Open access peer-reviewed chapter

Generalized Bessel Operator and Reproducing Kernel Theory

Written By

Fethi Soltani

Reviewed: 31 May 2022 Published: 24 May 2023

DOI: 10.5772/intechopen.105630

From the Edited Volume

Operator Theory - Recent Advances, New Perspectives and Applications

Edited by Abdo Abou Jaoudé

Chapter metrics overview

83 Chapter Downloads

View Full Metrics

Abstract

In 1961, Bargmann introduced the classical Fock space FC and in 1984, Cholewinsky introduced the generalized Fock space F2,νC. These two spaces are the aim of many works, and have many applications in mathematics, in physics and in quantum mechanics. In this work, we introduce and study the Fock space F3,νC associated to the generalized Bessel operator L3,ν. The space F3,νC is a reproducing kernel Hilbert space (RKHS). This is the reason for defining the orthogonal projection operator, the Toeplitz operators and the Hankel operators associated to this space. Furthermore, we give an application of the theory of extremal function and reproducing kernel of Hilbert space, to establish the extremal function associated to a bounded linear operator T:F3,νC→H, where H be a Hilbert space. Finally, we come up with some results regarding the extremal functions, when T is a difference operator and an integral operator, respectively. Finally, we remark that it is now natural to raise the problem of studying the Bessel-type Segal-Bargmann transform associated to the space F3,νC. This problem is difficult and will be an open topic. This topic requires more details for the harmonic analysis associated to the operator L3,ν. We have the idea to continue this research in a future paper.

Keywords

  • Bessel-type Fock space
  • Heisenberg-type uncertainty principle
  • Toeplitz operators
  • Hankel operators
  • Tikhonov regularization problem
  • extremal function

1. Introduction

In [1], Bargmann has studied the Fock space FC, is a Hilbert space consisting of entire functions on C, square integrable with respect to the measure

dmz1πez2dxdy,z=x+iy.

This space is equipped with the inner product

fgFCCfzgz¯dmz,

and has the reproducing kernel kzw=ez¯w. In [2], Cholewinsky has constructed a generalized Fock space F2,νC consisting of even entire functions on C, square integrable with respect to the measure

dm2,νzz2ν+1Kν1/2z2π2ν1/2Γν+1/2dz,

where Kν, ν>0, is the modified Bessel function of the second kind and index ν, called also the Macdonald function [3]. The generalized Fock space F2,νC is associated to the Bessel operator

L2,νd2dz2+2νzddz,

and has the reproducing kernel

k2,νzw=I2,νz¯w=n=0z¯w2nαn2ν,

where

αn2ν=22nn!Γn+ν+1/2Γν+1/2.

The study of several generalizations of the Fock spaces has a long and rich history in many different settings [4, 5, 6, 7, 8]. In this work, we will try to generalize Bessel-type Fock space, to give some properties concerning Toeplitz operators and Hankel operators of this space; and to establish Heisenberg-type uncertainty principle for this generalized Fock space. The generalized Bessel operator (or hyper-Bessel operator [9]) is the third-order singular differential operator given by

L3,νd3dz3+3νzd2dz23νz2ddz,

where ν is a nonnegative real number. When ν=0 this operator becomes the third derivative operator for which some analysis was studied by Widder [10] and for some special value of ν the operator L3,ν appeared as a radial part of the generalized Airy equation of a nonlinear diffusion type partial differential equation in d. Recently, in a nice and long paper, Cholewinski and Reneke [11] studied and extended, for the operator L3,ν, the well known theory related to some singular differential operator of second order for which the literature is extensive. Next, Fitouhi et al. [12, 13] established a harmonic analysis related to this operator (for examples the eigenfunctions, the generalized translation, the Fourier-Airy transform, the heat equation, the heat polynomials, the transmutation operators, …). Recently the Airy operator has gained considerable interest in various field of mathematics [9, 14] and in certain parts of quantum mechanics [15]. The results of this work will be useful when discussing the Fock space associated to this operator. This space is the background of some applications in this contribution. Especially, we give an application of the theory of extremal functions and reproducing kernels of Hilbert spaces, to examine the extremal function for the Tikhonov regularization problem associated to a bounded linear operator T:F3,νCH, where H be a Hilbert space. We come up with some results regarding the extremal functions associated to a difference operator D and to an integral operator P.

The examination of the extremal functions is studied in several directions, in the Fourier analysis [16, 17], in the Sturm-Liouville hypergroups [18], and in the Fourier-Dunkl analysis [19, 20]…

The contents of the paper are as follows. In Section 2, we study the Toeplitz operators and the Hankel operators on the Bessel-type Fock space F3,νC, and we establish Heisenberg-type uncertainty principle for this space. In Section 3, we give an application of the theory of reproducing kernels to the Tikhonov regularization problem for a difference operator and for an integral operator, respectively. In the last section, we summarize the obtained results and describe the future work.

Advertisement

2. Bessel-type Fock space

In this section we introduce the Toeplitz and the Hankel operators on the Bessel-type Fock space F3,νC. And we establish an uncertainty inequality of Heisenberg-type on the space F3,νC.

2.1 Toeplitz and Hankel operators on F3,νC

Let zC and ωk=e2k13, k=1,2,3. A function uz is called 3-even if uωkz=uz.

For λC, the initial problem

L3,νuz=λ3uz,u0=1,uk0=0,k=1,2

admits a unique analytic solution on C (see [11]), which will be denoted by I3,νλz and expanded in a power series as

I3,νλz=n=0λz3nαn3ν,E1

where

αn3ν=33nn!Γn+1/3Γn+ν+2/3Γ1/3Γν+2/3.

The function I3,νλz is 3-even and defined as the hypergeometric function [11],

I3,νλz=0F213ν+23λz33.

In particular, I3,νλzeλz and I3,0λz=cos3λz=n=0λz3n3n!.

In the following we denote by

  • Oν the function (see [11], p. 12) defined for t0 by

Oνt2t1/333ν+5/2Γ1/3Γν+2/30xν1/2et/xKν1/34x27dx,

where Kν is the Macdonald function.

  • dm3,νz, ν>0, the measure defined on C by

dm3,νz3πz2Oνz6dxdy,z=x+iy.

This weighted measure is use d by Cholewinski-Reneke in their interesting paper [11] for computing the ν-Airy heat function.

  • Lν2C, the space of measurable functions f on C satisfying

    fLν2C2Cfz2dm3,νz<.

  • H3,C, the space of 3-even entire functions on C.

Let ν>0. We define the Bessel-type Fock space F3,νC, to be the pre-Hilbert space of functions in H3,CLν2C, equipped with the inner product

fgF3,νCCfzgz¯dm3,νz,

and the norm fF3,νCfLν2C.

If f,gF3,νC with fz=n=0anz3n and gz=n=0bnz3n, then

fgF3,νC=n=0anbn¯αn3ν,E2

where αn3ν are the constants given by (1).

If fF3,νC, then fzez2/2fF3,νC, zC. The map ffz, zC, is a continuous linear functional on F3,νC. Thus from Riesz theorem [21], the space F3,νC has a reproducing kernel. The function k3,ν given for, by.

k3,νzw=I3,νz¯w,E3

is a reproducing kernel for the Bessel-type Fock space F3,νC, that is k3,νz.F3,νC, and for all fF3,νC, we have fk3,νz.F3,νC=fz.

The space F3,νC equipped with the inner product ..F3,νC is a reproducing kernel Hilbert space (RKHS); and the set z3nαn3νn forms a Hilbert’s basis for the space F3,νC.

In the next part of this section we study the Toeplitz operators and the Hankel operators on the Bessel-type Fock space F3,νC. These operators generalize the classical operators [5]. We consider the orthogonal projection operator Pν:Lν2CF3,νC defined for zC, by

Pνfz:=fk3,νz.Lν2C,

where k3,ν is the reproducing kernel given by (3). Then we have

PνPν=Pν,Pν=Pν,Pν=1,IPν1.

Let ϕLC. The multiplication operators Mϕ:Lν2CLν2C are the operators defined for zC, by

Mϕfzϕzfz.

The Bessel-type Toeplitz operators Tϕ:F3,νCF3,νC are the operators defined for zC, by

TϕfzPνMϕfz.

Let ϕLC. Then we have

Tϕϕ,Tϕ=Tϕ¯.

However, if ϕLC has compact support, then Tϕ is a compact operator.

Let ϕLC. The Bessel-type Hankel operators Hϕ:F3,νCLν2C are the operators defined for zC, by

HϕfzIPνMϕfz.

Let ϕ,φLC. Then we have

Hϕϕ,Hϕ=PνMϕ¯IPν,TϕφTϕTφ=Hϕ¯Hφ.

2.2 Heisenberg-type uncertainty principle for F3,νC

Let U3,νC be the prehilbertian space of 3-even entire functions, equipped with the inner product

fgU3,νCCfzgz¯z6dm3,νz.

If f,gU3,νC with fz=n=0anz3n and gz=n=0bnz3n, then

fgU3,νC=n=0anbn¯αn+13ν,fU3,νC2=n=0an2αn+13ν.

The space U3,νC is a Hilbert space with Hilbert’s basis z3nαn+13νn and reproducing kernel

J3,νzw=n=0z¯w3nαn+13ν=1z¯w3I3,νz¯w1.

Using the fact that

αn+13ν=3n+13n+13n+3ν+2αn3ναn3ν,n,E4

then the space U3,νC is a subspace of the Bessel-type Fock space F3,νC.

Let M be the multiplication operator defined by

Mfzz3fz.

Lemma 1. For fU3,νC then L3,νf and Mf belong to F3,νC. And for f,gU3,νC one has

L3,νfgF3,νC=fMgF3,νC.

Proof. Let fU3,νC with fz=n=0anz3n. By straightforward calculation we obtain

L3,νz3n=3n3n23n+3ν1z3n3,n.

Thus

L3,νfz=n=03n+13n+13n+3ν+2an+1z3n,E5

and by (4) we deduce that

L3,νfFνC2=n=13n3n23n+3ν1an2α3nνfU3,νC2.

On the other hand for Mf one has

Mfz=n=1an1z3n,

and

MfF3,νC2=n=1an12αn3ν=fU3,νC2.

Therefore L3,νf and Mf belong to F3,νC.

Let f,gU3,νC with fz=n=0anz3n and gz=n=0bnz3n. From relations (4) and (5) we have

L3,νfz=n=0an+1αn+13ναn3νz3n.

Therefore and according to (2) we obtain

L3,νfgF3,νC=n=0an+1bn¯αn+13ν=n=1anbn1¯α3nν=fMgF3,νC.

This completes the proof of the lemma. □

Let V3,νC be the prehilbertian space of 3-even entire functions, equipped with the inner product

fgV3,νCCfzgz¯z12dm3,νz.

If f,gV3,νC with fz=n=0anz3n and gz=n=0bnz3n, then

fgV3,νC=n=0anbn¯αn+23ν,fV3,νC2=n=0an2αn+23ν.

The space V3,νC is a Hilbert space with Hilbert’s basis z3nαn+23νn and reproducing kernel

K3,νzw=n=0z¯w3nαn+23ν=1z¯w6I3,νz¯wz¯w39ν+61.

The space V3,νC is a subspace of the space U3,νC.

Let L3,νM the commutator operator defined by

L3,νML3,νMML3,ν.

We easily have

Lemma 2. For fV3,νC we have L3,νMfF3,νC and

MfF3,νC2=L3,νfF3,νC2+L3,νMffF3,νC.

We will use the following result of functional analysis.

Lemma 3. (See [22, 23]). Let A and B be self-adjoint operators on a Hilbert space H (A=A, B=B). Then we have

AafHBbfH12ABffH,

for all fDomAB, and all a,b.

We obtain the following Heisenberg-type uncertainty principle.

Theorem 1. Let fV3,νC. For all a,b, we have.

L3,ν+MafF3,νCL3,νM+ibfF3,νCMfF3,νC2L3,νfF3,νC2.E6

Proof. Let us consider the following two operators on V3,νC by

A=L3,ν+M,B=iL3,νM.

By Lemmas 1 and 2, the operators A and B satisfies the following properties.

  1. For f,gV3,νC, we have

AfgF3,νC=fAgF3,νC,BfgF3,νC=fBgF3,νC.
V3,νCDomAB.
AB=2iL3,νM.

Thus the inequality (6) follows from Lemmas 2 and 3. □

Advertisement

3. Reproducing kernel theory

Let T:F3,νCH be a bounded linear operator from F3,νC into a Hilbert space H. By using the theory of reproducing kernels of Hilbert space and building on the ideas of Saitoh [24, 25, 26] we examine the extremal function associated to the operator T on the Bessel-type Fock space F3,νC.

Theorem 2. For any hH and for any λ>0, the problem

inffF3,νCλfF3,νC2+TfhH2E7

has a unique minimizer given by

fλ,Th=λI+TT1Th.E8

Proof. The problem (7) is solved elementarily by finding the roots of the first derivative dΦ of the quadratic and strictly convex function Φf=λfF3,νC2+TfhH2hH2. Note that for convex functions the equation dΦf=0 is a necessary and sufficient condition for the minimum at f. The calculation provides

dΦf=2λf+2TTfh,

and the assertion of the theorem follows at once. □

In this section we examine the extremal functions associated to a difference operator D; and to an integral operator P, respectively.

3.1 The difference operator

Let D be the difference operator defined by

Dfz1z3fzf0.

If fz=n=0anz3n, then Dfz=n=0an+1z3n.

In this subsection, we determine the extremal function fλ,D associated to the difference operator D on the space F3,νC.

Theorem 3.

  1. The operator D maps continuously from F3,νC into F3,νC, and

    DfF3,νCfF3,νC.

  2. For fF3,νC with fz=n=0anz3n, we have

    Dfz=n=1αn13ναn3νan1z3n,DDfz=n=1αn13ναn3νanz3n.

  3. For any hF3,νC and for any λ>0, the problem

    inffF3,νCλfF3,νC2+DfhF3,νC2

    has a unique minimizer given by

    fλ,Dhz=hΨzF3,νC,

    where

    Ψzw=n=0z¯3n+3w3nλαn+13ν+αn3ν,wC.

Proof.

  1. If fF3,νC with fz=n=0anz3n, then

    DfF3,νC2=n=0αn3νan+12n=1αn3νan2fF3,νC2.

  2. If f,gF3,νC with fz=n=0anz3n and gz=n=0bnz3n, then

    DfgF3,νC=n=0αn3νan+1bn¯=n=1αn13νanbn1¯=fDgF3,νC,

    where

    Dgz=n=1αn13ναn3νbn1z3n.

    And therefore

    DDfz=n=1αn13ναn3νanz3n.

  3. We put hz=n=0hnz3n and fλ,Dhz=n=0cnz3n. From (8) we have λI+DDfλ,Dhz=Dhz. By (ii) we deduce that

c0=0,cn=αn13νhn1λαn3ν+αn13ν,n.

Thus

fλ,Dhz=n=0αn3νhnz3n+3λαn+13ν+αn3ν=hΨzF3,νC,E9

where

Ψzw=n=0z¯3n+3w3nλαn+13ν+αn3ν.

This completes the proof of the theorem. □

The extremal function fλ,Dh possesses the following properties.

Theorem 4. If λ>0 and hF3,νC, then

fλ,Dhz12λI3,νz21/2hF3,νC,
fλ,DhF3,νC12λhF3,νC.

Proof. Let λ>0 and hF3,νC with hz=n=0hnz3n.

  1. From (9) we have

    fλ,DhzΨzF3,νChF3,νC.

    And by using the fact that x+y24xy we obtain

    ΨzF3,νC2=n=0αn3νz6n+6λαn+13ν+αn3ν214λn=0z6n+6αn+13ν14λI3,νz2.

    This proves (i).

  2. From (9) we have

fλ,DhF3,νC2=n=1αn3ναn13νhn1λαn3ν+αn13ν2.

Using the fact that x+y24xy we obtain

fλ,DhF3,νC214λn=1αn13νhn12=14λhF3,νC2.

This proves (ii) and completes the proof of the theorem. □

As in the same way of Theorem 4 we also obtain.

Remark 1. If λ>0 and hF3,νC, then

Dfλ,Dhz,fλ,DDhz12λI3,νz21/2fF3,νC.

The extremal function fλ,Dh possesses also the following approximation formulas.

Theorem 5. If λ>0 and hF3,νC, then

limλ0+Dfλ,DhhF3,νC2=0,

and

limλ0+Dfλ,Dhz=hz.

Proof. Let λ>0 and hF3,νC with hz=n=0hnz3n. From (9) we have

Dfλ,Dhzhz=n=0λαn+13νhnλαn+13ν+αn3νz3n.

Therefore

Dfλ,DhhF3,νC2=n=0αn3νλαn+13νhnλαn+13ν+αn3ν2

Using the fact that

αn3νλαn+13νhnλαn+13ν+αn3ν2αn3νhn2,

and

λαn+13νhnλαn+13ν+αn3νz3nhnz3n,

we obtain the results from dominated convergence theorem. □

As in the same way of Theorem 5 we also obtain.

Remark 2. If λ>0 and hF3,νC, then

limλ0+fλ,DDhhh0F3,νC2=0,

and

limλ0+fλ,DDhz=hzh0,

where h0z=h0.

3.2 The integral operator

Let P be the integral operator defined by

PfzfΦzF3,νC=CfwΦzw¯dm3,νw,

where

Φzw=n=0z¯3n+3w3nn+13αn3ν,wC.

If fz=n=0anz3n, then Pfz=n=1an1n3z3n.

In this subsection, we determine the extremal function fλ,P associated to the integral operator P on the space F3,νC.

Theorem 6.

  1. The operator P maps continuously from F3,νC into F3,νC, and

    PfF3,νC27ν+1fF3,νC.

  2. For fF3,νC with fz=n=0anz3n, we have

    Pfz=n=0αn+13νan+1n+13αn3νz3n,PPfz=n=0αn+13νann+13αn3νz3n.

  3. For any hF3,νC and for any λ>0, the problem

    inffF3,νCλfF3,νC2+PfhF3,νC2

has a unique minimizer given by

fλ,Phz=hΨzF3,νC,

where

Ψzw=n=1n3z¯3n1w3nλn3αn13ν+αn3ν,wC.

Proof.

  1. If fF3,νC with fz=n=0anzn, then

    PfF3,νC2=n=1αn3νn3an1227ν+1n=0αn3νan2=27ν+1fF3,νC2.

  2. If f,gF3,νC with fz=n=0anz3n and gz=n=0bnz3n, then

    PfgF3,νC=n=1αn3νan1n3bn¯=n=0αn+13νann+13bn+1¯=fPgF3,νC,

    where

    Pgz=n=0αn+13νbn+1n+13αn3νz3n.

    And therefore

    PPfz=n=0αn+13νann+13αn3νz3n.

  3. We put hz=n=0hnz3n and fλ,Phz=n=0cnz3n. From (8) we have λI+PPfλ,Phz=Phz. By (ii) we deduce that

    cn=n+13αn+13νhn+1λn+13αn3ν+αn+13ν,n.

Thus

fλ,Phz=n=1n3αn3νhnz3n3λn3αn13ν+αn3ν=hΨzF3,νC,E10

where

Ψzw=n=1n3z¯3n3w3nλn3αn13ν+αn3ν.

This completes the proof of the theorem. □

The extremal function fλ,Ph possesses the following properties.

Theorem 7. If λ>0 and hF3,νC, then

fλ,Phz12λI3,νz21/2hF3,νC,
fλ,PhF3,νC12λhF3,νC.

Proof. Let λ>0 and hF3,νC with hz=n=0hnz3n.

i. From (10) we have

fλ,PhzΨzF3,νChF3,νC.

And by using the fact that x+y24xy we obtain

ΨzF3,νC2=n=1n3αn3νz6n6λn3αn13ν+αn3ν214λn=0z6nαn3ν=14λI3,νz2.

This proves (i).

ii. From (10) we have

fλ,PhF3,νC2=n=0αn3νn+13αn+13νhn+1λn+13αn3ν+αn+13ν2.

Using the fact that x+y24xy we obtain

fλ,PhF3,νC214λn=0αn+13νhn+1214λhF3,νC2.

This proves (ii) and completes the proof of the theorem. □

As in the same way of Theorem 7 we also obtain.

Remark 3. If λ>0 and hF3,νC, then

Pfλ,Phz,fλ,PPhz332ν+1λI3,νz21/2hF3,νC.

The extremal function fλ,Ph possesses also the following approximation formulas.

Theorem 8. If λ>0 and hF3,νC, then

limλ0+Pfλ,Phhh0F3,νC2=0,

and

limλ0+Pfλ,Phz=hzh0.

Proof. Let λ>0 and hF3,νC with hz=n=0hnz3n. From (10) we have

Pfλ,Phzhzh0=n=1λn3αn13νhnλn3αn13ν+αn3νz3n.

Therefore

Pfλ,Phhh0F3,νC2=n=1αn3νλn3αn13νhnλn3αn13ν+αn3ν2

Using the fact that

αn3νλn3αn13νhnλn3αn13ν+αn3ν2αn3νhn2,

and

λn3αn13νhnλn3αn13ν+αn3νz3nhnz3n,

we obtain the results from dominated convergence theorem. □

As in the same of Theorem 8 we also obtain.

Remark 4. If λ>0 and hF3,νC, then

limλ0+fλ,PPhhF3,νC2=0,

and

limλ0+fλ,PPhz=hz.
Advertisement

4. Conclusion and perspectives

Bargmann [1] in 1961 introduced the classical Fock space FC and Cholewinsky [2] in 1984 introduced the generalized Fock space F2,νC. The Bessel-type Fock space F3,νC introduced in this work generalizes these analytic spaces. We are studied the Tikhonov regularization problem associated to this Hilbert space, and we are established the extremal function for this problem. Finally, in a future paper we have the idea to study the Bessel-type Segal-Bargmann transform, in which we will prove inversion formula, Plancherel formula and some uncertainty inequalities for this transform.

Advertisement

Conflicts of interest

The author declares that there is no conflict of interests regarding the publication of this paper.

References

  1. 1. Bargmann V. On a Hilbert space of analytic functions and an associated integral transform part I. Communications on Pure and Applied Mathematics. 1961;14:187-214
  2. 2. Cholewinski FM. Generalized Fock spaces and associated operators. SIAM Journal on Mathematical Analysis. 1984;15:177-202
  3. 3. Erdely A et al. Higher Transcendental Functions. Vol. 2. New York: McGraw-Hill; 1953
  4. 4. Baccar C, Omri S, Rachdi LT. Fock spaces connected with spherical mean operator and associated operators. Mediterranean Journal of Mathematics. 2009;6(1):1-25
  5. 5. Berger CA, Coburn LA. Toeplitz operators on the Segal-Bargmann space. Transactions of the American Mathematical Society. 1987;301:813-829
  6. 6. Nemri A, Soltani F, Abd-alla AN. New studies on the Fock space associated to the generalized Airy operator and applications. Complex Analysis and Operator Theory. 2018;12(7):1549-1566
  7. 7. Sifi M, Soltani F. Generalized Fock spaces and Weyl relations for the Dunkl kernel on the real line. Journal of Mathematical Analysis and Applications. 2002;270:92-106
  8. 8. Soltani F. Fock-type spaces associated to higher-order Bessel operator. Integral Transforms and Special Functions. 2018;29(7):514-526
  9. 9. Kiryakova V. Obrechkoff integral transform and hyper-Bessel operators via G-function and fractional calculus approach. Global Journal of Pure Applied Mathematics. 2005;1(3):321-341
  10. 10. Widder DV. The Airy transform. American Mathematical Monthly. 1979;86:271-277
  11. 11. Cholewinski FM, Reneke JA. The generalized Airy diffusion equation. Electronic Journal of Differential Equations. 2003;87:1-64
  12. 12. Fitouhi A, Mahmoud NH, Ahmed Mahmoud SA. Polynomial expansions for solutions of higher-order Bessel heat equations. Journal of Mathematical Analysis and Applications. 1997;206:155-167
  13. 13. Fitouhi A, Ben Hammouda MS, Binous W. On a third singular differential operator and transmutation. Far East Journal of Mathematical Sciences. 2006;21(3):303-329
  14. 14. Dimovski I, Kiryakova V. The Obrechkoff, integral transform: properties and relation to a generalized fractional calculus. Numerical Functional Analysis and Optimization. 2000;21(1–2):121-144
  15. 15. Ben Hammouda MS, Nemri A. Polynomial expansions for solutions of higher-order Besse heat equation in quantum calculus. Fractional Calculus and Applied Analysis. 2007;10(1):39-58
  16. 16. Matsuura T, Saitoh S, Trong D. Inversion formulas in heat conduction multidimensional spaces. Journal of Inverse and Ill-Posed Problems. 2005;13:479-493
  17. 17. Matsuura T, Saitoh S. Analytical and numerical inversion formulas in the Gaussian convolution by using the Paley-Wiener spaces. Applicable Analysis. 2006;85:901-915
  18. 18. Soltani F. Extremal functions on Sturm-Liouville hypergroups. Complex Analysis and Operator Theory. 2014;8(1):311-325
  19. 19. Soltani F. Extremal functions on Sobolev-Dunkl spaces. Integral Transforms and Special Functions. 2013;24(7):582-595
  20. 20. Soltani F. Multiplier operators and extremal functions related to the dual Dunkl-Sonine operator. Acta Mathematica Scientia. 2013;33B(2):430-442
  21. 21. Paulsen VI. An Introduction to the Theory of Reproducing Kernel Hilbert Spaces. Cambridge: Cambridge University Press; 2016
  22. 22. Folland G. Harmonic Analysis on Phase Space. Princeton, New Jersey: Princeton University Press; 1989
  23. 23. Gröchenig K. Foundations of Time-frequency Analysis. Boston: Birkhäuser; 2001
  24. 24. Saitoh S. Best approximation, Tikhonov regularization and reproducing kernels. Kodai Mathematical Journal. 2005;28:359-367
  25. 25. Saitoh S. Theory of reproducing kernels: Applications to approximate solutions of bounded linear operator equations on Hilbert spaces. In book: Selected papers on analysis and differential equations. Amer. Math. Soc. Transl. Series 2. 2010;230:107-134
  26. 26. Saitoh S, Sawano Y. Theory of reproducing kernels and applications. Developements in Mathematics. 2016;44:4

Written By

Fethi Soltani

Reviewed: 31 May 2022 Published: 24 May 2023