J. Pseudo-Differ. Oper. Appl. (2014) 5:215–230 DOI 10.1007/s11868-014-0092-6
Characterization of spaces of type W and pseudo-differential operators of infinite order involving fractional Fourier transform S. K. Upadhyay · Anuj Kumar · Jitendra Kumar Dubey
Received: 17 September 2013 / Revised: 13 March 2014 / Accepted: 15 March 2014 / Published online: 21 April 2014 © Springer Basel 2014
Abstract The characterization of W-type spaces is investigated and various properties of pseudo-differential operators are studied by using the fractional Fourier transform. Keywords Fractional Fourier transform · Pseudo-differential operator · Convex functions · Gel’fand and Shilov spaces Mathematics Subject Classification (2010)
46F12 · 46E35 · 35S05
1 Introduction The theory of pseudo-differential operators was introduced by Kohn–Nirenberg [8], Hörmander [6,7], Wong [13] and others. They applied pseudo-differential operators associated with different types of symbols on the Schwartz space S(Rn ) and studied many properties by exploiting the Fourier transformation. Cappiello [2], Zanghirati [14], Boutet de Monvel [1] and Upadhyay et al. [11,12] studied the concept of pseudodifferential operators of infinite order on Gevrey, Gelfand and Shilov types of spaces by using the Fourier transformation. The characterization of pseudo-differential operators associated with fractional Fourier transformation on the Schwartz space S(R) was
S. K. Upadhyay (B) DST-CIMS, Department of Applied Mathematics, Indian Institute of Technology, Banaras Hindu University, Varanasi 221005, India e-mail:
[email protected] A. Kumar · J. K. Dubey DST-CIMS, Banaras Hindu University, Varanasi 221005, India e-mail:
[email protected] J. K. Dubey e-mail:
[email protected]
216
S. K. Upadhyay et al.
given by Pathak et al. [9], Prasad and Kumar [10]. Our main aim of this paper is to discuss the properties of W-type spaces and to study the properties of pseudo-differential operators involving fractional Fourier transformation on the space W M (Rn ), where Rn = {(x1 , . . . , xn ) : x j ’s are real numbers} is usual Euclidean space. If x = (x1 , . . . , xn ) and y = (y1 , . . . , yn ). Then the inner product of x and y is defined by n x, y = x.y = x j .y j (1.1) j=1
and the norm of x is defined by ⎛ |x| = ⎝
n
⎞1
2 1 2 x 2j ⎠ = x12 + · · · + xn2 .
(1.2)
j=1
Fractional Fourier transform is the generalization of Fourier transform. It is frequently applied in signal processing, optics, quantum physics, mathematical sciences and many other scientific fields. The n-dimensional fractional Fourier transform with parameter α of f (x) on x ∈ Rn is denoted by (Fα f )(ξ ) [3] and defined as ˆ (1.3) f α (x) = (Fα f )(ξ ) = K α (x, ξ ) f (x)d x, ξ ∈ Rn Rn
where K α (x, ξ ) =
⎧ i(|x|2 +|ξ |2 ) cot α ⎪ −ix,ξ csc α ⎨Cα e 2 if
α = nπ
⎪ ⎩
α=
1 n (2π ) 2
e−ix,ξ
if
π 2,
∀ n ∈ Z,
and −n
inα
Cα = (2πi sin α) 2 e 2 1 = n . [π(1 − e−2iα )] 2 The corresponding inversion formula is given by f (x) = K α (x, ξ ) fˆα (ξ ) dξ, x ∈ Rn Rn
where the kernel K α (x, ξ ) = Cα e
−i(|x|2 +|ξ |2 ) cot α +ix,ξ csc α 2
,
(1.4)
Characterization of spaces of type W and pseudo-differential operators
217
and n
inα
Cα = (2πi sin α)− 2 e 2 − n 2 = π(1 − e−2iα ) . Now, the Eq. (1.3) can be written as, Fα [ f (x)](ξ ) = Cα e
i|ξ |2 cotα 2
e
−ix,ξ csc α
Rn n
= (2π ) 2 Cα e Replacing f (x) = e Fα [e
i|x|2 cot α 2
−i|x|2 cotα 2
i|ξ |2 cot α 2
f (x)e
f (x)e
i|x|2 cot α 2
i|x|2 cotα 2
dx
ˆ (ξ csc α).
(1.5)
φ(x) in (1.5), we obtain n
φ(x)](ξ ) = (2π ) 2 Cα e
i|ξ |2 cot α 2
[φ(x)ˆ](ξ csc α).
(1.6)
Now substituting ξ = sin α · w, where w ∈ Rn in (1.6), we obtain Fα [e
−i|x|2 cot α 2
n
φ(x)](sin α · w) = (2π ) 2 Cα e
i|sin α·w|2 cot α 2
[φ(x)ˆ](w).
(1.7)
−i|x|2 cot α 2 φ(x) , then (1.7) can be written as Let ψ = Fα e n
ψ(sin α · w) = (2π ) 2 Cα e
i|sin α·w|2 cot α 2
[φ(x)ˆ](w).
(1.8)
Now we recall the definition of W-type spaces from [4,5], which are given below. Let μ j and w j , j = 1, . . . , n, be continuous and increasing functions on [0, ∞) with μ j (0) = w j (0) = 0 and μ j (∞) = w j (∞) = ∞. We define x j M j (x j ) = μ j (ξ j )dξ j , (x j ≥ 0) (1.9) 0
y j j (y j ) =
w j (η j )dη j ,
(y j ≥ 0)
(1.10)
0
where j = 1, . . . , n. The functions M j (x j ) and j (y j ) are continuous, increasing and convex with M j (0) = j (0) = 0 and M j (∞) = j (∞) = ∞, we have (1.11) M j (−x j ) = M j (x j ), M j (x j ) + M j (x j ) ≤ M j (x j + x j ), j (−y j ) = j (y j ),
j (y j ) + j (y j ) ≤ j (y j + y j ).
(1.12)
218
S. K. Upadhyay et al.
We set μ(ξ ) = (μ1 (ξ1 )), . . . , (μn (ξn )), w(η) = (w1 (η1 )), . . . , (wn (ηn )). The space W M (Rn ) consists of all C ∞ -complex valued functions φ(x) on Rn which for some a ∈ Rn+ satisfy the inequality k Dx φ(x) ≤ Ck ex p [−M(ax)] ,
(1.13)
where, ex p [−M(ax)] = ex p [−M1 (a1 x1 ) − · · · − Mn (an xn )] , Dxk = Dxk11 · · · Dxknn and a1 , . . . , an , Ck are positive constants depending on the function φ(x) and the space W M,a (Rn ) consists of all infinitely differentiable functions φ(x) on x ∈ Rn , which for any δ ∈ Rn+ satisfy the inequality k Dx φ(x) ≤ Ck,δ ex p [−M(a − δ)x]
(1.14)
where ex p [−M(a − δ)x] = ex p −M1 (a1 − δ1 )x1 − · · · − M j (a j − δ j )x j − · · · − Mn (an − δn )xn and a, δ ∈ Rn+ , depend on the function φ(x). The space W (Cn ) consists of all entire analytic functions φ(z), where z = x + i y n , satisfy the inequality and x, y ∈ Rn , which for some b ∈ R+ k z φ(z) ≤ Ck ex p [(by)] , where
(1.15)
z k = z 1k1 . . . z nkn ,
ex p [(by)] = ex p 1 (b1 y1 ) + · · · + j (b j y j ) + · · · + n (bn yn ) , and Ck , b1 , . . . , bn are positive constants depending on the function φ(x), and the space W ,b consists of all entire analytic functions φ(z) such that for k ∈ Zn+ , ρ ∈ Rn+ , there exists a constants Ck,ρ > 0 such that k z φ(z) ≤ Ck,ρ ex p [(b + ρ)y] ,
(1.16)
where ex p [(b + ρ)y] = ex p 1 (b1 + ρ1 )y1 + · · · + j (b j + ρ j )y j + · · · + n (bn + ρn )yn ,
Characterization of spaces of type W and pseudo-differential operators
219
(Cn ) consists of all entire analytic functions φ(z) such that there The space W M n exist a, b ∈ R+ and C > 0 such that
|φ(z)| ≤ C ex p [−M[(ax)] + [(by)]] ,
(1.17)
where ex p [−M(ax)] and ex p [(by)] have usual meaning like (1.13) and (1.15), ,b and the space W M,a (Cn ) consists of all entire analytic functions φ(z) such that for n ρ, δ ∈ R+ and Cρ,δ > 0, |φ(z)| ≤ Cρ,δ ex p [−M[(a − δ)x] + [(b + ρ)y]] ,
(1.18)
where ex p [−M[(a − δ)x]] and ex p [[(b + ρ)y]] have usual meaning like (1.14) and (1.16), and the constants Cρ,δ , a, b and ρ, δ, depend only on the function φ(z). Let M j (x j ) and j (y j ) be the functions defined by (1.9) and (1.10), where the functions μ j (ξ j ) and w j (η j ) which occur in these equations are mutually inverse, that is μ j (w j (η j )) = η j and w j (μ j (ξ j )) = ξ j , then the corresponding functions M j (x j ) and j (y j ) are said to be the dual in sense of Young. In this case, the Young inequality (1.19) x j y j ≤ M j (x j ) + j (y j ), holds for any x j ≥ 0, y j ≥ 0. Definition 1.1 Let m ∈ R. Then we define the symbol class U m to be the space of all entire analytic functions θ (x, ξ ) ∈ C ∞ (Rn × Cn ) in ξ such that for any two multiindices μ and ϑ , there is a positive constant Cμ,ϑ depending upon μ and ϑ such that μ ϑ (1.20) (Dx Dξ θ )(x, ξ ) ≤ Cμ,ϑ (1 + |ξ |)m−|ϑ| ex p[(a0 t)] ∀ x ∈ Rn , ξ ∈ Cn and ξ = u + it. If we put t = 0 in (1.20), then the symbol class U m reduces to the symbol class introduced by Kohn and Nirenberg [8], see also Wong [13]. Definition 1.2 Let θ (x, ξ ) be a symbol belonging to U m , then the pseudo-differential operator Aθ,α associated with θ (x, ξ ) is defined as (Aθ,α φ)(x) =
K α (x, ξ ) θ (x, ξ )φˆ α (ξ ) dξ , φ ∈ W M (Rn )
(1.21)
Rn
where φˆ α (ξ ) is the fractional Fourier transform of φ, defined by (1.3). 2 Characterization of W-type spaces In this section we study the characterization of W-type spaces by using the fractional Fourier transformation.
220
S. K. Upadhyay et al.
Theorem 2.1 Let M(x) and (y) be a pair of functions which are dual in sense of Young . Then 1 Fα W M,a ⊂ W , a . (2.1) i|x|2 cot α
Proof Let e 2 φ(x) ∈ W M,a (Rn ) and σ = w + iτ . Then by the technique of Gel’fand and Shilov [5, pp. 20–21] and (1.8), we have 2 k k i|sin α·σ2| cot α (sin α · σ ) ψ(sin α · σ ) = Cα (sin α · σ ) e e−iσ,x φ(x)d x Rn
= Cα (sin α)|k| e
i|sin α·σ |2 cot α 2
(i)|k|
Dxk e−iσ,x φ(x)d x.
Rn
Now using the integration by parts, we get (sin α · σ )k ψ(sin α · σ ) = Cα (sin α)|k| e
i|sin α·σ |2 cot α φ(x) 2
(i)|k| (−1)|k|
e−iσ,x Dxk φ(x)d x
Rn
Thus, we get −n inα (sin α · σ )k ψ(sin α · σ ) ≤ (2π sin α) 2 e 2 |sin α||k| e−τ,x Dxk φ(x) d x ≤ D |sin α|
|k|− n2
Rn
e|τ,x| Dxk φ(x) d x
Rn n
≤ Ck,δ D |sinα||k|− 2 ≤ Ck,δ,sin α
ex p(|x| |τ |)ex p [−M(a − δ)x] d x Rn
ex p [|x| |τ | − M(a − δ)x] d x.
(2.2)
Rn
Using the young inequality (1.19), then the exponent of (2.2) can be evaluated as follows τ |x| |τ | − M [(a − δ)x] ≤ −M [(a − δ)x] + M [(a − 2δ)x] + . (a − 2δ) Hence (2.2) can be written as (sin α · σ )k ψ(sin α · σ ) τ (a−2δ) ≤ Ck,δ,sin α e ex p [−M(a − δ)x + M(a − 2δ)x] d x. Rn
Characterization of spaces of type W and pseudo-differential operators
Therefore,
(sin α · σ )k ψ(sin α · σ ) ≤ C ex p
221
τ (a − 2δ)
.
1 The quantity a−2δ can be represented in the form a1 + ρ, where ρ = (ρ1 , . . . , ρn ), ρ j arbitrarily small. Therefore the last expression can be written as, 1 k (sin α · σ ) ψ(sin α · σ ) ≤ Cex p [( + ρ)τ ] . a
Theorem 2.2 Let M(x) be the function which is dual in the sense of Young of (y). Then Fα W ,b ⊂ W M, 1 . b
−i|z|2 cotα
Proof Let e 2 φ(z) ∈ W ,b (Cn )and σ = w + iτ . Then by the technique of [5, pp. 21–22] and (1.8), we have −i|z|2 cotα k k Dw Fα e 2 φ(z) (sin α · w) ψ(sin α · w) = Dw ˆ i|sin α·w|2 cotα n k 2 φ(z) (w) = Dw (2π ) 2 Cα e ⎫ ⎧ ⎬ ⎨ i|sin α·w|2 cotα k 2 = C α Dw e−iw,z φ(z)d x e ⎭ ⎩ Rn ⎞ ⎛ k i|sin α·w|2 cot α r k−r ⎝ 2 Dw e Dw = Cα e−iw,z φ(z)d x ⎠ r r ≤k Rn k i|sin α·w|2 cot α i |sin α|2 cot α 2 = Cα Pr ,w e r 2 r ≤k × (i z)k−r e−iw,z φ(z)d x Rn
= C(α) ×
k i|sin α·w|2 cot α 2 Cη wη e r η≤r r ≤k
(i z)k−r e−iw,z φ(z)d x,
Rn
where C(α) = Cα | sin α|2 cot α and Pr
−i|sin α|2 cot α ,w 2
is a polynomial of degree r.
222
S. K. Upadhyay et al.
Therefore, k ψ(sin α Dw
k i|sin α·w|2 cot α 2 · w) = C(α) Cη e (i)|k−r |−|η| r η≤r r ≤k Dzη e−iw,z z k−r φ(z)d x. × Rn
k i|sin α·w|2 cot α 2 Cη e (i)|k−r |−|η| r η≤r r ≤k ×(−1)|η| e−iw,z Dzη z k−r φ(z) d x
= C(α)
Rn
k i|sin α·w|2 cotα 2 Cη e (i)|k−r |−|η| = C(α) r η≤r r ≤k η Dzη−s z k−r Dzs φ(z) d x ×(−1)|η| e−iw,z s s≤η Rn k i|sin α·w|2 cotα 2 Cη e (i)|k−r |−|η| = C(α) r η≤r r ≤k η (k − r )! × eiw,z z k−r −η+s Dzs φ(z)d x. s (k − r − η + s)! s≤η Rn
Thus, we have k k Cη Dw ψ(sin α · w) ≤ |C(α)| r η≤r r ≤k η (k − r )! −w,y k−r −η+s s × Dz φ(z) d x. e z s (k − r − η + s!) s≤η Rn
Now using the inequality |z| p ≤ |z||x|2+|z| and (1.14), we get +1 k [(b+ρ)]y −w,y Dw ψ(sin α · w) ≤ D(Ck−r −η+s+2,ρ+Ck−r −η+s,ρ )e e p+2
p
≤ Ck,r,η,s,ρ ex p[− w, y + (b + ρ)y]
Rn
dx (|x|2 + 1)
So that, k Dw ψ(sin α · w) ≤ Ck,r,η,s,ρ ex p[−w1 y1 + 1 [(b1 + ρ1 )y1 ] · · · − w j y j + j [(b j + ρ j )y j ] · · · − wn yn + n [(bn + ρn )yn ]]. (2.3)
Characterization of spaces of type W and pseudo-differential operators
223
Arguing as in [5, p. 22], we choose the sign of y j in such a way that the following equality holds wj + (b j + ρ j )y j . w j y j = w j y j = M j (b j + ρ j ) The j-th term of the exponent in the expression (2.3), becomes wj . −w j y j + [(b j + ρ j )y j ] = −M j (b j + ρ j ) Hence, we have k Dw ψ(sin α · w)
wj w1 ≤ Ck,ρ,r,η,s ex p −M1 − ··· − Mj (b1 + ρ1 ) (b j + ρ j ) wn − · · · − Mn (bn + ρn ) w . = Ck,ρ,r,η,s ex p −M (b + ρ)
1 In the above expression, we get b+ρ = small, then we obtain the inequality
1 b
− δ, where δ = (δ1 , . . . , δn ), δ j is arbitrarily
1 k −δ w . Dw ψ(sin α · w) ≤ Ck,ρ,r,η,s ex p −M b This implies that ψ(sin α · w) ∈ W M, 1 . b
Theorem 2.3 Let M0 (x) and 0 (y) be dual in sense of Young to the functions M(x) and (y), respectively. Then 0 , 1 ,b ⊂ W 1a . Fα W M,a M0 , b
−i|x|2 cotα
,b (Rn ) and σ = w + iτ .Then by the technique of Proof Let e 2 φ(x) ∈ W M,a Gel’fand and Shilov [5, pp. 23–24] and (1.8), we have i|sin α·σ |2 cot α 2 e−iσ,z φ(z)d x ψ(sin α · σ ) = Cα e
Rn
= Cα e
i|sin α·σ |2 cot α 2
Rn
e−iw+iτ,x+i y φ(x + i y)d x.
224
S. K. Upadhyay et al.
Thus, we get −n inα |ψ(sin α · σ )| ≤ (2πi sin α) 2 e 2 eiw+iτ,x+i y |φ(z)| d x Rn
−n inα ≤ (2πi sin α) 2 e 2 e−w,y−τ,x |φ(z)| d x Rn
Using (1.18), above expression yields |ψ(sin α · σ )| ≤ Cρ,δ,α e−w,y ex p [[(b+ρ)y]]
ex p [−M[(a − δ)x]+|τ | |x| ] d x.
Rn
Now using the arguments of [2, p. 24], we have 1 1 |ψ(sin α · σ )| ≤ Cρ,δ,α ex p −M0 ( − δ1 )w + 0 ( + ρ1 )τ . b a This shows that ψ(sin α · σ ) ∈ W
0 , a1 M0 , b1
.
3 Pseudo-differential operators of infinite order In this section the mapping properties of pseudo-differential operators defined by (1.21) on W M (Rn ) are obtained. Theorem 3.1 Let θ (x, ξ ) ∈ U m , where m ∈ R. Then Aθ,α maps W M (Rn ) into itself. Proof Let φ ∈ W M (Rn ). Then for non-negative integers μ and ν, we prove that sup ex p [M(ax)] Dxμ Aθ,α φ (x) < ∞.
x∈Rn
(3.1)
From (1.21) for φ ∈ W M (Rn ), we have Dxμ
Aθ,α φ (x) =
= Cα
Rn
=
Cα
Dxμ
K α (x, ξ ) θ (x, ξ )φˆ α (ξ ) dξ ,
Rn
−i(|x|2 +|ξ |2 ) cot α μ +ix,ξ csc α 2 Dx e θ (x, ξ ) φˆ α (ξ ) dξ
μ−r −i(|x|2 +|ξ |2 ) cot α μ r +ix,ξ csc α 2 Dx θ (x, ξ )φˆ α (ξ ) dξ. Dx e r r ≤μ
Rn
Characterization of spaces of type W and pseudo-differential operators
225
So that, x ν Dxμ Aθ,α φ (x) −i(|x|2 +|ξ |2 ) cot α μ r r r −r ix,ξ csc α 2 = x ν Cα e D e D x x r r r ≤μ
Rn
r ≤r
(x, ξ )φˆ α (ξ ) dξ μ r −i(|x|2 +|ξ |2 ) cot α i cot α ix,ξ csc α ν 2 e e = x Cα Pr x, r r 2 r ≤μ ×
Dxμ−r θ
Rn
r −r
×(iξ csc α)
r ≤r
Dxμ−r θ (x, ξ )φˆ α (ξ ) dξ,
α where Pr x, i cot is a polynomial of degree r. 2 x ν Dxμ Aθ,α φ (x) μ r −i(|x|2 +|ξ |2 ) cot α 2 = x ν Cα a cot α e x s (i csc α)|r −r | s r r r ≤μ r ≤r s≤r Rn ×eix,ξ csc α Dxμ−r θ (x, ξ ) ξ r −r φˆ α (ξ ) dξ. Since Dξν+s eix,ξ csc α = (i x csc α)ν+s eix,ξ csc α , we have x ν Dxμ Aθ,α φ (x) −i(|x|2 +|ξ |2 ) cot α μ r ν+s ix,ξ csc α 2 e e = Cα a cot α D s ξ r r r ≤μ r ≤r s≤r Rn ×(i cos α)|r −r |−|ν+s| Dxμ−r θ (x, ξ ) ξ r −r φˆ α (ξ ) dξ. Using integration by parts, we have x ν Dxμ Aθ,α φ (x) μ r = Cα as cot α(−1)|ν+s| (i csc α)|r −r |−|ν+s| r r r ≤μ r ≤r
s≤r
−i(|x|2 +|ξ |2 ) cot α ix,ξ csc α ν+s μ−r r −r ˆ 2 e φα (ξ ) dξ Dx θ (x, ξ ) ξ e Dξ
× Rn
= Cα
μ r
r ≤μ
r
r ≤r
r
s≤r
as cot α(−1)|ν+s| (icscα)|r −r |−|ν+s|
226
S. K. Upadhyay et al.
×
e
β≤ν+s
Rn ν+s−β
×Dξ = Cα
ν + s β −i|ξ |2 cot α 2 Dξ e Dxμ−r θ (x, ξ ) β
−i|x|2 cot α +ix,ξ csc α 2
! ξ r −r φˆ α (ξ ) dξ
μ r r ≤μ
×
e
r
r ≤r
r
as cot α
s≤r
−i|x|2 cot α −ix,ξ cscα 2
ν + s (−1)|ν+s| (i csc α)|r −r |−|ν+s| β
β≤ν+s
β β −i|ξ |2 ) cot α β−β μ−r 2 D e D D θ (x, ξ ) x ξ ξ β
β ≤β
Rn ν+s−β
×Dξ
! ξ r −r φˆ α (ξ ) dξ
Thus, we have x ν Dxμ Aθ,α φ (x) μ r ν + s = Cα (−1)|ν+s| (icscα)|r −r |−|ν+s| as cotα r r β r ≤μ r ≤r
β≤ν+s
s≤r
β −i(|x|2 +|ξ |2 ) cot α +ix,ξ csc α 2 × a cot α e s 1 β β ≤β
s1 ≤β
β−β
×ξ s1 Dξ
Rn
ν+s−β
Dxμ−r θ (x, ξ )Dξ
! ξ r −r φˆ α (ξ ) dξ.
So that using the arguments of [7, p. 121], we get ν μ x D Aθ,α φ (x) x
ν + s μ r |a |csc α||r −r |−|ν+s| ≤ Cα cotα| s r r β r ≤μ r ≤r
s≤r
β≤ν+s
β −x,t csc α s as1 cotα × e |ξ 1 | β β ≤β
s1 ≤β
Rn
! ν+s−β r −r ˆ β−β μ−r φα (ξ ) dξ ξ Dx θ (x, ξ ) Dξ × Dξ ν + s μ r |as cotα| |csc α||r −r |−|ν+s| ≤ Cα r r β r ≤μ r ≤r
s≤r
β≤ν+s
Characterization of spaces of type W and pseudo-differential operators
×
227
β as cot α e−x,t csc α D β−β D μ−r θ (x, ξ ) (1 + |ξ |s1 ) 1 x ξ β
β ≤β
s1 ≤β
Rn
! ν+s−β r −r ˆ φα (ξ ) (1 + |ξ |k )(1 + |ξ |k )−1 dξ. ξ × Dξ Now for k > 0, using (1 + |ξ |k )−1 ≤ 2k−1 (1 + |ξ |)−k where ξ = u + it, we obtain ν μ x D Aθ,α φ (x) x
ν + s μ r |csc α||r −r |−|ν+s| |as cot α| ≤ Cα r r β r ≤μ r ≤r
s≤r
β≤ν+s
β k−1 as1 cot α Cβ−β ,μ−r 2 × (1 + |ξ |)m−k−|β−β | β β ≤β
s1 ≤β
Rn
! ν+s−β r −r ˆ ×e[(a0 t)] (1 + |ξ |s1 )(1 + |ξ |k )e−x,t csc α Dξ ξ φα (ξ ) dξ. Therefore, we have ν μ x D Aθ,α φ (x) x
ν + s μ r |csc α||r −r |−|ν+s| |a ≤ Cα cot α| s r r β r ≤μ r ≤r
s≤r
β≤ν+s
β k−1 as1 cot α Cβ−β ,μ−r 2 × (1 + |ξ |)m−k−|β−β | β β ≤β
s1 ≤β
Rn
ν+s−β r −r ˆ φα (ξ ) ξ ×ex p[[(a0 t)] − x, t csc α] Dξ ν+s−β ν+s−β + ξ k+s1 Dξ ξ r −r φˆ α (ξ ) + ξ k Dξ ξ r −r φˆ α (ξ ) ν+s−β + ξ s1 Dξ ξ r −r φˆ α (ξ ) dξ.
(3.2)
Now using the definition of W (Cn ) and making reference to Gel’fand and Shilov [5, pp. 13–14], we obtain ν μ x D Aθ,α φ (x) x ≤ D C0,ν+s−β + Ck+s1 ,ν+s−β + Ck,ν+s−β + Cs1 ,ν+s−β × (1 + |ξ |)m−k−|β−β | ex p[[(a0 + b)t] − x, t csc α] dξ, Rn
228
S. K. Upadhyay et al.
where D be a positive constant depending on α, ν, μ, r, r , s, s1 and β, β . Now using the Young enquality (1.19) inside of above integral, we get ν μ x D Aθ,α φ (x) x −1 ≤ Dα ex p −M[(b + a0 ) x csc α] (1 + |ξ |)m−k−|β−β | dξ. Rn
The last integral is convergent for sufficiently large k. Thus above expression can be written as ν μ x D Aθ,α φ (x) ≤ Dm,k ex p −M[(b + a0 )−1 x csc α] . (3.3) x This implies that ex p[M[(b + a0 )−1 x csc α]]Dxμ Aθ,α φ (x) ≤ Dm,k,ν (1 + |x|ν )−1 . Therefore, sup ex p[M[(b + a0 )−1 x csc α]]Dxμ Aθ,α φ (x) ≤ Dm,k,ν sup (1 + |x|ν )−1 < ∞.
x∈Rn
x∈Rn
This implies that Aθ,α maps W M (Rn ) into itself.
Theorem 3.2 Let θ (x, ξ ) ∈ U m be a symbol, where m ∈ R. Then Aθ,α maps W M (Rn ) continuously into itself. Proof Using (3.2), we have ν μ x D Aθ,α φ j (x) x ν + s μ r |csc α||r −r |−|ν+s| |as cot α| ≤ Cα r r β r ≤μ β≤ν+s r ≤r s≤r β as cot α Cβ−β ,μ−r 2k−1 (1 + |ξ |)m−k−|β−β | × 1 β β ≤β s1 ≤β Rn ν+s−β r −r ˆ ξ ×e[(a0 t)] e−x,t csc α Dξ φα, j (ξ ) ν+s−β ν+s−β + ξ k+s1 Dξ ξ r −r φˆ α, j (ξ ) + ξ k Dξ ξ r −r φˆ α j (ξ ) s1 ν+s−β r −r ˆ + ξ Dξ φα, j (ξ ) dξ ξ ν + s μ r |a |csc α||r −r |−|ν+s| ≤ Cα cot α| s r r β r ≤μ r ≤r
s≤r
β≤ν+s
Characterization of spaces of type W and pseudo-differential operators
229
β as cot α Cβ−β ,μ−r 2k−1 1 β β ≤β s1 ≤β ν+s−β r −r ˆ ν+s−β φα, j (ξ ) + ξ k+s1 Dξ ξ ξ r −r φˆ α, j (ξ ) × sup ex p[−(bt)] Dξ ×
ξ ∈Cn
ν+s−β ν+s−β + ξ k Dξ ξ r −r φˆ α j (ξ ) + ξ s1 Dξ ξ r −r φˆ α, j (ξ ) × (1 + |ξ |)m−k−|β−β | e[−x,t csc α+[(b+a0 )t]] dξ. Rn
Now using the Young inequality (1.19), we have ν μ x D Aθ,α φ j (x) x ν + s μ r |csc α||r −r |−|ν+s| |as cot α| ≤ Cα r r β r ≤μ β≤ν+s r ≤r s≤r β x csc α k−1 as1 cot α Cβ−β ,μ−r 2 ex p −M × β (b + a0 ) β ≤β s1 ≤β ν+s−β r −r ˆ ν+s−β φα, j (ξ ) + ξ k+s1 Dξ ξ ξ r −r φˆ α, j (ξ ) × sup ex p[−(bt)] Dξ ξ ∈Cn
k ν+s−β r −r ˆ s1 ν+s−β r −r ˆ + ξ D ξ φα j (ξ ) + ξ Dξ φα, j (ξ ) ξ ξ × (1 + |ξ |)m−k−|β−β | dξ. Rn
Therefore, μ ex p M x csc α Dx Aθ,α φ j (x) (b + a0 ) μ r ν + s |csc α||r −r |−|ν+s| |a ≤ Cα,ν cot α| s r r β r ≤μ β≤ν+s r ≤r s≤r β as cot α Cβ−β ,μ−r 2k−1 sup ex p[−(bt)] × 1 β ξ ∈Cn β ≤β s1 ≤β ν+s−β r −r ˆ ν+s−β × Dξ φα, j (ξ ) + ξ k+s1 Dξ ξ ξ r −r φˆ α, j (ξ ) k ν+s−β r −r ˆ s1 ν+s−β r −r ˆ + ξ Dξ φα, j (ξ ) + ξ Dξ φα, j (ξ ) ξ ξ × (1 + |ξ |)m−k−|β−β | dξ. Rn
230
S. K. Upadhyay et al.
Now using the technique of Gel’fand and Shilov [5, p. 6], as φα, j → 0 when j → ∞ in W (Cn ). Therefore, ν+s−β r −r ˆ ν+s−β φα, j (ξ ) + ξ k+s1 Dξ sup ex p[−(bt)] Dξ ξ ξ r −r φˆ α, j (ξ )
ξ ∈Cn
ν+s−β ν+s−β + ξ k Dξ ξ r −r φˆ α, j (ξ ) + ξ s1 Dξ ξ r −r φˆ α, j (ξ ) → 0,
as j → ∞. This implies that the right hand side of last integral tends to 0 as j → ∞, which proves that Aθ,α φ j → 0 as j → ∞ in W M (Rn ).
Acknowledgments The first author is thankful to DST-CIMS, Banaras Hindu University, Varanasi, India for providing the research facilities and the second author is also thankful to DST-CIMS, Banaras Hindu University, Varanasi, India for awarding the Junior Research Fellowship since December (2012). The authors are thankful to referee for his valuable comments and suggestions regarding this paper.
References 1. Boutet de Monvel, L.: Operateurs pseudo-differentiels analytiques et operateurs d’ordre infini. Ann. Inst. Fourier Grenoble 22, 229–268 (1972) 2. Cappiello, M.: Pseudodifferential parametrices of infinite order for SG-hyperbolic problems. Rend. Sem. Mat. Univ. Pol. Torino 61(4), 411–441 (2003) 3. De Bie, H., De Schepper, N.: Fractional Fourier transforms of hyper complex signals. SIVip 6, 381–388 (2012) 4. Friedman, A.: Generalized Functions and Partial Differential Equations. Prentice Hall, Englewood Cliffs (1963) 5. Gel’fand, I.M., Shilov, G.E.: Generalized functions, Theory of Differential Equations, vol. 3. Academic Press, New York (1967) 6. Hörmander, L.: Linear Partial Differential Operators. Springer, Berlin, Heidelberg, New York (1976) 7. Hörmander, L.: Linear Partial Differential Operators. Actes Congres intern. Math. Tome 1, 121–133 (1970) 8. Kohn, J.J., Nirenberg, N.: On the algebra of pseudo-differential operators. Commun. Pure Appl. Math 18, 269–305 (1965) 9. Pathak, R.S., Prasad, A., Kumar, M.: Fractional Fourier transform of tempered distributions and generalized pseudo-differential operators. J. Pseudo-Differ.Oper. Appl. 3, 239–254 (2012) 10. Prasad, A., Kumar, M.: Product of two generalized pseudo-differential operators involving fractional Fourier transform. J. Pseudo-Differ. Oper. Appl. 2, 355–365 (2011) 11. Upadhyay, S.K.: W-spaces and pseudo-differential operators. Appl. Anal. 82(4), 381–397 (2003) 12. Upadhyay, S.K., Yadav, R.N., Debnath, L.: Infinite pseudo-differential operators on W M (R n ) space. Analysis 32, 163–178 (2012) 13. Wong, M.W.: Introduction to pseudo-differential operators. World Scientific Publishing, Singapore (1991) 14. Zanghirati, L.: Pseudodifferential operators of infinite order and Gevrey classes. Ann. Univ Ferrara, Sez. VII Sc. Mat. 31, 197–219 (1985)