Volume 2020, Issue 1 6319138
Research Article
Open Access

An Analytic Characterization of (p, q)-White Noise Functionals

Anis Riahi

Anis Riahi

Department of Mathematics, College of Science Al-Zulfi, Majmaah University, P.O. Box 66, Al-Majmaah 11952, Saudi Arabia mu.edu.sa

Department of Mathematics, Nabeul Preparatory Institute for Engineering Studies, Carthage University, Nabeul, Tunisia carthage.edu

Search for more papers by this author
Amine Ettaieb

Amine Ettaieb

Department of Mathematics, Higher School of Sciences and Technologies of Hammam-Sousse, MaPSFA Laboratory, Sousse University, Hammam Sousse 4011, Tunisia uc.rnu.tn

Search for more papers by this author
Wathek Chammam

Corresponding Author

Wathek Chammam

Department of Mathematics, College of Science Al-Zulfi, Majmaah University, P.O. Box 66, Al-Majmaah 11952, Saudi Arabia mu.edu.sa

Department of Electro Mechanics, Higher Institute of Indiustrial Systems of Gabès, Gabès 6072, Tunisia

Search for more papers by this author
Ziyad Ali Alhussain

Ziyad Ali Alhussain

Department of Mathematics, College of Science Al-Zulfi, Majmaah University, P.O. Box 66, Al-Majmaah 11952, Saudi Arabia mu.edu.sa

Search for more papers by this author
First published: 08 December 2020
Citations: 1
Academic Editor: Yongqiang Fu

Abstract

In this paper, a characterization theorem for the -transform of infinite dimensional distributions of noncommutative white noise corresponding to the (p, q)-deformed quantum oscillator algebra is investigated. We derive a unitary operator U between the noncommutative L2-space and the (p, q)-Fock space which serves to give the construction of a white noise Gel’fand triple. Next, a general characterization theorem is proven for the space of (p, q)-Gaussian white noise distributions in terms of new spaces of (p, q)-entire functions with certain growth rates determined by Young functions and a suitable (p, q)-exponential map.

1. Introduction

The white noise distribution theory originally aiming to extend Itô theory keeping contact with Lévy’s stochastic variational calculus [1] has been developed to an efficient infinite dimensional calculus with considerable applications to quantum physics, infinite dimensional harmonic analysis, infinite dimensional differential equations, quantum stochastic calculus, and mathematical finance, see, e.g., [28] and the references therein. This theory is based on the quantum decomposition of the Gaussian random variable 〈ω, ξ〉 given as the sum of creation and annihilation operators which satisfies the canonical commutative commutation relation. As a generalization by replacing the classical commutative notion of independence by some other type in a noncommutative probability space, we conclude that the noncommutative white noise theory is a generalization of classical white noise theory to the description of quantum systems. In the framework of the free probability, Alpay and Salomon [9] (see also [10]) constructed a noncommutative analog of the Kondratiev space. For q ∈ (−1,1), Bożejko et al. introduced q-analogs of Brownian motions and Gaussian processes in [11, 12], which are governed by classical independence for q = 1 and free independence for q = 0 introduced by Voiculescu et al. in [13].

The aim of the present paper is to introduce a proper mathematical framework of (p, q)-white noise calculus based on the noncommutative white noise corresponding to the (p, q)-deformed oscillator algebra [14]. More precisely, as a generalization by using the second-parameter refinement of the q-Fock space, formulated as the (p, q)-Fock space ℱp,q(ℋ) which is constructed via a direct generalization of Bożejko and Speicher’s framework, yielding the q-Fock space when p = 1, we introduce the noncommutative analogs of Gaussian processes (white noise measure) for the relation of the (p, q)-deformed quantum oscillator algebra. Next, we construct a white noise Gel’fand triple, and we derive the characterization of the space of generalized functions in terms of new spaces of (p, q)-entire functions with certain growth.

Our paper is organized as follows. Section 2 is devoted to study the (p, q)-white noise functionals with special emphasis on the chaos decomposition of the noncommutative L2-space with respect to the vacuum expectation τ based on orthogonalization of polynomials of (p, q)-white noise. In Section 3, firstly for a fixed Young function θ with particular condition, we construct a nuclear Gel’fand triple
(1)
of test and generalized functions, and we introduce the -transform which is our main analytical tool in working with these spaces and serves to prove a characterization of white noise functionals.

2. Noncommutative Orthogonal Polynomials of (p, q)-White Noise

We start with the real Gel’fand triple:
(2)
where is the space of rapidly decreasing functions and is the dual space, i.e., the space of tempered distributions. We denote by 〈⋅, ⋅〉 the canonical bilinear form on ℰ × ℰ and by |⋅|0 the norm of H. For notational convenience, the -bilinear form on ℰ × ℰ is denoted by the same symbol so that holds for ξ ∈ ℋ≔H (in general, the complexification of a real vector space X is denoted by X). In [15], Simon has proved that the space ℰ is a nuclear Fréchet space constructed from the Hilbert space H and the Harmonic oscillator operator A = 1 + t − (d2/dt2), i.e., ℰ = proj limss = ∩s≥0s, where for s ≥ 0, ℰs is the Hilbert space corresponding to the domain of As, i.e.,
(3)
We define ℰs to be the completion of H with respect to , and hence we obtain a chain of Hilbert space {ℰs, s}, and one can see that
(4)
Let Sn denote the symmetric group of all permutations on 〚1, n〛≔{1, …, n} and I(σ) denote the number of inversions of the permutation σSn defined by
(5)
where (A) denotes the cardinality of the set A. Analogously, the pair (i, j) with i < j is called a coinversion in σ if σ(i) < σ(j). The corresponding coinversion is encoded by (i, j) and contained in the set
(6)
with cardinality C(σ)≔(Cinv(σ)). Denote ℱ0(ℋ) = ⊕n≥0n the full Fock space over ℋ with the inner product 〈⋅, ⋅〉 and the linear span of vectors of the form ξ1 ⊗ ⋯⊗ξn ∈ ℋn, n, where ℋ⊗0 = Ω for the vacuum vector Ω = (1,0,0, …) ∈ ℱ0(ℋ). We equip with the inner product
(7)
Recall that for p and q, two real numbers such that 0 < q < p ≤ 1, the (p, q)-factorial is defined by
(8)
where [n]p,q is the (p, q)-deformation of the natural number n given by
(9)
Define the operator on by a linear extension of
(10)
and put
(11)
Define as the separable Hilbert space which coincides with ℋn as a set and has a scalar product:
(12)
Hence, the (p, q)-Fock space denoted ℱp,q(ℋ) is defined by
(13)
and if we denote the linear span of vectors of the form
(14)
one can see that 〈⋅,⋅〉p,q on satisfies the following useful relation:
(15)

For more details about the properties of the operator and the construction of the (p, q)-Fock space, see [16].

Definition 1. For each ξ ∈ ℋ, we define the (p, q)-creation operator a(ξ) and the (p, q)-annihilation operator a(ξ) on the dense subspace as follows:

(16)
where 〈⋅, ⋅〉 denotes the inner product on ℋ and the symbol means that fi has to be deleted in the tensor product.

The (p, q)-creation and (p, q)-annihilation operators fulfill the (p, q)-commutation relations of the (p, q)-deformed quantum oscillator algebra, i.e.,

(17)
where
(18)
such that N is the standard number operator defined by
(19)
and the commutator [⋅, ⋅] is defined by [B, C] = BCCB. For more details, one can see [16].

Now, we will introduce noncommutative analogs of Gaussian processes (white noise measure) for the relation of the (p, q)-deformed quantum oscillator algebra. For t, if we denote by bt and the standard pointwise annihilation and creation operators on ℱp,q(ℋ) defined by

(20)
where δt is the delta function at t and stands for the symmetric tensor product, then one can see that the (p, q)-creation and (p, q)-annihilation operators are given as the smeared operators in terms of bt and , i.e.,
(21)

Now, the (p, q)-white noise is defined by

(22)

Thus, by using (21), we deduce that ω(t) is an operator-valued distribution which satisfies

(23)

Moreover, for each , we define a monomial of ω by

(24)

Using the Cauchy–Schwarz inequality, we easily conclude that (24) indeed identifies a bounded linear operator in ℱp,q(ℋ).

Let denote the complex unital ∗-algebra generated by {〈ω, ξ〉, ξ ∈ ℋ}, i.e., the algebra of noncommutative polynomials in the variables 〈ω, ξ〉. Evidently, consists of all noncommutative polynomials in ω which are of the form:

(25)

In particular, elements of are linear operators acting on ℱp,q(ℋ).

Definition 2. Let τ be a vacuum state on defined by

(26)
where Ω is the vacuum vector in ℱp,q(ℋ). The inner product on is defined by
(27)
where is the adjoint operator of T2 in ℱp,q(ℋ). Moreover, the noncommutative L2-space L2(τ) is the Hilbert space obtained as the closure of with respect to the norm induced by the scalar product .

For n\{0}, we denote by the subset of consisting of all noncommutative polynomials of order ≤n, i.e., all given as in (25) with kn. Let denote the closure of in L2(τ), and let be the set of orthogonal polynomials of order n defined by

(28)
where ⊖ denotes the orthogonal difference in L2(τ).

Theorem 1. For each , define UTTΩ. Then, U is extended by continuity to a unitary operator U : L2(τ)⟶ℱp,q(ℋ) defined as follows:

(29)
where 〈:ωn : ξ(n)〉 is the orthogonal projection of the monomial 〈ωn, ξ(n)〉 onto and given recursively by
(30)

Furthermore, under the action of U, the operator of the left multiplication by the monomial 〈ω, ξ〉 in L2(τ) (denoted by Mω, ξ) becomes 〈ω, ξ〉, i.e.,
(31)

Proof. Firstly from equation (27), it is clear that U : L2(τ)⟶ℱp,q(ℋ) is extended by continuity to a unitary operator. Moreover, since is dense in L2(τ), we get the orthogonal decomposition

(32)
and for each n, one can see that . Hence, using the fact that U(〈:ωn : ξ(n)〉) is the orthogonal projection in ℱp,q(ℋ) of
(33)
on , we obtain
(34)

On the other hand, we have

(35)
and this gives 〈:ω : ξ〉 = 〈ω, ξ〉. Therefore, (23) and (34) yield
(36)
from which (30) follows.

3. (p, q) White Noise Gel’fand Triple and Characterization Theorem

Recall that a Young function is a continuous, convex, and increasing function
(37)
such that
(38)
Define a weight sequence by
(39)
where θ is a Young function and Ep,q is the (p, q)-exponential function defined by
(40)
Let be the sequence associated with the (p, q)-polar function θ of θ, defined by
(41)
For simplicity of notation, we denote
(42)
where is constructed as for by replacing ℋ by the space ℰs, given by equation (3). Now, suppose a pair s ≥ 0 and γ > 0 is given, then for with , we put
(43)
Hence, we obtain a projective system of Hilbert spaces
(44)
where
(45)
Finally, we define the nuclear space ℱp,q,θ(ℰ) by
(46)

Definition 3. The space of (p, q)-white noise test functions is defined as a projective system of Hilbert space , where is the set of function φ of the form

(47)
such that
(48)

Moreover, if is the set of functions Φ of the form

(49)
equipped with Hilbertian norm
(50)
the space of (p, q)-white noise generalized functions is defined by
(51)

Theorem 2. Assume that the Young function θ satisfies the following condition:

(52)

Then, we obtain the so-called (p, q)-white noise Gel’fand triple of Hilbert spaces

(53)
with the -bilinear form on given by
(54)
where 〈⋅,⋅〉p,q is the canonical -bilinear form on which is compatible with the inner product of defined by equation (12).

Proof. Let . By definition, we have

(55)

On the other hand, condition (52) guarantees the existence of two constant numbers a > 0 and b > 0 such that

(56)

Then, by a simple calculus, one can see that

(57)

Hence, by using the fact that

(58)
and the inequality (57), we obtain
(59)

Therefore, for γ < (2be)−1, we have . Thus, (55) becomes

(60)
which means that ℱp,q,θ(ℰ) ⊂ ℱp,q(ℋ) and the inclusion is continuous. On the other hand, if we put , where U is the isomorphism given in Theorem 1, we obtain the following diagram:
(61)

Moreover, one can see that is the dual of with respect to L2(τ), and we obtain the nuclear Gel’fand triple given by (53). From here the statement follows.

Now our goal is to derive a characterization of the space of (p, q)-white noise generalized functions by using a suitable space of (p, q)-entire functions with certain growth determined by using the Young functions and a suitable (p, q)-exponential map.

Let (ℬ, ‖⋅‖) be a complex Banach space. Define the space ℬ by

(62)

Then, (ℬ, ‖⋅‖) becomes a Banach space.

Definition 4. Let be a fixed Hilbert space. A -valued function F is said to be (p, q)-entire function on , if there exists with such that

(63)
where the series in the right hand side of (63) converges uniformly on every bounded subset of .

For s and γ > 0, let Γp,q,θ,γ(ℰs,) be the space of (p, q)-entire functions g on the complex Hilbert space such that

(64)

Note that {Γp,q,θ,γ(ℰs,); s, γ > 0} becomes a projective system of Banach spaces as s and γ ↓ 0. Then, we can define

(65)

This is called the space of (p, q)-entire functions on with (θ, p, q)-exponential growth of minimal type. Similarly, {Γp,q,θ,γ(ℰs,); s, γ > 0} becomes a inductive system of Banach spaces as s and γ ↓ 0. the space of (p, q)-entire functions on ℰs, with (θ, p, q)-exponential growth of finite type is defined by

(66)

Lemma 1. Let be given by

(67)
where and . Then, for any s ≥ 0, there exists s > s such that the canonical embedding is of the Hilbert–Schmidt type, and for γ > 0, we get
(68)

Proof. Fixing s > s ≥ 0 and such that

(69)

By definition, the series in the right hand side of (67) converges uniformly on every bounded subset of . Then, for every R > 0, we have the following Cauchy’s integral formula:

(70)

Therefore, by using the fact that , for γ > 0, we get

(71)
which gives
(72)

Let now be an orthonormal basis of . Then, we get

(73)

This provides the desired inequality.

Lemma 2. For each ξ ∈ ℰ, the generating function of the noncommutative polynomials defined by

(74)
is an element in . Moreover, for all s ≥ 0 and γ > 0, there exist c, γ > 0, and s > s such that
(75)

Proof. Let s ≥ 0 and γ > 0, and then for any ξ ∈ ℰs,, we have

(76)

This proves that , and we obtain

(77)

On the other hand, if we choose s > s such that the embedding is of the Hilbert–Schmidt type and γ > 0 such that

(78)
and by using Lemma 1 and (77), we get
(79)

This implies that .

As a consequence, we can define the -transform of a distribution , at ξ ∈ ℰ, as follows:

(80)

Moreover, by using (15) and (49), we get

(81)

Theorem 3. Assume that the sequences θn,p,q and satisfy

(82)
for some constant β > 0. Then, the -transform realizes a topological isomorphism from onto the space .

Proof. Let , and then there exist s ≥ 0 and γ > 0 such that , and we have

(83)

On the other hand, by inequality (75), there exist c, γ > 0, and s > s such that

(84)
which yields
(85)

This proves the continuity and injectivity of the -transform.

Conversely, given , then there exist s ≥ 0 and γ > 0 such that with Taylor expansion

(86)

Put , and then (15) and (81) yield

(87)

Using the same technics as in Lemma 1, we immediately prove that for all s > s such that is of the Hilbert–Schmidt type the following inequality holds:

(88)

Thus, under condition (82), we obtain

(89)

On the other hand, for all γ > 0 such that , one can see that the series converges. This proves that acts surjectively and that is continuous.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

The authors extend their appreciation to the Deanship of Scientific Research at Majmaah University for funding this work under project number (RGP-2019-1).

    Data Availability

    All data required for this paper are included within this paper.

      The full text of this article hosted at iucr.org is unavailable due to technical difficulties.