Research article

Abelian and Tauberian results for the fractional Fourier cosine (sine) transform

  • Received: 04 January 2024 Revised: 13 March 2024 Accepted: 25 March 2024 Published: 28 March 2024
  • MSC : 40E05, 43A50, 46F10, 46F12

  • In this paper, we presented Tauberian type results that intricately link the quasi-asymptotic behavior of both even and odd distributions to the corresponding asymptotic properties of their fractional Fourier cosine and sine transforms. We also obtained a structural theorem of Abelian type for the quasi-asymptotic boundedness of even (resp. odd) distributions with respect to their fractional Fourier cosine transform (FrFCT) (resp. fractional Fourier sine transform (FrFST)). In both cases, we quantified the scaling asymptotic properties of distributions by asymptotic comparisons with Karamata regularly varying functions.

    Citation: Snježana Maksimović, Sanja Atanasova, Zoran D. Mitrović, Salma Haque, Nabil Mlaiki. Abelian and Tauberian results for the fractional Fourier cosine (sine) transform[J]. AIMS Mathematics, 2024, 9(5): 12225-12238. doi: 10.3934/math.2024597

    Related Papers:

    [1] Zehba Raizah, Arshad Khan, Saadat Hussain Awan, Anwar Saeed, Ahmed M. Galal, Wajaree Weera . Time-dependent fractional second-grade fluid flow through a channel influenced by unsteady motion of a bottom plate. AIMS Mathematics, 2023, 8(1): 423-446. doi: 10.3934/math.2023020
    [2] Yong-Ki Ma, Marimuthu Mohan Raja, Kottakkaran Sooppy Nisar, Anurag Shukla, Velusamy Vijayakumar . Results on controllability for Sobolev type fractional differential equations of order $ 1 < r < 2 $ with finite delay. AIMS Mathematics, 2022, 7(6): 10215-10233. doi: 10.3934/math.2022568
    [3] Yudhveer Singh, Devendra Kumar, Kanak Modi, Vinod Gill . A new approach to solve Cattaneo-Hristov diffusion model and fractional diffusion equations with Hilfer-Prabhakar derivative. AIMS Mathematics, 2020, 5(2): 843-855. doi: 10.3934/math.2020057
    [4] Rongbo Wang, Qiang Feng, Jinyi Ji . The discrete convolution for fractional cosine-sine series and its application in convolution equations. AIMS Mathematics, 2024, 9(2): 2641-2656. doi: 10.3934/math.2024130
    [5] Yue Liang . Study on the controllability of delayed evolution inclusions involving fractional derivatives. AIMS Mathematics, 2024, 9(7): 17984-17996. doi: 10.3934/math.2024876
    [6] Ahmed Salem, Kholoud N. Alharbi . Fractional infinite time-delay evolution equations with non-instantaneous impulsive. AIMS Mathematics, 2023, 8(6): 12943-12963. doi: 10.3934/math.2023652
    [7] Anumanthappa Ganesh, Swaminathan Deepa, Dumitru Baleanu, Shyam Sundar Santra, Osama Moaaz, Vediyappan Govindan, Rifaqat Ali . Hyers-Ulam-Mittag-Leffler stability of fractional differential equations with two caputo derivative using fractional fourier transform. AIMS Mathematics, 2022, 7(2): 1791-1810. doi: 10.3934/math.2022103
    [8] Kehong Zheng, Fuzhang Wang, Muhammad Kamran, Rewayat Khan, Ali Sikandar Khan, Sadique Rehman, Aamir Farooq . On rate type fluid flow induced by rectified sine pulses. AIMS Mathematics, 2022, 7(2): 1615-1627. doi: 10.3934/math.2022094
    [9] Li Chen, Dmitry V. Dolgy, Taekyun Kim, Dae San Kim . Probabilistic type 2 Bernoulli and Euler polynomials. AIMS Mathematics, 2024, 9(6): 14312-14324. doi: 10.3934/math.2024696
    [10] J. Kayalvizhi, A. G. Vijaya Kumar, Ndolane Sene, Ali Akgül, Mustafa Inc, Hanaa Abu-Zinadah, S. Abdel-Khalek . An exact solution of heat and mass transfer analysis on hydrodynamic magneto nanofluid over an infinite inclined plate using Caputo fractional derivative model. AIMS Mathematics, 2023, 8(2): 3542-3560. doi: 10.3934/math.2023180
  • In this paper, we presented Tauberian type results that intricately link the quasi-asymptotic behavior of both even and odd distributions to the corresponding asymptotic properties of their fractional Fourier cosine and sine transforms. We also obtained a structural theorem of Abelian type for the quasi-asymptotic boundedness of even (resp. odd) distributions with respect to their fractional Fourier cosine transform (FrFCT) (resp. fractional Fourier sine transform (FrFST)). In both cases, we quantified the scaling asymptotic properties of distributions by asymptotic comparisons with Karamata regularly varying functions.



    The fractional Fourier transform (FrFT), a broader version of the traditional Fourier transform (FT), was introduced seven decades ago by Namias [13]. However, it has only recently gained traction in fields such as signal processing, optics, and quantum mechanics [3,11,26]. Notably, it can be employed with real-world data such as one-dimensional signals (e.g., electrocardiogram or ECG data) and two-dimensional signals like geographical information (e.g., satellite images) [10]. Many scientists defined FrFT differently and enriched its theory properties [14]. Specially, in [7], FrFT of real order has been introduced using the Mittag-Leffler function. This transform plays the same role for the fractional derivatives as FT plays for the ordinary derivatives and is reduced into the FT, particularly for α=1 in the usual sense. FrFT is parameterized by α and effectively rotates a signal by an angle α within the time-frequency plane. Offering versatility, the FrFT facilitates the transformation of functions into various intermediate domains lying between time and frequency.

    Alongside the FT, cosine (CT) and sine (ST) transforms play crucial roles in signal processing by expanding functions over cosine and sine basis functions, despite some differences compared to the FT. The idea of fractionalization of the CT and ST was proposed in [9]. There, authors selected the real and imaginary components of the FrFT kernel to function as the kernels for the fractional Fourier cosine transform (FrFCT) and the fractional Fourier sine transform (FrFST), respectively. However, they acknowledged that their fractional transforms lack index additivity and do not qualify as genuine fractional versions of CT and ST. In [2], FrFCT and FrFST that are additive on the index and preserve the similar relationships with the fractional FT were introduced. Their fractional FT corresponds to a rotation of the Wigner distribution and the ambiguity function. Subsequently, in [16] the discrete version of the FrFCT and FrFST based on the eigen decomposition of discrete CT and discrete ST kernels are defined. Since then, FrFST and FrFSTs have significantly evolved, expanded to spaces of generalized functions [18], and became powerful tools in mathematical analysis, physics, signal processing, etc. [1,20,21].

    Distribution theory is a power tool in applied mathematics and the extension of integral transforms to generalized function spaces is an important subject, especially recently, when new transforms have been discovered and their connection with the old ones has been established. When studying the theory of distributions, one quickly learns that distributions do not have point values, which naturally imposes the idea of incorporating the asymptotic analysis to the field of generalized functions. The concept of quasi-asymptotics, as introduced in [25], serves to extend classical asymptotic methods within the framework of Schwartz distributions, finding applications across various fields, particularly in mathematical physics. There are many results where the asymptotic behavior of distributions is analyzed true to the behavior of various integral transforms in a form of the Abel and Tauberian type theorems, see [8,19,22,23] and references therein.

    The primary objective of this paper is to employ the FrFCT (FrFST) for a thorough investigation into the quasi-asymptotic properties of even (odd) distributions. Our study is structured around several theorems of the Abelian and Tauberian type, utilizing the asymptotic behavior of the FrFCT (FrFST) to investigate the quasi-asymptotic behavior of even (odd) distributions. The first section introduces the spaces of even (odd) tempered distributions. Subsequently, the FrFCT and FrFST are introduced through the FrFT and the Fourier cosine (sine) transform, establishing connections between all three transforms. The main result, presented in Section 2 with Theorem 2.3, asserts that if an even (odd) tempered distribution exhibits quasi-asymptotics at zero, then its FrFCT (FrFST) quasi-asymptotically oscillates at infinity. With an additional boundedness assumption, we demonstrate the converse. The second result, presented in Theorem 2.10, establishes the boundedness of the FrFCT (FrFST) of a distribution, assuming the distribution is quasi-asymptotically bounded.

    We employ the standard notation S(R) to denote the space of rapidly decreasing smooth functions f satisfying the condition:

    ρ(f)k=suptR,pk(1+|t|2)k/2|f(p)(t)|<,kN0=N0. (1.1)

    The dual space of S(R) is the space of tempered distributions, denoted by S(R). The subsets Se(R) and So(R) consist of all even and odd functions, respectively, within S(R).

    An example of an element in Se(R) is given by ex2, as ex2S(R) and is even. Conversely, xex2 belongs to So(R) since, according to [12], if φ is an even differentiable function, then Cφ is odd for any constant C.

    An even (odd) tempered distribution is defined as a continuous linear functional on the vector space Se(R) (So(R)). The spaces of such distributions are denoted as Se(R) and So(R), respectively. It is noteworthy that Se(R) and So(R) constitute broader classes than S(R), and specifically, Se(R)S(R) and So(R)S(R). Moreover,

    Se(R)So(R)=S(R).

    The FT of a function fS(R) is defined:

    ˆf(ξ)=F(f)(ξ)=12πRf(x)eixξdx. (1.2)

    The map fˆf is a continuous bijection from S(R) to S(R), and can be extended by duality to S(R).

    For an even function f, the Fourier cosine transform (FCT) is given by

    ˆfc(ξ)=Fc(f)(ξ)=2π0f(t)cos(tξ)dt,  ξR,

    and for an odd function f, the Fourier sine transform (FST) is given by

    ˆfs(ξ)=Fs(f)(ξ)=2π0f(t)sin(tξ)dt,  ξR.

    If ϕSe(R)(So(R)), then Fc(ϕ)Se(R)(Fs(ϕ)So(R)). The mapping Fc(Fs) is a continuous isomorphism from Se(R)(So(R)) to Se(R)(So(R)) [5].

    The development of the FrFCT and FrFST of distributions follows the conventional approach employed in FTs. Consequently, it is necessary to highlight certain fundamental properties of the FrFCT and FrFST for functions (refer to [12,Chapter 3]).

    The FrFCT (FrFST) of an even (odd) tempered distribution f can be defined by

    Fcf,φ=f,Fcφ,(Fsf,φ=f,Fsφ)

    for arbitrary φSe(R)(So(R)). Thus, Fc(Fs) is a continuous mapping from Se(R)(So(R)) to Se(R)(So(R)).

    Let fL1(R). Recall from [13] that the FrFT of order α is defined by

    Fα(f(x))(ξ)={Rf(x)Kα(x,ξ)dx,ξR,αkπ,kN,   f(ξ),α=2kπ,    f(ξ),α=(2k+1)π, (1.3)

    where

    Kα(x,ξ)=Cαei(x2+ξ22axξb) (1.4)

    is the kernel of FrFT, a=cotα, b=cscα, and Cα=1ia2π. The kernel Kα(x,ξ) is a continuously differentiable function in both variables, x and ξ. The function Fαf exhibits 2π periodicity with respect to α, and, thus, we will consistently consider α within the interval [0,2π). Notice that when nZ, Fnπ/2f=Fnf, where Fn is the n-th power of the FT (1.2), i.e., Fα is the sth power of the FT for s=2α/π, for α within the interval [0,2π), [13]. So, the FT is of order 1, while the identity operator is of order 0. Negative orders correspond to inverse transforms. For instance, applying the FrFT of order 1/2 twice yields the FT.

    In the research conducted by Pathak [15,Thrm. 3.1], it has been established that the FrFT constitutes a continuous mapping from the Schwartz space S(R) to itself. Furthermore, this mapping can be extended to the space of tempered distributions. Specifically, according to the definition provided in [15,Def. 3.1], the generalized FrFT, denoted as Fαf for fS(R), is expressed as follows:

    Fαf,φ=f,Fαφ,φS(R). (1.5)

    For fS(R), the inverse FrFT Fα is given with [3]

    f(x)=RFαf(ξ)Kα(x,ξ)dξ,xR. (1.6)

    From the linearity of the FrFT and the reversion property Fα(f(x))(ξ)=Fα(f(x))(ξ), we have

    Fα(f(x)±f(x))(ξ)=Fα(f(x))(ξ)±Fα(f(x))(ξ),

    and we conclude that the FrFT of an even function is even, while the FrFT of an odd function is odd.

    In [6], the FrFCT is denoted as Fcα(f(x))(ξ)=f(x)Kcα(x,ξ)dx, and the FrFST is represented as Fsα(f(x))(ξ)=f(x)Kcα(x,ξ)dx. It is important to observe that if the function f is odd, then Fcα(f(x))(ξ)=0. Similarly, if f is an even function, then the corresponding FrFCT simplifies to the one-sided FrFCT, given by:

    Fcα(f(x))(ξ)=0f(x)Kcα(x,ξ)dx.

    A similar consideration can also be repeated for the FrFST.

    Here, we follow the notion from [3,18], and the definitions from there. This means that if we restrict ourself to one-sided functions (f(x)=0 for x<0), we can define the FrFCT of a function fL1(R) as

    Fcα(f(x))(ξ)=Fα(f(x)+f(x))(ξ)=0f(x)Kcα(x,ξ)dx, (1.7)

    where

    Kcα(x,ξ)={Cαeix2+ξ22acos(xξb),αkπ,kZ=N0(N),   2πcos(xξ),α=π/2,    δ(xξ),α=kπ, (1.8)

    and a=cotα, b=cscα, and Cα=2(1ia)π.

    The inverse FrFCT is given by

    f(x)=0Fcαf(ξ)Kcα(x,ξ)dξ,xR. (1.9)

    Similarly, the FrFST of a function fL1(R) is defined as

    Fsα(f(x))(ξ)=Fα(f(x)f(x))(ξ)=0f(x)Ksα(x,ξ)dx, (1.10)

    where

    Ksα(x,ξ)={Cαei(απ/2)x2+ξ22asin(xξb),αkπ,kZ=N0(N),   2πsin(xξ),α=π/2,    δ(xξ),α=kπ, (1.11)

    with the same constants as above. The corresponding inverse FrFST is given by

    f(x)=0Fsαf(ξ)Ksα(x,ξ)dξ,xR. (1.12)

    The kernel Kcα(x,ξ)(Ksα(x,ξ)) is a continuously differentiable function in both variables, x and ξ. For α=(2k1)π2, the FrFCT and FrFST are reduced to the FCT and FST, respectively.

    To establish the connection between the FrFT of a causal, one-sided function f(x), where f(x)=0 for x<0, and the FrFCT and FrFST of this function in an alternative manner, we can formulate the following expression.

    2Fαf(±ξ)=Fcαf(ξ)±eiαFsαf(ξ).

    The expression Fcαf(ξ) can be associated with the even component of Fαf(ξ), while Fsαf(ξ) is linked to its odd component. In a broader context, it can be concluded that for determining the FrFCT of a causal, one-sided function f(x), one can equivalently determine the FrFT of the symmetrically extended two-sided function f(x)+f(x). Similarly, to ascertain the FrFST of such a function, one can alternatively determine the FrFT of the anti-symmetrically extended two-sided function ejα(f(x)f(x)). Both cases involve restricting the analysis to ξ0. Moreover, Fcαf(ξ) and Fαf(ξ) are periodic with period π [2].

    In [6], authors studied the FrFCT (resp., FrFST) for the space Se(R) (resp., So(R)). They demonstrated a Parseval-type relationship, an inversion formula, and the continuity properties of the FrFCT and FrFST. They, and also Prasad and Sihgh in [18]HY__HY, Thrm 3.1 and Thrm 3.2], have shown that the FrFCT is the continuous linear mapping of Se(R) onto itself, and that the FrFST is the continuous linear mapping of So(R) onto itself. This allows them to define the generalized FrFCT (FrFST) of distribution f from Se(R) (So(R)) by

    Fcαf,φ=f,Fcαφ,(Fsαf,φ=f,Fsαφ) (1.13)

    for all φSe(R)(So(R))). Moreover, the FrFCT (FrFST) of a distribution fSe(R)(So(R)) is a continuous linear map of Se(R)(So(R)) onto itself.

    We need the relation between the FrFCT and the FCT:

    Fcαf(ξ)=0f(x)Kcα(x,ξ)dx=Cαeiξ2a/20eix2a/2cos(xξb)f(x)dx (1.14)
    =1iaeiξ2a/2Fc(eix2a/2f(x))(ξb).

    Similarly, it holds for the FrFST and the FST.

    We will analyze the properties of a distribution by comparing it to regularly varying functions, specifically focusing on the quasi-asymptotic behavior outlined in [17,22,23]. A real-valued function, measurable, defined, and positive within an interval (0,D] (or [D,)), where D>0, is termed "a slowly varying function" at the origin (or at infinity) if it satisfies:

    limε0+L(dε)L(ε)=1( resp.limhL(dh)L(h)=1)for each d>0. (2.1)

    If L is a slowly varying function at zero, then ˜L()=L(1/) is also a slowly varying function in a neighborhood of , and vice versa.

    Considering L as a function exhibiting slow variation at the origin, it is relevant to recall the definition from [4] that a distribution fS(R) manifests quasi-asymptotic behavior, or quasi-asymptotics, of degree mR at the point x0R concerning L. This characterization is established if there exists uS(R) such that, for every φS(R), the following condition is satisfied:

    limε0+f(x0+εx)εmL(ε), φ(x)=u(x),φ(x). (2.2)

    The quasi-asymptotic behavior is conveniently denoted as:

    f(x0+εx)εmL(ε)u(x)asε0+inS(R),

    and should be consistently interpreted within the framework of the weak topology S(R), i.e., in the sense of (2.2).

    The form of u is not arbitrary; it must exhibit homogeneity with a degree of homogeneity m, i.e., u(dx)=dmu(x), for all d>0 [17,25]. Additionally, if (2.2) holds for each φD(R), it must also hold for each φS(R) [22,Thrm. 6.1]. Consequently, the quasi-asymptotic behavior at finite points is considered a local property. The quasi-asymptotics of distributions at infinity concerning a slowly varying function L at infinity is similarly defined. The notation f(hx)hmL(h)u(x) as h in S(R) is employed in this case.

    Furthermore, we may explore quasi-asymptotics within alternative distribution spaces. The relationship f(x0+εx)εmL(ε)u(x) as ε0+ in B(R) implies that (2.2) holds for each φB(R). Similarly, the quasi-asymptotics at infinity in B(R) satisfy this condition, where B(R) denotes any space of test functions on R, and B(R) represents their dual.

    The next lemma, as presented in [4,Lemma 3.1], establishes a connection between quasi-asymptotic behavior at the point and the corresponding oscillation at that same point.

    Lemma 2.1. If

    f(εx)/(εmL(ε)),φ(x),converges as  ε0+,φS, (2.3)

    then,

    eiC(εx)2/2f(εx)/(εmL(ε)),φ(x), converges as  ε0+,φS, (2.4)

    where CR and ε(0,1). On the contrary, if condition (2.4) is satisfied and there exists ε0(0,ε) such that the family

    {f(εx)/(εmL(ε)):ε(0,ε0)} is bounded in S(R), (2.5)

    then (2.4)(2.3).

    Moreover, (2.3) is equivalent to

    eih2ξ22Cˆf(hξ)/(hm2˜L(h)),γ(ξ),converges as h,γS, (2.6)

    where h>h0>0.

    Theorem 2.2. Let fSe(R)(So(R)). The statements below are equivalent:

    f(x/ε)εmL(ε)u(x)asεinSe(R+)(So(R)), (2.7)

    and

    ˆfc(hξ)hm1˜L(h)ˆuc(ξ)ashinSe(R+),(ˆfs(hξ)hm1˜L(h)ˆus(ξ)ashinSo(R+)), (2.8)

    where ˜L(h) is slowly varying at , (˜L()=L(1)).

    Proof. Let φSe(R). The relation obtained by using the Parseval identity

    ˆfc(hξ)hm1˜L(h),φ(ξ)=ˆfc(ξ)hm˜L(h),φ(ξh)=f(x)hm˜L(h),ˆφc(xh)=f(xh)hm˜L(h),ˆφc(x),

    immediately implies the assertion.

    The following theorem asserts that if fSe(R)(fSo(R)) has quasi-asymptotics at zero, then Fcαf(Fsαf) quasi-asymptotically oscillates as it approaches infinity. The reverse is true, with the additional assumption (2.5).

    Theorem 2.3. Let fSe(R)(So(R)), L be of the form (2.1) at 0+, uSe(R)(So(R)) be a homogeneous function, and mR. If

    f(εx)εmL(ε)u(x)   as   ε0+   in  Se(R)(So(R)), (2.9)

    then

    eia(hξ)2/2Fcαf(hξ)1iabm+1hm1˜L(h)ˆuc(ξ)   as   h,(eia(hξ)2/2Fsαf(hξ)1iabm+1hm1˜L(h)ˆus(ξ)   as   h), (2.10)

    in Se(R)(So(R)). Conversely, if (2.5) holds, then (2.10) (2.9).

    Proof. The notation with 1/ε will be maintained in the proof instead of using h. We will present the proof for the FrFCT, since the proof for the FrFST is analogous.

    Let φSe(R). Using (1.13) and (1.14), we obtain:

    eia(ξε)2/2Fcαf(ξε),φ(ξ)=εeiaξ2/2(Fcαf)(ξ),φ(εξ)
    =εCα2πbFc(eiax2/2f(x))(ξ),φ(εξb)=1iaeiax2/2f(x),ˆφc(xbε)
    =1iaεbeia(εtb)2/2f(εtb),ˆφc(t).

    The essential relation is the next one (see [4]):

    1εm+1L(ε)eia(ξε)2/2Fcαf(ξε),φ(ξ)=L(εb)L(ε)1iabm+1(εb)mL(εb)eia(εtb)2/2f(εtb),ˆφc(t). (2.11)

    From (2.1), we obtain:

    limε0+eia(ξε)2/2Fcαf(ξε)εm+1L(ε),φ(ξ)=
    =1iabm+1limε0+1(εb)mL(εb)eia(εtb)2/2f(εtb),ˆφc(t)
    =1iabm+1(limε0+1(εb)mL(εb)f(εtb),ˆφc(t)+limε0+1(εb)mL(εb)f(εtb),(eia(εtb)2/21)ˆφc(t)).

    The second limit is zero because

    limε0+(eia(εtb)2/21)ˆφc(t)=0,

    and {f(εtb)(εb)mL(εb):ε(0,ε0)} is bounded in Se(R).

    Now, by (2.9) and the Banach Steinhaus theorem, we have

    limε0+eia(ξε)2/2Fcαf(ξε)εm+1L(ε),φ(ξ)=1iabm+1limε0+f(εt)εmL(ε),ˆφc(t)
    =1iabm+1u(x),ˆφc(x)=1iabm+1ˆuc(ξ),φ(ξ).

    This proves (2.9)(2.10).

    For the opposite implication, we assume that (2.10) holds and we use (2.11).

    limε0+1εm+1L(ε)f(εx),φ(x)
    =limε0+1εm+1L(ε)eia(xε)2/2f(εx),φ(x)limε0+1εm+1L(ε)(eia(xε)2/21)f(εx),φ(x).

    Now, by the boundedness condition (2.5), we have that

    1εm+1L(ε)(eia(xε)2/21)f(εx),φ(x)0,asε0+,φSe.

    For the last part, we note that for any p>0:

    Fc(eipx22)(ξ)=eiπ4peiξ22p,ξR+.

    Applying the FT to Eq (2.3), we find that there exists a suitable constant ˜CC such that:

    Fc(eiC(εx)2)(ξ)Fc(f(εx))(ξ)/(εmL(ε)),(ˆφc)(ξ)
    =˜C1εm+2L(ε)eiξ2/(2ε2C)Fc(f)(ξ/ε),(ˆφc)(ξ).

    If we put h=1/ε and γ=ˆφc, we obtain (2.6). The implication (2.6)(2.4) follows in the same way. The result is now obvious by (2.11).

    Remark 2.4. Now, we take tanβ=ε2tanα, and ε>0, with Cα,β defined as:

    Cα,β(ξ)=cosβcosαeiα2eiβ2exp(iξ22cotα(1cos2βcos2α)).

    We have the following scalar property for the FrFCT (FrFST):

    Fcα(f(εx))(ξ)=Cα,β(ξ)Fcβ(f(x))(sξε), (2.12)

    and

    Fsα(f(εx))(ξ)=Cα,β(ξ)eiβFsβ(f(x))(sξε), (2.13)

    where s=sinβsinα, [2], then, we can obtain the same result with very few modifications.

    Example 2.5. It is known that every bounded function in R defines a tempered distribution, so the function f(x)=H(x)(2+sin1x)S(R), where H(x) is Heviside's function. It is known that it does not have regular asymptotics at 0, but f(εx)2asε0+ (m=0, L(ϵ)=1 and u(x)=2). Since ˆus(ξ)=2ξ2π, by Theorem 2.3 we have

    eia(hξ)2/2Fsαf(hξ)1hξ22(1ia)πb  as   h

    with respect to ˜L(ξ)1.

    Example 2.6. For f(x)=sinxS0(R), it is known that f(εx)ε1δ(x)asε0+, for m=1,L(ε)1,u(x)=δ(x), then by ˆus(ξ)=ˆδs(ξ)=0 and Theorem 2.3, we have

    eia(hξ)2/2Fsαf(hξ)0   as   h

    with respect to ˜L(ξ)1.

    Example 2.7. Let f(x)=2πδ(x1), so it follows that f(εx)ε2πδ(x1) as ε0+, then by ˆus(ξ)=Fs(2πδ(ξ1))=2πeiξ and Theorem 2.3, we have

    eia(hξ)2/2Fsαf(hξ)2π(1ia)b2h2eiξ  as   h

    with respect to ˜L(ξ)1.

    We require an additional concept from quasi-asymptotic analysis, specifically the idea of quasi-asymptotic boundedness as in [24]. Recall from [24] that the distribution fS(R) is quasi-asymptotically bounded at x0R of degree mR, in relation to the slowly varying function L at the origin if

    f(x0+εx)=O(εmL(ε))asε0+inS(R).

    The above relation should be interpreted in the sense of the weak topology of S(R), namely,

    f(x0+εx),φ(x)=O(εmL(ε))asε 0+, (2.14)

    for every φS(R). This means there exists a constant C>0 such that

    |f(x0+εx),φ(x)|C|εmL(ε)|,

    for x sufficiently close to x0.

    The notion of quasi-asymptotic boundedness, with respect to a degree mR in correlation with the slowly varying function at infinity L, is explicated in an analogous manner. Concurrently, our examination extends to the contemplation of quasi-asymptotic boundedness within a designated mathematical space: Se(R), φSe(R) (So(R), φSo(R)).

    Please note that L exhibits slow variation at the origin if, and only if, there exist measurable functions u and w defined on an interval (0,A], where u is bounded and possesses a finite limit at 0. Additionally, ω is continuous on [0,A] with ω(0)=0, such that the following representation holds for L(x) within the interval (0,A]:

    L(x)=exp(u(x)+Axω(ξ)ξdξ),x(0,A].

    In the context of our exploration for suitable modifications of L concerning quasi-asymptotics, it is reasonable to assume that L is defined across the entire interval (0,) and maintains nonnegativity or even positivity throughout. The analysis entails extending the functions u and ω to (0,) through a chosen method.

    For example, when addressing functions with slow variation at the origin, the condition ξ1ω(ξ)L1([1,)) implies the existence of positive constants ˜C and C such that the following inequalities hold for x>1:

    ˜C<L(x)<C.

    Remark 2.8. It is readily apparent that when f belongs to the space Sc(R) (f belonging to So(R)) and exhibits quasi-asymptotic boundedness at zero concerning εmL(ε), then eiC(x)2f(x) also demonstrates quasi-asymptotic boundedness at zero, considering the same slowly varying function. Here, x is a real number, and the condition |eiC(x)2|=1 holds.

    Remark 2.9. It is clear that for φSe(R), it follows that xpφSe(R),pN, and

    ˆφ(p)c(t)={(1)p/2Fc(xpφ(x))(t),pis even; (1)(p1)/2Fc(xpφ(x))(t),pis odd.

    Similar remark holds for ˆφs(t) in So(R).

    We have the following result.

    Theorem 2.10. Let fSc(R) (fSo(R)) be a quasi-asymptotically bounded at zero with respect to a slowly varying function at infinity L, that is,

    |f(εx),φ(x)|CφεmL(ε),ε0,

    where φSc(R) (φSo(R)), and Cφ>0 depends of φ, then the FrFCT (FrFST) for f is a bounded function at 0, i.e., there exists a constant C>0 such that

    |eia(ξε)2/2Fcαf(ξε),φ(ξ)|εm+1L(ε)C(|eia(ξε)2/2Fsαf(ξε),φ(ξ)|εm+1L(ε)C).

    Proof. Let f be a quasi-asymptotically bounded at zero with respect to L. From (2.11) and Remarks 2.8 and 2.9, we have that there exist kN0 and M>0 such that

    1εm+1L(ε)|eia(ξε)2/2Fcαf(ξε),φ(ξ)|=|1iabL(ε)εmeia(εtb)2/2f(εtb),ˆφc(t)|
    MCαˆφc(t)k<≤MCαsupxR,pk(1+|t|2)k/2|ˆφ(p)c(t)|<.

    In conclusion, in this research we extended recent inquiries into the analysis of integral transforms within specific distributional spaces. Our approach integrates the concept of quasi-asymptotic behavior, as introduced by Zavialov in [27], and we quantify the scaling asymptotic properties of distributions by asymptotic comparisons with Karamata regularly varying functions. In this paper, we characterized the quasi-asymptotic behavior of even (resp., odd) distributions within the context of a Tauberian theorem applied to the FrFCT (resp., FrFST), and by Thrm. 2.3, we established that distributions exhibiting quasi-asymptotic behavior at zero manifest quasi-asymptotic oscillations at infinity through their corresponding FrFCT or FrFST. Additionally, our second result presented by Thrm. 2.10 sheds light on the boundedness of these transforms concerning quasi-asymptotically bounded distributions.

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

    The authors S. Haque and N. Mlaiki would like to thank Prince Sultan University for paying the publication fees for this work through TAS LAB.

    The authors declare no conflict of interest.



    [1] N. Abbas, S. Nadeem, A. Sial, W. Shatanawi, Numerical analysis of generalized Fourier's and Fick's laws for micropolar Carreaufluid over a vertical stretching Riga sheet, Zamm Z. Angew. Math. Me., 103 (2023), e202100311. https://doi.org/10.1002/zamm.202100311 doi: 10.1002/zamm.202100311
    [2] T. Alieva, M. J. Bastiaans, Fractional cosine and sine transforms in relation to the fractional Fourier and Hartley transforms, Seventh International Symposium on Signal Processing and Its Applications, Paris, France, 2003,561–564. https://doi.org/10.1109/ISSPA.2003.1224765
    [3] L. B. Almeida, The fractional Fourier transform and time-frequency representations, IEEE T. Signal Proces., 42 (1994), 3084–3091. https://doi.org/10.1109/78.330368 doi: 10.1109/78.330368
    [4] S. Atanasova, S. Maksimović, S. Pilipović, Abelian and Tauberian results for the fractional Fourier and short-time Fourier transforms of distributions, Integr. Transf. Spec. F., 35 (2024), 1–16. https://doi.org/10.1080/10652469.2023.2255367 doi: 10.1080/10652469.2023.2255367
    [5] P. K. Banerji, S. K. Al-Omari, L. Debnath, Tempered distributional sine (cosine) transform, Integr. Transf. Spec. F., 17 (2006), 759–768. https://doi.org/10.1080/10652460600856534 doi: 10.1080/10652460600856534
    [6] H. Hudzik, P. Jain, R. Kumar, On generalized fractional cosine and sine transforms, Georgian Math. J., 25 (2018), 259–270. https://doi.org/10.1515/gmj-2018-0030 doi: 10.1515/gmj-2018-0030
    [7] G. Jumarie, Fourier's transform of fractional order via Mittag-Leffler function and modified Riemann-Liouville derivative, J. Appl. Math. Inform., 26 (2008), 1101–1121. https://doi.org/10.1016/j.aml.2009.05.011 doi: 10.1016/j.aml.2009.05.011
    [8] S. Kostadinova, S. Pilipović, K. Saneva, J. Vindas, The ridglet transform of distribution and quasi-asymptotic behavior of distributions, In: Pseudo-differential operators and generalized functions, Cham: Birkhäuser, 2015,185–197. https://doi.org/10.1007/978-3-319-14618-8_13
    [9] A. W. Lohmann, D. Mendlovic, Z. Zalevsky, R. G. Dorsch, Some important fractional for signal processing, Opt. Commun., 125 (1996), 18–20. https://doi.org/10.1016/0030-4018(95)00748-2 doi: 10.1016/0030-4018(95)00748-2
    [10] A. R. Maddukuri, Fractional Fourier transform and scaling problem in signals and images, Master Thesis, DITE, Sweden, 2018.
    [11] A. C. McBride, F. H. Kerr, On Namias's fractional Fourier trans-forms, IMA J. Appl. Math., 39 (1987), 159–175. https://doi.org/10.1093/imamat/39.2.159 doi: 10.1093/imamat/39.2.159
    [12] E. O. Milton, Fourier transforms of odd and even tempered distributions, Pac. J. Math., 50 (1974), 563–572. http://doi.org/10.2140/pjm.1974.50.563 doi: 10.2140/pjm.1974.50.563
    [13] V. Namias, The fractional order Fourier transform and its application to quantum mechanics, IMA J. Appl. Math., 25 (1980), 241–265. https://doi.org/10.1093/imamat/25.3.241 doi: 10.1093/imamat/25.3.241
    [14] H. M. Ozaktas, Z. Zalevsky, M. A. Kutay, The fractional Fourier transform with applications in optics and signal processing, New York: Wiley, 2001.
    [15] R. S. Pathak, A. Prasad, M. Kumar, Fractional Fourier transform of tempered distributions and generalized pseudo differential operator, J. Pseudo-Differ. Oper. Appl., 3 (2012), 239–254. https://doi.org/10.1007/s11868-012-0047-8 doi: 10.1007/s11868-012-0047-8
    [16] S. C. Pei, M. H. Yeh, Thediscrete fractional cosine and sine transforms, IEEE T. Signal Proces., 49 (2001), 1198–1207. https://doi.org/10.1109/78.923302 doi: 10.1109/78.923302
    [17] S. Pilipović, B. Stanković, A. Takači, Asymptotic behavior and Stieltjes transformation of distribution, Teubner-Texte zur Mathematik 11, Leipzig, 1990.
    [18] A. Prasad, M. K. Singh, Pseudo-differential operators involving fractional Fourier cosine (sine) transform, Filomat, 31 (2017), 1791–1801. https://doi.org/10.2298/FIL1706791P doi: 10.2298/FIL1706791P
    [19] S. Pilipović, J. Vindas, Multidimensional Tauberian theorems for vector-valued distributions, Publications de l'Institut Mathematique, 95 (2014), 1–28. https://doi.org/10.2298/PIM1409001P doi: 10.2298/PIM1409001P
    [20] F. A. Shah, W. Z. Lone, K. S. Nisar, T. Abdeljawad, On the class of uncertainty inequalities for the coupled fractional Fourier transform, J. Inequal. Appl., 2022 (2022), 133. https://doi.org/10.1186/s13660-022-02873-2 doi: 10.1186/s13660-022-02873-2
    [21] W. Shatanawi, N. Abbas, T. A. M. Shatnawi, F. Hasan, Heat and mass transfer of generalized fourier and Fick's law for second-grade fluid flow at slendering vertical Riga sheet, Heliyon, 9 (2023), E14250, https://doi.org/10.1016/j.heliyon.2023.e14250 doi: 10.1016/j.heliyon.2023.e14250
    [22] J. Vindas, S. Pilipović, Structural theorems for quasi-asymptotics of distributions at the origin, Math. Nachr., 282 (2009), 1584–1599. https://doi.org/10.1002/mana.200710090 doi: 10.1002/mana.200710090
    [23] J. Vindas, Structural theorems for quasi-asymptotics of distributions at infinity, Publications de l'Institut Mathematique, 84 (2008), 159–174. https://doi.org/10.2298/PIM0898159V doi: 10.2298/PIM0898159V
    [24] J. Vindas, The structure of quasi-asymptotics of Schwartz distributions, Banach Center Publications, 88 (2010), 297–314. https://doi.org/10.4064/bc88-0-24 doi: 10.4064/bc88-0-24
    [25] V. S. Vladimirov, Y. N. Drozinov, B. I. Zavialov, Tauberian theorems for generalized functions, Dordrecht: Springer, 1988. https://doi.org/10.1007/978-94-009-2831-2
    [26] N. Wiener, Hermitian polynomials and Fourier analysis, J. Math. Phys., 8 (1929), 70–73. https://doi.org/10.1002/sapm19298170 doi: 10.1002/sapm19298170
    [27] B. I. Zavialov, "Quasi-asymptotic" behavior of generalized functions, and self-simulation of electromagnetic form factors, Theor. Math. Phys., 19 (1974), 426–432. https://doi.org/10.1007/BF01035942 doi: 10.1007/BF01035942
  • This article has been cited by:

    1. Yannan Sun, Wenchao Qian, Fast algorithms for nonuniform Chirp-Fourier transform, 2024, 9, 2473-6988, 18968, 10.3934/math.2024923
  • Reader Comments
  • © 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1061) PDF downloads(44) Cited by(1)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog