Fractional Evolution Equation With Cauchy Data in L Spaces: Research Open Access
Fractional Evolution Equation With Cauchy Data in L Spaces: Research Open Access
Fractional Evolution Equation With Cauchy Data in L Spaces: Research Open Access
1 Introduction
Today, fractional calculus involves the investigation of so-called integral operators and
fractional derivatives over real or complex domains and their applications. There are many
types of mathematical models that require the use of noninteger order derivatives. Re-
cently, it has been widely used in modeling practical models because of its ability to provide
approximation and ignore the influence of external forces such as in physics, engineer-
ing, mechanics science, biology, and some other areas [1–5]. There are several versions of
noninteger derivatives, but perhaps the two types of derivatives, Caputo and Riemann–
Liouville, are of most interest to mathematicians [6–15].
Let be a bounded domain in Rd (d ≥ 1) with sufficiently smooth boundary ∂. In
this paper, we are interested in the following evolution equation with a time-fractional
derivative:
⎧
⎨ Dα u + u = F(t, x; u), in × (0, T],
C t
(1.1)
⎩u|∂ = 0, in ,
© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit
to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The
images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise
in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright
holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 2 of 22
where F is the source function and u describes the distribution of the temperature at po-
sition x and time t. In (1.1), α ∈ (1, 2) is the fractional order and C Dαt denotes the Caputo
fractional derivative with respect to t (time-fractional) and is defined by (see [16, 17])
⎧ t
⎨ Dα u(t, x) = 1
(t
2
– z)1–α ∂∂z2u (z, x) dz, for 1 < α < 2,
C t Γ (2–α) 0
(1.3)
⎩C Dα u(t, x) = α
∂ u
(t, x), for α = 1, 2,
t ∂t α
and Γ is the gamma function. Note that if α = 2, then Eq. (1.1) represents a Cauchy prob-
lem for elliptic equation. There are some main reasons why we are interested in studying
Problem (1.1)–(1.2)
⎧
⎪
⎨ut + u = F(t, x; u), in × (0, T],
⎪
u|∂ = 0, in , (1.4)
⎪
⎪
⎩
u(x, 0) = g(x), in .
By changing the variable v(x, t) = u(x, T –t), we turn this problem into a final value problem
⎧
⎪
⎪vt – u = F(t, x; v), in × (0, T],
⎨
v|∂ = 0, in , (1.5)
⎪
⎪
⎩
v(x, T) = g(x), in .
The backward in time problem for parabolic equation introduced as above is one of the
classic inverse problems involving many applications and has been studied extensively
over the past 50 years. Interesting work on Problem (1.5) can be consulted by (1.5).
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 3 of 22
Let us continue to add justification to our interest in elliptic equations with the Caputo
derivative. We observe that the elliptic problem with the Caputo derivative is a subbranch
of the elliptic equation with a nonlocal condition. According to the work of Bitsadze and
Samarskii [23], the nonlocal elliptic equations have many applications in the theory of
plasma. Furthermore, it is difficult for elliptic equations with classical derivatives to de-
scribe the past process. When the phenomenon needs to involve factors and past informa-
tion, the Caputo derivative plays an important role. In addition, many problems in dynami-
cal processes, electrochemistry, and signal processing lead to elliptic differential equations
of fractional order.
• In the interesting paper, Jin and Rundell [24] considered the ill-posedness of Problem
(1.1)–(1.2) in the simple case F = 0; however, the authors have not provided the approx-
imation and error estimate. The first results investigating the above equation seem to be
of [25] and [26]. Our present paper has generalized the previous results by [25] and [26]
in the sense that the observed data belongs to Lp . It is a strong claim that our paper is part
of a series of investigations on the ill-posedness of fractional diffusion equations, which
have been published in a recent series of works by Baleanu and colleagues [27–29].
2 Preliminary results
This section provides some notation and the functional spaces which will be used through-
out this article. Recall that the spectral problem
⎧
⎨e (x) = –λ e (x), x ∈ ,
j j j
⎩ej (x) = 0, x ∈ ∂,
Definition 2.1 (Hilbert scale space) We recall the Hilbert scale space Hs () given as fol-
lows:
2
∞
H () = f ∈ L ()
s 2 2s
λj f (x)ej (x) dx < ∞
j=1
for any s ≥ 0. It is well known that Hs () is a Hilbert space corresponding to the norm
∞
2 1/2
f Hs () = λ2s
j f (x)ej (x) dx , f ∈ Hs ().
j=1
d 2d
Lp () → Hμ () if < μ ≤ 0 and p ≥ ,
4 d – 4μ
(2.1)
d 2d
H () → L () if 0 ≤ s < and p ≤
s p
.
4 d – 4s
Also, for M, n > 0, we introduce the n-order Gevrey class GnM () of L2 -functions, see
e.g. [44], defined by the spectrum of the Laplacian as follows:
∞ 2
1
GnM () := ψ ∈ L () : 2
λnj exp
2Mλj ψ(x)ej (x) dx < ∞ ,
α
j=1
∞
zm
Eα,θ (z) = , z ∈ C, (2.2)
m=0
Γ (αm + θ )
The following lemmas provide upper and lower bounds of the Mittag-Leffler functions
Eα,1 (z), Eα,2 (z), Eα,α (z) by the exponential functions.
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 5 of 22
Lemma 2.2 (see [45]) Let 0 < α0 < α1 < 2 and α ∈ [α0 , α1 ]. Then there exist constants
M1 , M2 > 0 and z > 0 such that
M1 1 M2 1
exp z α ≤ Eα,1 (z) ≤ exp z α ,
α α
1 1
M1 exp(z α ) M2 exp(z α )
1 ≤ Eα,2 (z) ≤ 1 ,
α zα α zα
1 1
M1 exp(z α ) M2 exp(z α )
≤ Eα,α (z) ≤ .
α z1– α1 α z1– α1
Lemma 2.3 (see [26]) Let α ∈ [α0 , α1 ] with 1 < α0 < α1 < 2 and z ∈ [0, T]. Then there exists
a positive constant M > 0 independent of z such that
M 1
Eα,1 λj zα ≤ exp λjα z , (2.3)
α
M –1 1
zEα,2 λj zα ≤ λj α exp λjα z , (2.4)
α
M 1–α 1
zα–1 Eα,α λj zα ≤ λj α exp λjα z . (2.5)
α
Next, let us give the explicit fomula of the mild solution to Problem (1.1)–(1.2). Sup-
pose that Problem (1.1)–(1.2) has a solution u(t, x) = ∞ j=1 [ u(t, x)ej (x) dx]ej (x). Then the
function uj (t) = u(t, x)ej (x) dx solves the following ordinary differential equation:
⎧
⎪
⎪ D α
u (t) – λ u (t) = G(t, x, u(t, x))ej (x) dx,
⎨ C t j
j j
By applying the method in [17, Sect. 2], we obtain the solution of (2.6) as follows:
uj (t) = Eα,1 λj t α ψ(x)ej (x) dx + tEα,2 λj t α ϕ(x)ej (x) dx
t
α–1
α
+ (t – ν) Eα,α λj (t – ν) G ν, x, u(ν, x) ej (x) dx dν.
0
∞
u(t, x) = Eα,1 λj t α ψ(x)ej (x) dx ej (x)
j=1
∞
α
+ tEα,2 λj t ϕ(x)ej (x) dx ej (x) (2.7)
j=1
∞
t
+ (t – ν)α–1 Eα,α λj (t – ν)α G ν, x, u(ν, x) ej (x) dx dν ej (x).
j=1 0
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 6 of 22
⎧
⎪
⎪
α
in × (0, T],
⎨C Dt u + u = 0,
u|∂ = 0, in , (3.1)
⎪
⎪
⎩
u(x, 0) = ψ(x), ut (x, 0) = ϕ(x), in .
The solution of the homogeneous problem has the following series representation:
∞
u(t, x) = Eα,1 λj t α ψ(x)ej (x) dx ej (x)
j=1
(3.2)
∞
α
+ tEα,2 λj t ϕ(x)ej (x) dx ej (x).
j=1
D
(β)
Wβ (t, x) = Eα,1 λj t α ψβ (x)ej (x) dx ej (x)
j=1
(3.3)
D
(β)
+ tEα,2 λj t α ϕβ (x)ej (x) dx ej (x),
j=1
Theorem 3.1 Let β, δ > 0, 0 < p < d4 , and 1 < m < 2. We presume that the Cauchy data
(ψ, ϕ) is disturbed by the noisy data (ψβ , ϕβ ) ∈ Lm () × Lm () such that
1 1 2p – m–2
u(t, ·) – Wβ (t, ·) 2d ≤ (T + 1)C1 C(d, m, p) exp |C| α T D(β) αd D(β) d 2m β
L d–4p ()
– 2δ
C )–δ D(β) d uL∞ (0,T;Hp+δ ()) ,
+ C(d, p)(
dm–2d
Proof Since 1 < m < 2, we find that Lm () → H 4m (). This embedding allows us to give
that
ψ – ψβ dm–2d + ϕ – ϕβ dm–2d
H 4m () H 4m ()
≤ βC(d, m).
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 7 of 22
It is obvious that
u(t, ·) – Wβ (t, ·) p ≤ u(t, ·) – vβ (t, ·) p + vβ (t, ·) – Wβ (t, ·) p , (3.6)
H () H () H ()
D
(β)
α
vβ (t, x) = Eα,1 λj t ψ(x)ej (x) dx ej (x)
j=1
(3.7)
D
(β)
α
+ tEα,2 λj t ϕ(x)ej (x) dx ej (x).
j=1
In view of Parseval’s equality, the second quantity on the right-hand side of (3.6) is
bounded by
vβ (t, ·) – Wβ (t, ·) p ≤ J1 (t) + J2 (t), (3.8)
H ()
where
D(β)
α
J1 (t) = Eα,1 λj t ψβ (x) – ψ(x) ej (x) dx ej (x) (3.9)
j=1 Hp ()
and
D(β)
α
J2 (t) = tEα,2 λj t ϕβ (x) – ϕ(x) ej (x) dx ej (x) . (3.10)
j=1 Hp ()
Our upcoming task is to provide the upper bounds for the two components J1 (t) and
J2 (t). By looking at Parseval’s equality for the term J1 (t), we have the following equality:
2
D
(β)
2p 2
J1 (t)2 = λj Eα,1 λj t α ψβ (x) – ψ(x) ej (x) dx
j=1
D
2
(β)
2p– dm–2d dm–2d
= λj 2m Eα,1 λj t α 2 λ 2m ψβ (x) – ψ(x) ej (x) dx . (3.11)
j
j=1
2 4p m–2
J1 (t)2 ≤ |C1 |2 exp 2|C| α1 T D(β) αd D(β) d – m
D
(β) 2
dm–2d
× λj 2m
ψβ (x) – ψ(x) ej (x) dx
j=1
2 4p – m–2
≤ |C1 |2 exp 2|C| α T D(β) αd D(β) d m ψ – ψβ 2
1
dm–2d , (3.15)
H 4m ()
dm–2d
where |C1 |2 = | M
α
|2 |C|2p– 2m . Using (3.5), we find that
1 1 2p – m–2
J1 (t) ≤ C1 exp |C| α T D(β) αd D(β) d 2m ψ – ψβ dm–2d
H 4m ()
1 1 2p – m–2
≤ C1 C(d, m) exp |C| α T D(β) αd D(β) d 2m β. (3.16)
2
D
(β)
2p 2
J2 (t)2 = λj t 2 Eα,2 λj t α ϕβ (x) – ϕ(x) ej (x) dx
j=1
D
2
(β)
2p– dm–2d
2m
2 dm–2d
=t 2
λj Eα,2 λj t α λj 2m ϕβ (x) – ϕ(x) ej (x) dx
j=1
2 4p – m–2
≤ T 2 |C1 |2 exp 2|C| α T D(β) αd D(β) d m ϕ – ϕβ 2
1
dm–2d . (3.17)
H 4m ()
1 1 2p – m–2
J2 (t) ≤ TC1 C(d, m) exp |C| α T D(β) αd D(β) d 2m β. (3.18)
vβ (t, ·) – Wβ (t, ·) p
H ()
(3.19)
1 1 2p – m–2
≤ (T + 1)C1 C(d, m) exp |C| α T D(β) αd D(β) d 2m β.
We continue by dealing with the first term on the right-hand side of (3.6). It is clear to see
that
2
u(t, ·) – vβ (t, ·)2 p = λ –2δ 2p+2δ
λ u(t, x)ej (x) dx . (3.20)
H () j j
j>D (β)
2/d
λj ≥ C D(β) ,
C j2/d >
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 9 of 22
≤ (
4δ
we find that λ–2δ
j C)–2δ |D(β)|– d , which allows us to provide that
u(t, ·) – vβ (t, ·)2 p
H ()
2
– 4δ 2p+2δ
C )–2δ D(β) d
≤ ( λj u(t, x)ej (x) dx
j>D (β)
– 4δ 2 – 4δ
C )–2δ D(β) d u(t, ·)Hp+δ () ≤ (
≤ ( C )–2δ D(β) d u2L∞ (0,T;Hp+δ ()) . (3.21)
– 2δ
u(t, ·) – vβ (t, ·) p ≤ (
C )–δ D(β) d uL∞ (0,T;Hp+δ ()) . (3.22)
H ()
u(t, ·) – Wβ (t, ·) p ≤ vβ (t, ·) – Wβ (t, ·) p + u(t, ·) – vβ (t, ·) p
H () H () H ()
1 1 2p – m–2
≤ (T + 1)C1 C(d, m) exp |C| α T D(β) αd D(β) d 2m β
– 2δ
C )–δ D(β) d uL∞ (0,T;Hp+δ ()) .
+ ( (3.23)
2d
Hp () → L d–4p ().
u(t, ·) – Wβ (t, ·) 2d ≤ C(d, p)u(t, ·) – Wβ (t, ·)Hp ()
L d–4p ()
1 1 2p – m–2
≤ (T + 1)C1 C(d, m, p) exp |C| α T D(β) αd D(β) d 2m β
– 2δ
C )–δ D(β) d uL∞ (0,T;Hp+δ ()) .
+ C(d, p)(
⎧
⎪
⎪
α
in × (0, T],
⎨C Dt u + u = G(x, t),
u|∂ = 0, in , (4.1)
⎪
⎪
⎩
u(x, 0) = ψ(x), ut (x, 0) = ϕ(x), in .
∞
u(t, x) = Eα,1 λj t α ψ(x)ej (x) dx ej (x)
j=1
∞
+ tEα,2 λj t α ϕ(x)ej (x) dx ej (x)
j=1
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 10 of 22
∞
t
+ (t – ν)α–1 Eα,α λj (t – ν)α G(ν, x)ej (x) dx dν ej (x).
j=1 0
D
(β)
α
Vβ (t, x) = Eα,1 λj t ψβ (x)ej (x) dx ej (x)
j=1
D
(β)
+ tEα,2 λj t α ϕβ (x)ej (x) dx ej (x)
j=1
D (β) t
α–1 α
+ (t – ν) Eα,α λj (t – ν) Gβ (ν, x)ej (x) dx dν ej (x).
j=1 0
Theorem 4.1 Let β, δ > 0, 0 < p < d4 , and 1 < m < 2. We presume that the Cauchy data
(ψ, ϕ, G) is disturbed by the noisy data (ψβ , ϕβ , Gβ ) ∈ Lm () × Lm () × L2 (0, T; Lm ())
such that
1 1 2p – m–2
u(t, ·) – Vβ (t, ·) 2d exp |C| α T D(β) αd D(β) d 2m β
L d–4p ()
CM 1 2 2p + 2–2α – m–2
+ exp |C| α T D(β) αd D(β) d dα 2m β
α
– 2δ
C)–δ D(β) d uL∞ (0,T;Hp+δ ()) .
+ C(d, p)(
D
(β)
vβ (t, x) = Eα,1 λj t α ψ(x)ej (x) dx ej (x)
j=1
D
(β)
+ tEα,2 λj t α ϕ(x)ej (x) dx ej (x)
j=1
D (β) t
+ (t – ν)α–1 Eα,α λj (t – ν)α G(ν, x)ej (x) dx dν ej (x).
j=1 0
vβ (t, ·) – Vβ (t, ·) p ≤ J1 (t) + J2 (t) + J3 (t), (4.2)
H ()
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 11 of 22
where J1 , J2 are defined respectively in (3.9), (3.10), and the third term J3 is defined by
D(β)
t
J3 (t) = (t – ν)α–1 Eα,α λj (t – ν)α
0
j=1
(4.3)
× Gβ (ν, x) – G(ν, x) ej (x) dx dν ej (x) .
Hp ()
M 1–α 1
(t – ν)α–1 Eα,α λj (t – ν)α ≤ λj α exp λjα (t – ν) . (4.4)
α
This inequality together with Parseval’s equality leads to
2
D
(β) t
J3 (t)2 = 2p
λj (t – ν)α–1 Eα,α λj (t – ν)α Gβ (ν, x) – G(ν, x) ej (x) dx dν
j=1 0
2 D(β)
M 2p+ 2–2α
≤ λ α
α j=1 j
2
t α1
× exp 2λj (t – ν) Gβ (ν, x) – G(ν, x) ej (x) dx dν . (4.5)
0
1 2
exp 2λjα (t – ν) ≤ exp 2|C| α T D(β) αd .
1
(4.6)
which leads to
2
2 4p 4–4α m–2
J3 (t)2 ≤ CM exp 2|C| α1 T D(β) αd D(β) d + dα – m
α
T
× Gβ (ν, ·) – G(ν, ·)2 dm–2d dν
0 H 4m ()
CM 2 2 4p 4–4α m–2
≤ exp 2|C| α1 T D(β) αd D(β) d + dα – m
α
dm–2d
L2 0, T; Lm () → L2 0, T; H 4m () ,
we derive that
Gβ – G2 dm–2d ≤ C(d, m)Gβ – G2L2 (0,T;Lm ()) ≤ β 2 C(d, m). (4.9)
L2 (0,T;H 4m ())
2 2p 2–2α m–2
J3 (t) CM exp |C| α1 T D(β) αd D(β) d + dα – 2m β. (4.10)
α
Here, the hidden constant depends on d, m. Combining (3.16), (3.18), (4.10) and looking
at (4.2), we arrive at
1 2p m–2
vβ (t, ·) – Vβ (t, ·) p exp |C| α1 T D(β) αd D(β) d – 2m β
H ()
CM 1 2 2p + 2–2α – m–2
+ exp |C| α T D(β) αd D(β) d dα 2m β. (4.11)
α
– 2δ
u(t, ·) – vβ (t, ·) p ≤ (
C )–δ D(β) d uL∞ (0,T;Hp+δ ()) . (4.12)
H ()
CM 1 2 2p + 2–2α – m–2
+ exp |C| α T D(β) αd D(β) d dα 2m β
α
– 2δ
C )–δ D(β) d uL∞ (0,T;Hp+δ ()) .
+ C(d, p)(
∞
P1 (t)v = Eα,1 λj t α v(x)ej (x) dx ej (x) (5.1)
j=1
and
∞
P2 (t)v = tEα,2 λj t α v(x)ej (x) dx ej (x), (5.2)
j=1
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 13 of 22
and
∞
P3 (t)v = t α–1 Eα,α λj t α v(x)ej (x) dx ej (x) (5.3)
j=1
for any v ∈ L2 (). We rewrite the mild solution as the following form:
t
u(t, x) = P1 (t)ψ + P2 (t)ϕ + (t – ν)α–1 P3 (t – ν)G ν, x, u(ν, x) dν. (5.4)
0
By using the Fourier method, we approximate Problem (2.7) by a new integral equation.
Before providing this integral equation, we need to introduce the operator SR , which is the
orthogonal projection onto the eigenspace span {ej , λj ≤ R} for any R > 0. Let any function
θ ∈ L2 (), then we provide the truncation operator SR θ as follows:
λj ≤R
SR θ = θ (x)ej (x) dx ej (x). (5.5)
j=1
Let us assume that the Cauchy data (ψ, ϕ) is noisy by the observation data (ψβ , ϕβ ) ∈
Lp () × Lp () such that
Since this observation data (ψβ , ϕβ ), we can construct a regularized solution as follows:
t
Zβ (t, x) = P1 (t)SRβ ψβ + P2 (t)SRβ ϕβ + P3 (t – ν)SRβ G ν, x, Zβ (ν, x) dν. (5.7)
0
Integral equation as above is called “regularized problem”. Our main purpose in this sec-
tion is to
• Show the existence and uniqueness of the mild solution to regularized problem (5.7);
• Estimate the upper bound of the regularized solution Zβ and the sought solution u on
Lm space.
Theorem 5.1 Let the observed data (ψβ , ϕβ ) ∈ Lp () × Lp (). Then Problem (5.7) has a
2d
unique solution Zβ ∈ L∞
a (0, T; L
d–4r ()). In addition, we get the following bound:
M dp–2d 1
≤ 2 (Rβ ) 4p exp Rβα T ψβ Lp ()
r–
Zβ ∞ 2d
La (0,T;L d–4r ()) α
–1/α M r–
dp–2d 1
+ 2λ1 (Rβ ) 4p exp Rβα T ϕβ Lp () (5.8)
α
1
for any a > Rβα and 0 ≤ r < d4 .
Proof Let w ∈ Hq (). Then, for any q ≥ q, we have the following estimate by a simple
computation:
2 12
2q –2q 2 2q
P1 (t)SR w q = λj Eα,1 λj t α λj w(x)ej (x) dx . (5.9)
β H ()
λj ≤Rβ
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 14 of 22
and
1–α 1
P3 (t)SR w q ≤ λ α M (Rβ )q –q exp R α t wHq () . (5.13)
β H () 1 α β
2d
Let a > 0 be a real constant. Denote by L∞ a (0, T; L
d–4r ()) the subspace of L∞ (0, T;
2d
L d–4r ()) associated with the following norm:
f a := max exp –a(T – t) f (·, t)
2d
2d , ∀f ∈ L∞ 0, T; L d–4r () .
0≤t≤T L d–4r ()
2d 2d
Let us define a nonlinear map M : L∞
a (0, T; L
d–4r ()) → L∞ (0, T; L d–4r ()) by
a
t
Mz(t, x) = P1 (t)SRβ ψβ + P2 (t)SRβ ϕβ + P3 (t – ν)SRβ G ν, x, z(ν, x) dν. (5.14)
0
2d
We aim to show that M(z = 0) ∈ L∞
a (0, T; L
d–4r ()) for 0 ≤ r < d . In view of the embedding
4
2d
Hr () → L d–4r () and (5.11) with q = r ≥ 0 and q = dp–2d
4p
< 0, we find that
P1 (t)SR ψβ 2d P1 (t)SRβ ψβ Hr ()
β
L d–4r ()
M dp–2d 1
≤ (Rβ ) 4p exp Rβα t ψβ
r–
dp–2d .
α H 4p ()
dp–2d
Lp () → H 4p () (5.15)
2d
From the two observations above, we can deduce that M(z = 0) ∈ L∞ a (0, T; L
d–4r ()).
2d
Let any two functions z1 , z2 ∈ L∞
a (0, T; L d–4r ()) for any a > 0. Then we have the follow-
ing identity:
t
Mz1 (t, x) – Mz2 (t, x) = P3 (t – ν)SRβ G ν, x, z1 (ν, x) – G ν, x, z2 (ν, x) dν. (5.18)
0
P3 (t – ν)SR G ν, x, z1 (ν, x) – G ν, x, z2 (ν, x) r
β H ()
1–α M 1
≤ λ1 α (Rβ )q –q exp Rβα (t – ν) G ν, ·, z1 (ν, ·) – G ν, ·, z2 (ν, ·) L2 ()
α
1–α M 1
≤ Lg λ1 (Rβ )q –q exp Rβα (t – ν) z1 (ν, ·) – z2 (ν, ·)L2 ()
α
α
1–α M 1
≤ Lg λ1 α (Rβ )q –q exp Rβα (t – ν) z1 (ν, ·) – z2 (ν, ·) 2d , (5.19)
α L d–4r ()
where we have used the global Lipschitz function of G and the Sobolev embedding
2d
L2 () → L d–4r (). It follows from (5.18) that
exp(–at)Mz1 (t, ·) – Mz2 (t, ·)Hr ()
1–α
MLg λ1 α (Rβ )q –q
≤
α
t
1
× exp Rβα – a (t – ν) exp(–aν)z1 (ν, ·) – z2 (ν, ·) 2d dν. (5.20)
0 L d–4r ()
Recall that
z1 – z2 2d := ess sup exp(–aν)z1 (ν, ·) – z2 (ν, ·) 2d ,
L∞
a (0,T;L d–4r ()) 0<ν<T L d–4r ()
and we get
exp(–at)Mz1 (t, ·) – Mz2 (t, ·) 2d
L d–4r ()
exp(–at)Mz1 (t, ·) – Mz2 (t, ·)Hr ()
1–α
MLg λ1 α (Rβ )q –q t 1
≤ exp Rβ – a (t – ν) dν z1 – z2
α
2d
α 0 L∞
a (0,T;L d–4r ())
1–α
MLg λ1 α (Rβ )q –q
≤ z1 – z2 ∞ 2d ,
aα La (0,T;L d–4r ())
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 16 of 22
1 2d
for any a > Rβα . Also we apply the embedding Hr () → L d–4r () to find that
1–α
MLg λ1 α (Rβ )q –q
Mz1 – Mz2 ∞ 2d ≤ z1 – z2 ∞ 2d . (5.21)
La (0,T;L d–4r ()) aα La (0,T;L d–4r ())
1–α
2MLg λ1 α (Rβ )q –q
a≥ ,
α
1–α
MLg λ1 α (Rβ )q –q
we know that aα
≤ 1/2. Hence, together with the fact that
2d
M(z = 0) ∈ L∞
a 0, T; L
d–4r () ,
2d
we can say that M is a contraction in L∞a (0, T; L
d–4r ()). By applying the Banach fixed
2d
point theorem, we can deduce that M has a fixed point Zβ ∈ L∞ a (0, T; L
d–4r ()). In the
1
MZβ – P1 (t)SR ψβ – P2 (t)SR ϕβ 2d ≤ Zβ ∞ 2d . (5.22)
L∞
β β
a (0,T;L d–4r ()) 2 La (0,T;L d–4r ())
Due to the fact that MZβ = Zβ and combining (5.16) and (5.17), we find that
Zβ 2d
L∞
a (0,T;L d–4r ())
= MZβ 2d
L∞
a (0,T;L d–4r ())
1
≤ P1 (t)SRβ ψβ + P2 (t)SRβ ϕβ 2d + Zβ ∞ 2d
L∞
a (0,T;L d–4r ()) 2 La (0,T;L d–4r ())
M dp–2d 1
≤ (Rβ ) 4p exp Rβα T ψβ Lp ()
r–
α
–1/α M r–
dp–2d 1 1
+ λ1 (Rβ ) 4p exp Rβα T ϕβ Lp () + Zβ ∞ 2d .
α 2 La (0,T;L d–4r ())
The following theorem provides the error between the regularized solution and the ex-
act solution in the space of Lp type when the noisy data in Lp .
Theorem 5.2 Let us assume that Problem (1.1)–(1.2) has a unique solution u ∈ L∞ (0, T;
Hr+δ ()) ∩ L∞ (0, T; GsMT ()) for any δ > 0 and 0 ≤ r < d4 and s > r. Let the observed data
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 17 of 22
(ψβ , ϕβ ) ∈ Lp () × Lp () satisfy (5.6) and 1 < p < 2. Let us choose Rβ such that
dp–2d
r– 4p 1
lim Rβ = ∞, lim (Rβ ) exp Rβα T β = 0. (5.23)
β→0 β→0
dp–2d 1
Zβ (t, ·) – u(t, ·) 2d
r–
max (Rβ ) 4p exp Rβα T β, (Rβ )r–s , (Rβ )–δ . (5.24)
L d–4r ()
–δ
r–s
α(r– dp–2d
4p )
1 1 1
max log , log , log βμ .
β β β
t
SR u(t, x) = P1 (t)SR ψ + P2 (t)SR ϕ + (t – ν)α–1 P3 (t – ν)SR G ν, x, u(ν, x) dν. (5.25)
0
Zβ (t, ·) – u(t, ·) 2 ≤ Zβ (t, ·) – SR u(t, ·) 2 + SR u(t, ·) – u(t, ·) 2
L () β L () β L ()
Let us first treat the term (I). Indeed, by the definition of Zβ as in (5.7) and (5.25), we arrive
at
Zβ (t, ·) – SR u(t, ·) 2
β L ()
≤ P1 (t)SRβ ψβ – P1 (t)SRβ ψ L2 () + P2 (t)SR ϕβ – P2 (t)SR ϕ L2 ()
t
+ (t – ν)α–1 P3 (t – ν)SRβ G ν, x, Zβ (ν, x) – G ν, x, u(ν, x) dν. (5.27)
0
P1 (t)SR ψβ – P1 (t)SR ψ 2
β β L ()
≤ P1 (t)SRβ ψβ – P1 (t)SRβ ψ 2d
L d–4r ()
M dp–2d 1
P1 (t)SRβ (ψβ – ψ)Hr () ≤ (Rβ ) 4p exp Rβα t ψβ – ψ dp–2d
–
α H 4p ()
M dp–2d 1 M dp–2d 1
(Rβ ) 4p exp Rβα t ψβ – ψLp () (Rβ ) 4p exp Rβα t β,
– –
α α
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 18 of 22
where we have used (5.15) to get the last estimate. By a similar explanation, we obtain the
estimate for the second term on the right-hand side of (5.27) as follows:
P2 (t)SR ϕβ – P2 (t)SR ϕ 2
β β L ()
≤ P2 (t)SRβ ϕβ – P2 (t)SRβ ϕ 2d
L d–4r ()
P2 (t)SRβ ϕβ – P2 (t)SRβ ϕ Hr ()
–1/α M –
dp–2d 1
≤ λ1 (Rβ ) 4p exp Rβα t ϕβ – ϕ dp–2d
α H 4p ()
–1/α M
dp–2d
– 4p α1
–1/α M –
dp–2d 1
λ1 (Rβ ) exp Rβ t ϕβ – ϕL () λ1 (Rβ ) 4p exp Rβα t β.
p
α α
We continue to consider the third term on the right-hand side of (5.27). Indeed, we get
t
(t – ν)α–1 P3 (t – ν)SR G ν, x, Zβ (ν, x) – G ν, x, u(ν, x) dν
0 L2 ()
t
1–α M 1
≤ λ1 α exp R α (t – ν) G ν, ·, Zβ (ν, ·) – G ν, ·, u(ν, ·) 2 dν
β L ()
α 0
t
1–α M 1
≤ Lg λ1 α exp Rβα (t – ν) Zβ (ν, ·) – u(ν, ·)L2 () dν. (5.28)
α 0
12
1 1 2
u(t) – SR u(t, ·) 2 = exp –2(T – t)λjα λ–2s α 2s
β L () j exp 2(T – t)λj λj u(t), ej
λj >Rβ
1
≤ (Rβ )–s exp –(T – t)Rβα B, (5.30)
M dp–2d 1 1
≤ λ–1/α + 1 (Rβ )– 4p exp R α t β + (Rβ )–s exp –(T – t)R α B
1 α β β
t
1–α M 1
+ Lg λ1 α exp Rβα (t – ν) Zβ (ν, ·) – u(ν, ·)L2 () dν. (5.32)
α 0
1
Multiplying both sides of the above expression by exp(Rβα (T – t)), we obtain that
1
exp Rβα (T – t) Zβ (t, ·) – u(t, ·)L2 ()
M(λ–1/α + 1) –
dp–2d 1
≤ 1
(Rβ ) 4p exp Rβα T β + (Rβ )–s B
α
1–α
MLg λ1 α t 1
+ exp Rβα (T – ν) Zβ (ν, ·) – u(ν, ·)L2 () dν. (5.33)
α 0
1
exp Rβα (T – t) Zβ (t, ·) – u(t, ·)L2 ()
1–α
M(λ–1/α + 1) dp–2d
– 4p α1 MLg λ1 α
≤ 1
(Rβ ) –s
exp Rβ T β + (Rβ ) B exp t . (5.34)
α α
Zβ (t, ·) – u(t, ·) 2
L ()
1 M(λ–1/α + 1) –
dp–2d 1
exp Rβα (t – T) 1
(Rβ ) 4p exp Rβα T β + (Rβ )–s B . (5.35)
α
Our next aim is to considering the quantity Zβ (t, ·) – u(t, ·)Hr () . Using the triangle
inequality, we arrive at
Zβ (t, ·) – u(t, ·) r ≤ Zβ (t, ·) – SR u(t, ·) r + SR u(t, ·) – u(t, ·) r . (5.36)
H () β H () β H ()
≤ (Rβ )r Zβ (t, ·) – SRβ u(t, ·)L2 () . (5.37)
2
= –2δ 2r+2δ
λj λj u(x, t)ej (x) dx
λj >Rβ
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 20 of 22
where
2
∞ 2r+2δ
uL∞ (0,T;Hr+δ ()) ≥ λj u(x, t)ej (x) dx .
j=1
Combining (5.36), (5.37), and (5.38) together with (5.35), we get that
Zβ (t, ·) – u(t, ·) r
H ()
≤ (Rβ )r Zβ (t, ·) – SRβ u(t, ·)L2 () + (Rβ )–δ uL∞ (0,T;Hr+δ ())
α1 M(λ–1/α + 1) dp–2d
r– 4p α1
≤ exp Rβ (t – T) 1
(Rβ ) r–s
exp Rβ T β + (Rβ ) B
α
+ (Rβ )–δ uL∞ (0,T;Hr+δ ()) . (5.39)
2d
Using the embedding Hr () → L d–4r (), we deduce that
Zβ (t, ·) – u(t, ·) 2d Zβ (t, ·) – u(t, ·)Hr ()
L d–4r ()
r–
dp–2d 1
max (Rβ ) 4p exp Rβα T β, (Rβ )r–s , (Rβ )–δ . (5.40)
6 Conclusion
We consider the Cauchy problem for an evolution equation with the Caputo fractional
derivative. The most important feature of the paper is that the problem is not well posed
in the sense of Hadamard. Throughout this work, the homogeneous case, the inhomoge-
neous case, and the nonlinear case are investigated in turn. For each case, we construct
a regularized solution by using the Fourier truncation method. In addition, using Cauchy
data in Lp spaces, p = 2, by some embeddings linking the Hilbert scale-spaces and the
Lebesgue spaces, we derive error estimates between regularized solutions and exact solu-
tions.
Acknowledgements
The author Le Dinh Long is thankful to Van Lang University. This research is supported by Industrial University of Ho Chi
Minh City (IUH) under Grant Number 130/HD-DHCN.
Funding
Not applicable.
Declarations
Competing interests
The authors declare no competing interests.
Author contributions
All authors contributed equally. All the authors read and approved the final manuscript.
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 21 of 22
Author details
1
Industrial University of Ho Chi Minh City, Ho Chi Minh City, Vietnam. 2 Department of Mathematics, Cankaya University,
06530 Balgat, Ankara, Turkey. 3 Institute of Space Sciences, R76900 Magurele-Bucharest, Romania. 4 Department of
Mathematics, Texas AM University-Kingsville, 700 University Blvd., MSC 172, Kingsville, TX, USA. 5 Division of Applied
Mathematics, Science and Technology Advanced Institute, Van Lang University, Ho Chi Minh City, Vietnam. 6 Faculty of
Applied Technology, School of Engineering and Technology, Van Lang University, Ho Chi Minh City, Vietnam.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
1. Yang, X.J., Gao, F., Srivastava, H.M.: Exact travelling wave solutions for the local fractional two-dimensional
Burgers-type equations. Comput. Math. Appl. 73(2), 203–210 (2017)
2. Yang, X.J., Gao, F., Ju, Y., Zhou, H.W.: Fundamental solutions of the general fractional-order diffusion equations. Math.
Methods Appl. Sci. 41(18), 9312–9320 (2018)
3. Yang, X.J., Gao, F., Ju, Y.: General Fractional Derivatives with Applications in Viscoelasticity. Elsevier/Academic Press,
London (2020)
4. Atangana, A., Baleanu, D.: Caputo-Fabrizio derivative applied to groundwater flow within confined aquifer. J. Eng.
Mech. 143(5), D4016005 (2017)
5. Carvalho-Neto, P.M., Planas, G.: Mild solutions to the time fractional Navier–Stokes equations in RN . J. Differ. Equ.
259(7), 2948–2980 (2015)
6. Odibat, Z., Baleanu, D.: Numerical simulation of initial value problems with generalized Caputo-type fractional
derivatives. Appl. Numer. Math. 156, 94–105 (2020)
7. Baleanu, D., Wu, G.C., Zeng, S.D.: Chaos analysis and asymptotic stability of generalized Caputo fractional differential
equations. Chaos Solitons Fractals 102, 99–105 (2017)
8. Tuan, N.H., Au, V.V., Xu, R.: Semilinear Caputo time-fractional pseudo-parabolic equations. Commun. Pure Appl. Anal.
20(2), 583 (2021)
9. Berdyshev, A.S., Kadirkulov, B.J., Nieto, J.J.: Solvability of an elliptic partial differential equation with boundary
condition involving fractional derivatives. Complex Var. Elliptic Equ. 59(5), 680–692 (2014)
10. Vinothkumar, C., Deiveegan, A., Nieto, J.J., Prakash, P.: Similarity solutions of fractional parabolic boundary value
problems with uncertainty. Commun. Nonlinear Sci. Numer. Simul. 102, 105926 (2021)
11. Baleanu, D., Ghassabzade, F.A., Nieto, J.J., Jajarmi, A.: On a new and generalized fractional model for a real cholera
outbreak. Alex. Eng. J. 61(11), 9175–9186 (2022)
12. Mahmudov, N.I., Nieto, J.J.: Exponential stability and stabilization of fractional stochastic degenerate evolution
equations in a Hilbert space: subordinate principle. Evol. Equ. Control Theory 11(6), 1997–2015 (2022).
https://doi.org/10.3934/eect.2022008
13. Patela, R., Shuklab, A., Nieto, J.J., Vijayakumard, V., Jadon, S.S.: New discussion concerning to optimal control for
semilinear population dynamics system in Hilbert spaces. Nonlinear Anal., Model. Control 27, 1–17 (2022)
14. Nguyen, A.T., Caraballo, T., Tuan, N.H.: On the initial value problem for a class of nonlinear biharmonic equation with
time-fractional derivative. Proc. R. Soc. Edinb., Sect. A, Math. 152, 989–1031 (2021)
15. Tuan, N.H., Khoa, V.A., Minh, M.N., Tran, T.: Reconstruction of the electric field of the Helmholtz equation in three
dimensions. J. Comput. Appl. Math. 309, 56–78 (2017)
16. Kian, Y., Yamamoto, M.: On existence and uniqueness of solutions for semilinear fractional wave equations. Fract. Calc.
Appl. Anal. 20(1), 117–138 (2017)
17. Sakamoto, K., Yamamoto, M.: Initial value/boundary value problems for fractional diffusion-wave equations and
applications to some inverse problems. J. Math. Anal. Appl. 382, 426–447 (2011)
18. Khoa, V.A., Truong, M.T.N., Duy, N.H.M., Tuan, N.H.: The Cauchy problem of coupled elliptic sine-Gordon equations
with noise: analysis of a general kernel-based regularization and reliable tools of computing. Comput. Math. Appl.
73(1), 141–162 (2017)
19. Tuan, N.H., Thang, L.D., Lesnic, D.: A new general filter regularization method for Cauchy problems for elliptic
equations with a locally Lipschitz nonlinear source. J. Math. Anal. Appl. 434, 1376–1393 (2016)
20. Amar, M., Andreucci, D., Bisegna, P., Gianni, R.: Exponential asymptotic stability for an elliptic equation with memory
arising in electrical conduction in biological tissues. Eur. J. Appl. Math. 20(5), 431–459 (2009)
21. Turmetov, B., Nazarova, K.: On fractional analogs of Dirichlet and Neumann problems for the Laplace equation.
Mediterr. J. Math. 16(3), Paper No. 59, 17 pp (2019)
22. Turmetov, B.: On some boundary value problems for nonhomogenous polyharmonic equation with boundary
operators of fractional order. Acta Math. Sci. Ser. B Engl. Ed. 36(3), 831–846 (2016)
23. Bicadze, A.V., Samarskii, A.A.: Some elementary generalizations of linear elliptic boundary value problems. Dokl. Akad.
Nauk SSSR 185, 739–740 (1969). (Russian), Engl. Transl. from Russian Soviet Math. Dokl., 10, 1969, 398–400
24. Jin, B., Rundell, W.: A tutorial on inverse problems for anomalous diffusion processes. Inverse Probl. 31(3), 035003
(2015). 40 pp
25. Au, V.V., Phuong, N.D., Tuan, N.H., Zhou, Y.: Some regularization methods for a class of nonlinear fractional evolution
equations. Comput. Math. Appl. 78(5), 1752–1771 (2019)
26. Tuan, N.H., Xuan, T.D., Triet, N.A., Lesnic, D.: On the Cauchy problem for a semilinear fractional elliptic equation. Appl.
Math. Lett. 83, 80–86 (2018)
27. Long, L.D., Luc, N.H., Tatar, S., Balenu, D., Can, N.H.: An inverse source problem for pseudo-parabolic equation with
Caputo derivative. J. Appl. Math. Comput. 68, 739–765 (2022)
28. Luc, N.H., Baleanu, D., Agarwal, R.P., Long, L.D.: Identifying the source function for time fractional diffusion with
non-local in time conditions. Comput. Appl. Math. 40(5), 1–21 (2021)
Phuong et al. Boundary Value Problems (2022) 2022:100 Page 22 of 22
29. Luc, N.H., Long, L.D., Hang, L.T.D., Baleanu, D., Can, N.H.: Identifying the initial condition for space-fractional Sobolev
equation. J. Appl. Anal. Comput. 11(5), 1–21 (2021)
30. Tuan, N.H., Caraballo, T.: On initial and terminal value problems for fractional nonclassical diffusion equations. Proc.
Am. Math. Soc. 149(1), 143–361 (2021)
31. Tuan, N.H., Nguyen, A.T., Yang, C.: Global well-posedness for fractional Sobolev-Galpern type equations. Discrete
Contin. Dyn. Syst. 42(6), 2637–2665 (2022)
32. Tuan, N.A., Tuan, N.H., Yang, C.: On Cauchy problem for fractional parabolic-elliptic Keller-Segel model. Adv. Nonlinear
Anal. 12(1), 97–116 (2023)
33. Tuan, N.H., Au, V.V., Nguyen, A.T.: Mild solutions to a time-fractional Cauchy problem with nonlocal nonlinearity in
Besov spaces. Arch. Math. (Basel) 118(3), 305–314 (2022)
34. Tuan, N.H., Foondun, M., Thach, T.N., Wang, R.: On backward problems for stochastic fractional reaction equations
with standard and fractional Brownian motion. Bull. Sci. Math. 179, Paper No. 103158, 58 pp (2022)
35. Zhao, C., Caraballo, T., Łukaszewicz, G.: Statistical solution and Liouville type theorem for the
Klein-Gordon-Schrödinger equations. J. Differ. Equ. 281, 1–32 (2021)
36. Tuan, N.H., Caraballo, T.: On initial and terminal value problems for fractional nonclassical diffusion equations. Proc.
Am. Math. Soc. 149, 143–161 (2021)
37. Zhao, C., Wang, J., Caraballo, T.: Invariant sample measures and random Liouville type theorem for the
two-dimensional stochastic Navier-Stokes equations. J. Differ. Equ. 317, 474–494 (2022)
38. Tuan, N.A., Caraballo, T., Tuan, N.H.: On the initial value problem for a class of nonlinear biharmonic equation with
time-fractional derivative. Proc. R. Soc. Edinb. A 152, 989–1031 (2022)
39. Caraballo, T., Mchiri, L., Rhaima, M.: Ulam–Hyers–Rassias stability of neutral stochastic functional differential equations.
Stochastics 94, 959–971 (2022)
40. Xu, J., Zhang, Z., Caraballo, T.: Mild solutions to time fractional stochastic 2D-Stokes equations with bounded and
unbounded delay. J. Dyn. Differ. Equ. 34, 583–603 (2022)
41. Caraballo, T., Guo, B., Tuan, N.H., Wang, R.: Asymptotically autonomous robustness of random attractors for a class of
weakly dissipative stochastic wave equations on unbounded domains. Proc. R. Soc. Edinb. A 151, 1700–1730 (2021)
42. Bao, N.T., Caraballo, T., Tuan, N.H., Zhou, Y.: Existence and regularity results for terminal value problem for nonlinear
fractional wave equations. Nonlinearity 34, 1448–1503 (2021)
43. Xu, J., Zhang, Z., Caraballo, T.: Non-autonomous nonlocal partial differential equations with delay and memory. J.
Differ. Equ. 270, 505–546 (2021)
44. Cao, C., Rammaha, M.A., Titi, E.S.: The Navier-Stokes equations on the rotating 2-D sphere: Gevrey regularity and
asymptotic degrees of freedom. Z. Angew. Math. Phys. 50, 341–360 (1999)
45. Dang, D.T., Nane, E., Nguyen, D.M., Tuan, N.H.: Continuity of solutions of a class of fractional equations. Potential Anal.
49(3), 423–478 (2018)