1. Introduction
Consider the first-order singular differential-difference operator on the real line

where
and
. For
, we regain the differential-difference operator

which is referred to as the Dunkl operator with parameter
associated with the reflection group
on
. Such operators have been introduced by Dunkl [1] -[3] in connection with a generalization of the classical theory of spherical harmonics. The one-dimensional Dunkl operator
plays a major role in the study of quantum harmonic oscillators governed by Wigner’s commutation rules [4] -[6] .
The authors have developed in [7] [8] a new harmonic analysis on the real line related to the differential-dif- ference operator
in which several classical analytic structures such as the Fourier transform, the translation operators, the convolution product, ..., were generalized. With the help of the translation operators tied to
, we construct in this paper generalized modulus of smoothness in the Hilbert space
. Next, we define Sobolev type spaces and K-functionals generated by
. Using essentially the properties of the Fourier transform associated to
, we establish the equivalence between K-functionals and modulus of smoothness.
In the classical theory of approximation of functions on
, the modulus of smoothness are basically built by means of the translation operators
. As the notion of translation operators was extended to various contexts (see [9] [10] and the references therein), many generalized modulus of smoothness have been discovered. Such generalized modulus of smoothness are often more convenient than the usual ones for the study of the connection between the smoothness properties of a function and the best approximations of this function in weight functional spaces (see [11] -[13] and references therein).
In addition to modulus of smoothness, the K-functionals introduced by J. Peetre [14] have turned out to be a simple and efficient tool for the description of smoothness properties of functions. The study of the connection between these two quantities is one of the main problems in the theory of approximation of functions. In the classical setting, the equivalence of modulus of smoothness and K-functionals has been established in [15] . For various generalized modulus of smoothness these problems are studied, for example, in [16] -[19] . It is pointed out that the results obtained in [16] emerge as easy consequences of those stated in the present paper by simply taking
.
2. Preliminaries
In this section, we develop some results from harmonic analysis related to the differential-difference operator
. Further details can be found in [7] [8] . In all what follows assume
and n a non-negative integer.
The one-dimensional Dunkl kernel is defined by
(1)
where
![]()
is the normalized spherical Bessel function of index
. It is well-known that the functions
,
, are solutions of the differential-difference equation
(2)
Furthermore, we have the Laplace type integral representations:
(3)
(4)
where
(5)
The following properties will be useful for the sequel.
Lemma 1 1) For all
,
.
2) There is
such that
for all
with
.
3) For all
,
.
4) For all
,
![]()
Proof. Assertions (1) and (2) are proved in [16] . By (1), (4) and the fact that
![]()
we have
![]()
Clearly the integral above is null only for
, which proves assertion (3). Let us check assertion (4). Using (3) and the fact that
(6)
![]()
we get
![]()
By (6),
![]()
Moreover,
![]()
which concludes the proof.
Notation 1 Put
![]()
We denote by
the class of measurable functions f on
for which
![]()
the space of
functions f on
, which are rapidly decreasing together with their derivatives, i.e., such that for all
,
![]()
The topology of
is defined by the semi-norms
,
.
the subspace of
consisting of functions f such that
![]()
the space of tempered distributions on
.
the topological dual of
.
Clearly
is a linear bounded operator from
into itself. Accordingly, if
define
by
![]()
For
and
, let
be defined by
![]()
Definition 1 The generalized Fourier transform of a function
is defined by
![]()
Remark 1 If
then
reduces to the Dunkl transform with parameter
associated with the reflection group
on
(see [3] ).
Theorem 1 The generalized Fourier transform
is a topological isomorphism from
onto
. The inverse transform is given by
![]()
where
![]()
Theorem 2 1) For every
we have the Plancherel formula
![]()
2) The generalized Fourier transform
extends uniquely to an isometric isomorphism from
onto
.
Definition 2 The generalized Fourier transform of a distribution
is defined by
![]()
Theorem 3 The generalized Fourier transform
is one-to-one from
onto
.
Lemma 2 If
then the functional
![]()
is a tempered distribution
. Moreover,
(7)
with
.
Proof. The fact that
follows readily by Schwarz inequality. Let
. It is easily checked that
![]()
where
. So using Theorem 2 we get
![]()
which completes the proof.
Lemma 3 Let
and
. Then for
we have
(8)
(9)
Proof. Identity (8) may be found in [7] . If
then
![]()
But by (8),
![]()
So
![]()
which ends the proof.
Notation 2 From now on assume
. Let
be the Sobolev type space constructed by the dif- ferential-difference operator
, i.e.,
![]()
More explicitly,
if and only if for each
, there is a function in
abusively denoted by
, such that
.
Proposition 1 For
we have
(10)
Proof. From the definition of
we have
![]()
By (7) and (9),
![]()
with
. Again by (7),
![]()
with
. Identity (10) is now immediate.
Definition 3 The generalized translation operators
,
, tied to
are defined by
![]()
where
![]()
with
given by (5).
Proposition 2 Let
and
. Then
and
(11)
Furthermore,
(12)
3. Equivalence of K-Functionals and Modulus of Smoothness
Definition 4 Let
and
. Then
The generalized modulus of smoothness is defined by
![]()
where
![]()
I being the unit operator.
The generalized K-functional is defined by
![]()
The next theorem, which is the main result of this paper, establishes the equivalence between the generalized modulus of smoothness and the generalized K-functional:
Theorem 4 There are two positive constants
and
such that
![]()
for all
and
.
In order to prove Theorem 4, we shall need some preliminary results.
Lemma 4 Let
and
. Then
(13)
(14)
Proof. The result follows easily by using (11), (12) and an induction on m.
Lemma 5 For all
and
we have
(15)
Proof. By (10), (14), Lemma 1 (4) and Theorem 2 we have
![]()
which is the desired result.
Notation 3 For
and
define the function
![]()
Proposition 3 Let
and
. Then
1) The function
is infinitely differentiable on
and
(16)
for all
.
2) For all
,
and
(17)
where
![]()
Proof. The fact that
follows from the derivation theorem under the integral sign. Identity (16) follows readily from (2) and the relationship
![]()
which is proved in [7] . Assertion (2) is a consequence of (16) and Theorem 2.
Lemma 6 There is a positive constant
such that
![]()
for any
and
.
Proof. By (17) and Theorem 2, we have
![]()
By Lemma 1 (2) there is a constant
which depends only on
and n such that
![]()
for all
with
. From this, (14) and Theorem 2 we get
![]()
which achieves the proof.
Corollary 1 For all
and
we have
![]()
where c is as in Lemma 6.
Lemma 7 There is a positive constant
such that
![]()
for every
and
.
Proof. By (17) and Theorem 2 we have
![]()
Put
![]()
By L’Hôpital’s rule,
![]()
This when combined with Lemma 1 (3) entails
. Moreover,
![]()
Therefore
![]()
by virtue of (14) and Theorem 2.
Corollary 2 For any
and
we have
![]()
where C is as in Lemma 7.
Proof of Theorem 4. 1) Let
and
. By (13) and (15), we have
![]()
Calculating the supremum with respect to
and the infimum with respect to all possible functions
we obtain
![]()
with
.
2) Let
be a positive real number. As
it follows from the definition of the K-functional and Corollaries 1 and 2 that
![]()
Since
is arbitrary, by choosing
we get
![]()
with
. This concludes the proof.