1+2α+2γ)g3(B2B1-μg3g22B1b(1+2α+2γ)(1+α)2),μ≤σ1B1b(1+2α+2γ)g3,σ1≤μ≤σ2B1b(1+2α+2γ)g3(-B2B1+μg3g22B1b(1+2α+2γ)(1+α)2),μ≥σ2,
where
(92)σ1=g22g3(B2-B1)(1+α)2B12b(1+2α+2γ),σ2=g22g3(B2+B1)(1+α)2B12b(1+2α+2γ).
Each of these results is sharp.
When g corresponds to the Owa-Srivastava operator given in (82), we obtain
(93)g2=Γ(3)Γ(2-ρ)Γ(3-ρ)=22-ρ,(94)g3=Γ(4)Γ(2-ρ)Γ(4-ρ)=6(2-ρ)(3-ρ).
For g2 and g3 given by (93) and (94), respectively, Theorem 15 reduces to the following result.
Theorem 16.
Let
(95)α>0,0≤γ≤1,b>0.
Also let
(96)φ(z)=1+B1z+B2z2+B3z3+⋯,
where
(97)B1>0,B2≥0.
If f(z) given by (1) belongs to the function class ℛbg(α,γ;φ), then
(98)|a3-μa22|≤{(2-ρ)(3-ρ)B1b6(1+2α+2γ)(B2B1-μ3(2-ρ)2(3-ρ)B1b(1+2α+2γ)(1+α)2),μ≤σ1(2-ρ)(3-ρ)B1b6(1+2α+2γ),σ1≤μ≤σ2(2-ρ)(3-ρ)B1b6(1+2α+2γ)(-B2B1+μ3(2-ρ)2(3-ρ)B1b(1+2α+2γ)(1+α)2),μ≥σ2,
where
(99)σ1=2(3-ρ)3(2-ρ)(B2-B1)(1+α)2B12b(1+2α+2γ),σ2=2(3-ρ)3(2-ρ)(B2+B1)(1+α)2B12b(1+2α+2γ).
Each of these results is sharp.
Conflict of Interests
The author declares that there is no conflict of interests regarding the publication of this paper.
FeketeM.SzegöG.Eine bemerkung über ungerade schlichte funktionen193388589Abdel-GawadH. R.ThomasD. K.The Fekete-Szegö problem for strongly close-to-convex functions1992114234534910.2307/2159653MR1065939ZBL0741.30008Al-AmiriH. S.Certain generalizations of prestarlike functions1979283325334MR55728310.1017/S1446788700012283ZBL0434.30010ChoiJ. H.KimY. C.SugawaT.A general approach to the Fekete-Szegö problem2007593707727MR234482410.2969/jmsj/05930707ZBL1132.30007ChonweerayootA.ThomasD. K.UpakarnitikasetW.On the Fekete-Szegö theorem for close-to-convex functions1992661826MR1276678ZBL0793.30006DarusM.ThomasD. K.On the Fekete-Szegö theorem for close-to-convex functions1996443507511MR1422157ZBL0868.30015DarusM.ThomasD. K.On the Fekete-Szegö theorem for close-to-convex functions1998471125132MR1606344ZBL0922.30009KanasS.LeckoA.On the Fekete-Szegö problem and the domain of convexity for a certain class of univalent functions1990104957MR1114742ZBL0741.30012KeoghF. R.MerkesE. P.A coefficient inequality for certain classes of analytic functions196920812MR023292610.1090/S0002-9939-1969-0232926-9ZBL0165.09102KoepfW.On the Fekete-Szegö problem for close-to-convex functions19871011899510.2307/2046556MR897076ZBL0635.30020LondonR. R.Fekete-Szegö inequalities for close-to-convex functions1993117494795010.2307/2159520MR1150652ZBL0771.30007MaW. C.MindaD.A unified treatment of some special classes of univalent functionsProceedings of the Conference on Complex Analysis (Tianjin, 1992)1994Cambridge, Mass, USAInternational Press157169Conference on Proceedings Lecture Notes for Analysis, IMR1343506BansalD.Fekete-Szegö problem for a new class of analytic functions20112011514309510.1155/2011/143096MR2821966ZBL1220.30013BansalD.Upper bound of second Hankel determinant for a new class of analytic functions201326110310710.1016/j.aml.2012.04.002MR2971408ZBL1250.30006MurugusundaramoorthyG.MageshN.Coefficient inequalities for certain classes of analytic functions associated with Hankel determinant2009138589MR2578118MacGregorT. H.Functions whose derivative has a positive real part1962104532537MR014067410.1090/S0002-9947-1962-0140674-7ZBL0106.04805RavichandranV.GangadharanA.DarusM.Fekete-Szegö inequality for certain class of Bazilevic functions2004152171180MR2133419ZBL1073.30011OwaS.On the distortion theorems. I19781815359MR0507718ZBL0401.30009SrivastavaH. M.OwaS.1989New York, NY, USAHalsted Press (Ellis Horwood Limited, Chichester); John Wiley and SonsOwaS.SrivastavaH. M.Univalent and starlike generalized hypergeometric functions19873951057107710.4153/CJM-1987-054-3MR918587ZBL0611.33007