Turkish Journal of Mathematics

Volume 47 Number 5 Article 21

1-1-2023

Boundary value problem for a loaded fractional diffusion equation

ARSEN V. PSKHU

MURAT I. RAMAZANOV

MINZILYA KOSMAKOVA

Follow this and additional works at: https://journals.tubitak.gov.tr/math

Part of the Mathematics Commons

Recommended Citation
PSKHU, ARSEN V.; RAMAZANOV, MURAT I.; and KOSMAKOVA, MINZILYA (2023) "Boundary value problem
for a loaded fractional diffusion equation," Turkish Journal of Mathematics: Vol. 47: No. 5, Article 21.
https://doi.org/10.55730/1300-0098.3450
Available at: https://journals.tubitak.gov.tr/math/vol47/iss5/21

This Article is brought to you for free and open access by TÜBİTAK Academic Journals. It has been accepted for
inclusion in Turkish Journal of Mathematics by an authorized editor of TÜBİTAK Academic Journals. For more
information, please contact [email protected].
 Turkish Journal of Mathematics Turk J Math
(2023) 47: 1585 – 1594
http://journals.tubitak.gov.tr/math/
© TÜBİTAK
Research Article doi:10.55730/1300-0098.3450

Boundary value problem for a loaded fractional diffusion

equation

Arsen V. PSKHU1, Murat I. RAMAZANOV2, Minzilya T. KOSMAKOVA2∗

1
Institute of Applied Mathematics and Automation, Kabardino-Balkarian Scientific C enter o f R AS, N alchik, Russia
2
Buketov Karaganda University, Karaganda, Kazakhstan

Received: 23.11.2021 • Accepted/Published Online: 15.06.2023 • Final Version: 18.07.2023

Abstract: In this paper we consider a boundary value problem for a loaded fractional diffusion equation. The loaded
term has the form of the Riemann-Liouville fractional derivative or integral. The BVP is considered in the open right
upper quadrant. The problem is reduced to an integral equation that, in some cases, belongs to the pseudo-Volterra
type, and its solvability depends on the order of differentiation in the loaded term and the behavior of the support line of
the load in a neighborhood of the origin. All these cases are considered. In particular, we establish sufficient conditions
for the unique solvability of the problem. Moreover, we give an example showing that violation of these conditions can
lead to nonuniqueness of the solution.

Key words: Fractional diffusion equation, loaded equation, fractional derivative, integral equation, Wright function

1. Introduction
Consider the equation
 
∂2  σ 
α
D0y − u(x, y) + λ D0y u(x, y) x=γ(y) = f (x, y), (0 < α ≤ 1, σ ≤ α) (1.1)
∂x2

α σ σ
where D0y (as well as D0y ) stands for the Riemann-Liouvlle fractional derivative (D0y may be a derivative or
integral) with respect to y of order α ∈ (0, 1) (σ ≤ α ) with the origin at the point x = 0 ; γ(y) is a continuous
function that is positive for y > 0.
Operators of fractional order integration and differentiation are defined as follows (see, e.g., [14])
Z y
1
ν
D0y u(x, y) = u(t)(y − t)−ν−1 dt (ν < 0); 0
D0y u(x, y) = u(x, y); (1.2)
Γ(−ν) 0

and
∂ ν−1
ν
D0y u(x, y) = D u(x, y) (0 < ν ≤ 1).
∂y 0y

Over the past three decades, fractional diffusion equations have been the subject of intense research
attention, starting from the works [6, 29]. Papers [1, 3, 4, 11, 12, 16, 20–24] give an idea of the variety of
∗Correspondence:
[email protected]
2010 AMS Mathematics Subject Classification: 35R11; 35K20

1585
This work is licensed under a Creative Commons Attribution 4.0 International License.
 PSKHU et al./Turk J Math

approaches to the study of these equations. Overviews can be found in [4]. The monographs [5, 7, 18] reflect
basic approaches and contain numerous bibliographies concerning the issue.
Equation (1.1) belongs to the class of loaded equations [15]. Of interest are boundary value problems for
the loaded heat equation, when the loaded term is represented in the form of a fractional derivative [8, 9, 25].
In [25] the loaded term is the trace of a fractional derivative on the manifold x = tω . It is proved that there
is continuity in the order of the fractional derivative. In [9] the loaded term is represented in the form of a
fractional derivative with respect to the time variable. In [8] it is shown that the existence and uniqueness of
solutions to the integral equation depend on the order of the fractional derivative in the loaded term. Integral
equations of Volterra type also arise in the study of boundary value problems in degenerate domains [2, 10].
In the present paper, we solve a boundary value problem in the right upper quadrant for the equation (1.1).
The problem is reduced to an integral equation that can belong to the pseudo-Volterra type. We investigate its
solvability that essentially depends on σ and the behavior of γ(y) in a neighborhood of the point y = 0 . The
main results are given in the form of two theorem that establish sufficient conditions for the unique solvability
of the problem, in different cases. In addition, we give an example showing that violation of these conditions
can lead to the nonuniqueness of the solution.
The article is structured as follows. In Section 2, we formulate the problem we are going to solve.
Section 3 contains notations and several auxiliary assertions. In Section 4, the problem is reduced to an integral
equation. In Sections 5 and 6, we consider various cases of the obtained integral equation and formulate the
corresponding results on the solvability of the problem. Section 7 gives an example of a nontrivial solution to
the homogeneous problem.

2. Section
In what follows Ω denotes a domain that coincides with the open right upper quadrant, i.e.

Ω = {(x, y) : x > 0, y > 0}.

We call a function u(x, y) a regular solution of the equation (1.1) in the domain Ω if y 1−µ u(x, y) ∈ C Ω
for some µ > 0; in Ω, u(x, y) has continuous derivatives with respect to x of the first and second order; the
α−1
function D0y u(x, y) is continuously differentiable as function of y for a fixed x at interior points of Ω and is
continuous up to the positive semiaxis {0 < x, y = 0}; and u(x, y) satisfies the equation (1.1) at all points of
Ω.

Problem 2.1 Find a regular solution of the equation (1.1) in the domain Ω satisfying the conditions
α−1
lim D0y u(x, y) = τ (x) (x > 0) (2.1)
y→0

and
u(0, y) = φ(y) (y > 0) (2.2)
where τ (x) and φ(y) are given continuous functions.

3. Preliminaries
Set
    
y µ−1 |x − s| x+s
Gµ (x, s, y) = Φ −β, µ; − − Φ −β, µ; − β . (3.1)
2 yβ y

1586
 PSKHU et al./Turk J Math

Here and in what follows,

α
β= (3.2)
2
and Φ is the Wright function ([27, 28]),
∞
X zk
Φ (a, b; z) = (a > −1).
k! Γ(ak + b)
k=0

Also, we will use the letter C to denote positive constants, which are different in different cases, indicating
in parentheses the parameters on which they depend, if necessary, C = C(α, β, ...) .
It is known that ([19, 28])
   
ρ( 12 −µ) (βt)ρ 1
Φ(−β, µ; −t) = C t exp − 1+O (t → ∞) (3.3)
ρ−1 tρ

where β ∈ (0, 1), ρ−1 = 1 − β ,

d
Φ(−β, µ; x) = Φ(−β, µ − β; x), (3.4)
dx
    
c c
ν
D0y y µ−1 Φ −β, µ; − β = y µ−ν−1 Φ −β, µ − ν; − β (c > 0), (3.5)
y y
and
Z ∞
1
Φ (−β, µ; −x) dx = . (3.6)
0 Γ(β + µ)
The asymptotic expansion (3.3) gives that
 
x
y µ−1 Φ −β, µ; − β ≤ C x−θ y βθ+µ−1 , (3.7)
y

where

0, (−µ) ̸∈ N ∪ {0},
θ≥
−1, (−µ) ∈ N ∪ {0}.

The formulas (3.4), (3.5), and (3.6) yield that

 s=x+ε
ν ∂ y µ−β−1
D0y Gµ (x, s, y) = Gµ−ν (x, s, y), lim Gµ (x, s, y) = , (3.8)
ε→0+ ∂x s=x−ε Γ(µ − β)

   
∂2 ∂2
α
D0y − Gµ (x, s, y) = 0, α
D0y − Gµ (x, s, y) = 0 (x ̸= s), (3.9)
∂x2 ∂s2
and
Z ∞   
1 x
Gµ (x, s, y)ds = y µ+β−1
− Φ −β, µ + β; − β . (3.10)
0 Γ(µ + β) y

Let us recall the form of solutions to the problem (2.1) and (2.2) for the equation (1.1) without the loaded
term, i.e. for the equation
 
∂2
α
D0y − u(x, y) = f (x, y). (3.11)
∂x2

1587
 PSKHU et al./Turk J Math

Lemma 3.1 Let

τ (x) ∈ C[0, ∞), y 1−µ φ(y) ∈ C[0, ∞), y 1−µ f (x, y) ∈ C(Ω), (3.12)

 2
  2

lim τ (x) exp −ωx 2−α = 0, lim y 1−µ f (x, y) exp −ωx 2−α = 0, (3.13)
x→∞ x→∞

for some µ > 0 and every ω > 0 , and let the function f (x, y) can be represented in the form

−δ ∗
f (x, y) = D0y f (x, y) (3.14)

for some δ > β and y 1−µ f ∗ (x, y) ∈ C(Ω) .

Then there exists a unique solution of the problem (3.11), (2.1), and (2.2); it has the form
Z y Z ∞
u(x, y) = f (s, t) Gβ (x, s, y − t) ds dt+ (3.15)
0 0

Z y Z ∞
+ φ(t) G0 (x, 0, y − t) dt + τ (s) Gβ (x, s, y) ds,
0 0

where Gµ (x, s, y) is defined by (3.1).

Proof The fact that a regular solution to the problem (3.11), (2.1), and (2.2) has the form (3.15), can be
adapted from [17, Theorem 8.1]. In particular, this presentation proves the uniqueness of the problem. A direct
verification shows that the conditions (3.12), (3.13), and (3.14) are sufficient for the function (3.15) to be a
solution of (3.11), (2.1), and (2.2). For example, by (3.8), (3.9), and (3.14), it is easy to see that

Z y Z ∞ Z y Z ∞
∂2 ∂ ∂
f (s, t) Gβ (x, s, y − t) ds dt = f ∗ (s, t) Gβ+µ (x, s, y − t) ds dt =
∂x2 0 0 ∂x 0 0 ∂x

Z Z Z ∞
y
(y − t)µ−1 ∗ y
=− f (x, t) ds dt + f ∗ (s, t)Gµ−β (x, s, y − t) ds dt =
0 Γ(µ) 0 0

Z y Z ∞
= −f (x, y) + D0y
α
f (s, t)Gβ (x, s, y − t) ds dt.
0 0

This proves that

 Z y Z ∞
∂2
D0y −
α
f (s, t) Gβ (x, s, y − t) ds dt = f (x, y).
∂x2 0 0

Remark 3.2 It should be noted that condition (3.14) is superfluous if the function f (x, y) does not depend on
x . Indeed, if f (x, y) = f (y) then, by (3.10), we have
Z y Z ∞ Z y  
−α x
f (s, t) Gβ (x, s, y − t) ds dt = D0y f (y) − f (t) (y − t) α−1
Φ −β, α; − dt.
0 0 0 (y − t)β

1588
 PSKHU et al./Turk J Math

and
 Z y  
∂2 x
α
D0y − f (t) (y − t)α−1 Φ −β, α; − dt = 0.
∂x2 0 (y − t)β
Consequently,
 Z y Z ∞
∂2
α
D0y − f (s) Gβ (x, s, y − t) ds dt = f (y).
∂x2 0 0

4. Reduction to integral equation

Let g(y) denote the loaded term in (1.1), namely
 σ 
g(y) = D0y u(x, y) x=γ(y) (4.1)

By Lemma 3.1, the solution of the equation

 
∂2
α
D0y − u(x, y) = f (x, y) − λg(y)
∂x2

that satisfies (2.1) and (2.2) can be written in the form

Z y Z ∞
u(x, y) = F (x, y) − λ g(t) Gβ (x, s, y − t) ds dt. (4.2)
0 0

where
Z y Z ∞
F (x, y) = f (s, t) Gβ (x, s, y − t) ds dt+
0 0
Z y Z ∞
+ φ(t) G0 (x, 0, y − t) dt + τ (s) Gβ (x, s, y) ds.
0 0

By (1.2), (3.2), and (3.10), the equality (4.2) is equivalent to

Z y  
−α x
u(x, y) = F (x, y) − λD0y g(y) + λ g(t) (y − t)α−1 Φ −β, α; − dt. (4.3)
0 (y − t)β

Formulas (3.5) and (4.1) give

Z y  
γ(y)
g(y) = h(y) − λD0y
σ−α
g(y) + λ g(t) (y − t)α−σ−1 Φ −β, α − σ; − dt, (4.4)
0 (y − t)β

where
 σ−α 
h(y) = D0y F (x, y) x=γ(y) .

Moreover, if the conditions (3.12), (3.13), and (3.14) are satisfied, then

y 1−µ h(y) ∈ C[0, ∞) for some µ > 0. (4.5)

The formula (4.4) is valid for every σ .

It should be noted that, in some cases, the integral equation (4.4) does not belong to the Volterra type,
and its solvability depends on σ and the behavior of γ(y) in a neighborhood of the point y = 0 . We next
consider these cases. In particular, we investigate separately the cases σ < α and σ = α .

1589
 PSKHU et al./Turk J Math

5. The case σ < α

It follows from (1.2) and (4.4) that g(y) is a solution of the integral equation
Z y
g(y) + λ g(t) H(y − t) dt = h(y), (5.1)
0

where
  
1 γ(y)
H(y) = y α−σ−1 − Φ −β, α − σ; − β .
Γ(α − σ) y

Moreover, the inequality (3.7) gives that

|H(y)| ≤ Cy α−σ−1 .

This means that the integral equation (5.1) has a solution that is unique. Moreover, it is easy to check that
this solution is continuous for y > 0 and has a singularity at the point y = 0 , the same as the function h(y)
has. Therefore, y 1−µ g(y) ∈ C[0, ∞) for some µ > 0.
Thus, taking into account Lemma 3.1 and Remark 3.2, we can formulate the result obtained in the
considered case.

Theorem 5.1 Let σ < α and the conditions of Lemma 3.1 imposed on f (x, y), τ (x) , and φ(y) be satisfied.
Then the problem (1.1), (2.1), and (2.2) has a unique regular solution that can be given by (4.2), where g(y) is
the solution of (5.1).

6. The case σ = α
Assuming that λ ̸= −1, from (4.4), we get
Z y  
λ g(t) γ(y)
g(y) = h(y) + Φ −β, 0; − dt. (6.1)
1+λ 0 y−t (y − t)β

By (3.7) we have
 
1 γ(y) −θ
Φ −β, 0; − ≤ C [γ(y)] (y − t)βθ−1 (θ ≥ −1). (6.2)
y−t (y − t)β

(Notice that Φ (−β, δ; −x) is positive for x > 0 and δ ≥ 0, [13, 26].)
Take θ = −1 in (6.2) and let
γ(y) ≥ Cy δ . (6.3)

Then, it follows from (6.2) that

 
1 γ(y)
Φ −β, 0; − ≤ Cy −δ (y − t)β−1 . (6.4)
y−t (y − t)β

(Note that, according to our convention above, in the last two inequalities, the letter C denotes different
constants.)

1590
 PSKHU et al./Turk J Math

Let K denote the integral operator in (6.1), i.e.

Z y  
g(t) γ(y)
K g(y) = Φ −β, 0; − dt. (6.5)
0 y−t (y − t)β

By (1.2), (4.5), and (6.4), we get

−β µ−1 Γ(µ)
K h(y) ≤ Cy −δ D0y y ≤C · y µ+β−δ−1
Γ(µ + β)

and, consequently,

 n Y
n−1
Γ(µ + k(β − δ))
−β
K n h(y) ≤ C n y −δ D0y y µ−1 ≤ C n · y µ+n(β−δ)−1
Γ(µ + β + k(β − δ))
k=0

This means that if δ < β then the series

∞
X λn
n
K n h(y)
n=0
(1 + λ)

is convergent, and its sum defines a unique solution of (6.1).

Thus, the result for this case can be formulated as follows.

Theorem 6.1 Let σ = α , λ ̸= −1, and (3.12), (3.13), and (3.14) be satisfied. If γ(y) satisfies (6.3) for some
δ < β , then the problem (1.1), (2.1), and (2.2) has a unique regular solution that is given by (4.2), where g(y)
can be found as a solution of (6.1).

7. Spectral case
By (1.2), (3.5), and (6.5), it is easy to check that
Z   Z y  
y
tε γ(y) (y − t)ε γ(y)
Ky =ε
Φ −β, 0; − dt = Φ −β, 0; − β dt =
0 y−t (y − t)β 0 t t
    
−ε−1 1 γ(y) γ(y)
= Γ(ε + 1) D0t Φ −β, 0; − β = Γ(ε + 1) y Φ −β, ε + 1; − β
ε
.
t t t=y y

Letting σ = α and
γ(y) = Ay β (A > 0),

we get that the function

g(y) = y ε

is a solution of the homogeneous equation

Z y  
λ g(t) γ(y)
g(y) = Φ −β, 0; − dt
1+λ 0 y−t (y − t)β

1591
 PSKHU et al./Turk J Math

where λ is defined by the equality

λ
1= Γ(ε + 1) Φ (−β, ε + 1; −A)
1+λ
or
−1
λ = [Γ(ε + 1) Φ (−β, ε + 1; −A) − 1] . (7.1)

It is known ([18, Lemma 2.2.4]) that

d
Φ (−β, δ; −z) < 0 (z > 0, δ ≥ β)
dz

and, moreover,
1
Φ (−β, δ; 0) = .
Γ(δ)

This yields that there exists λ0 < 0 such that for any λ < λ0 one can find ε and A that turn (7.1) into equality.
Thus, by (4.3), we get that, for λ , ε , and A satisfying (7.1), the function
  
1 x
u(x, y) = y ε+α
− Φ −β, ε + α + 1; − β
Γ(ε + α + 1) y

is a solution of the homogeneous problem

 
∂2  α 
α
D0y − u(x, y) + λ D0y u(x, y) x=Ayε = 0,
∂x2

α−1
lim D0y u(x, y) = 0, u(0, y) = 0.
y→0

Acknowledgment
This research was funded by the Science Committee of the Ministry of Education and Science of the Republic
of Kazakhstan (grant no. AP09259780, 2021–2023).

References

[1] Eidelman SD, Kochubei AN. Cauchy problem for fractional diffusion equations. Journal of Differential Equations
2004; 199 (2): 211-255. https://doi.org/10.1016/j.jde.2003.12.002
[2] Jenaliyev MT, Ramazanov MI, Kosmakova MT, Tuleutaeva ZM. On the solution to a two-dimensional
heat conduction problem in a degenerate domain. Eurasian Mathematical Journal 2020; 11 (3): 89–94.
https://doi.org/10.32523/2077-9879-2020-11-3-89-94
[3] Kemppainen J. Boundary behavior of the layer potentials for the time fractional diffusion equation in Lipschitz
domains. Journal of Integral Equations and Applications 2011; 23 (4): 541-563. https://doi.org/10.1216/JIE-2011-
23-4-541
[4] Kilbas AA. Partial fractional differential equations and some of their applications. Analysis 2010; 30 (1): 35-66.
https://doi.org/10.1524/anly.2010.0934
[5] Kilbas AA, Srivastava HM, Trujillo JJ. Theory and applications of fractional differential equations. Book Series:
North-Holland Mathematics Studies, Vol. 204, Amsterdam: Elsevier, 2006.

1592
 PSKHU et al./Turk J Math

[6] Kochubei AN. Diffusion of fractional order. Differential Equations 1990; 26 (4): 660-670.
[7] Kochubei A, Luchko Y. (Eds.) Handbook of fractional calculus with applications. Volume 2: Fractional Differential
Equations, Berlin, Germany: De Gruyter, 2019.
[8] Kosmakova MT, Iskakov SA, Kasymova LZh. To solving the fractionally loaded heat equation. Bulletin of the
Karaganda University-Mathematics 2021; 101 (1): 65-77. https://doi.org/10.31489/2021M1/65-77
[9] Kosmakova MT, Ramazanov MI, Kasymova LZh. To solving the heat equation with fractional load. Lobachevskii
Journal of Mathematics 2021; 42 (12): 2854-2866. https://doi.org/10.1134/S1995080221120210
[10] Kosmakova MT, Ramazanov MI, Tokesheva AS, Khairkulova AA. (2016). On the non-uniqueness of solution to
the homogeneous boundary value problem for the heat conduction equation in an angular domain. Bulletin of
the Karaganda University-Mathematics 2016; 84 (4): 80-87.
[11] Mainardi F. The fundamental solutions for the fractional diffusion-wave equation. Applied Mathematics Letters
1996; 9: 23-28. https://doi.org/10.1016/0893-9659(96)00089-4
[12] Mamchuev Murat O. Solutions of the main boundary value problems for the time-fractional telegraph equa-
tion by the green function method. Fractional Calculus and Applied Analysis 2017; 20 (1): 190-211.
https://doi.org/10.1515/fca-2017-0010
[13] Mikusiński J. On the function whose Laplace transform is exp(−sα λ) , 0 < α < 1 . Studia Mathematica 1959;
18 (2): 191-198. https://doi.org/10.4064/sm-18-2-191-198
[14] Nakhushev AM. Fractional calculus and its applications. Moscow: Fizmatlit, 2003.
[15] Nakhushev AM. Loaded equations and their applications. Moscow: Nauka, 2012.
[16] Pskhu AV. Fractional diffusion equation with discretely distributed differentiation operator. Siberian Electronic
Mathematical Reports 2016; 13: 1078-1098. https://doi.org/10.17377/semi.2016.13.086
[17] Pskhu AV. Green function of the first boundary-value problem for the fractional diffusion wave equation in
a multidimensional rectangular domain. Itogi Nauki i Tekhniki. Seriya ”Sovremennaya Matematika i ee Prilozheniya.
Tematicheskie Obzory” VINITI 2019; 167: 52-61. https://doi.org/10.36535/0233-6723-2019-167-52-61
[18] Pskhu AV. Partial differential equations of fractional order. Moscow: Nauka, 2005.
[19] Pskhu AV. Solution of a boundary value problem for a fractional partial differential equation. Differential Equations
2003; 39 (8): 1150-1158. doi: https://doi.org/10.1023/B:DIEQ.0000011289.79263.02
[20] Pskhu AV. Solution of boundary value problems for the fractional diffusion equation by the Green function method.
Differential Equations 2003; 39 (10): 1509-1513. https://doi.org/10.1023/B:DIEQ.0000017925.68789.e9
[21] Pskhu AV. Solution of the first boundary value problem for a fractional-order diffusion equation. Differential
Equations 2003; 39 (9): 1359-1363. https://doi.org/10.1023/B:DIEQ.0000012703.45373.aa
[22] Pskhu AV. Stabilization of solutions to the Cauchy problem for fractional diffusion-wave equation. Journal of
Mathematical Sciences 2020; 250 (5): 800-810. https://doi.org/10.1007/s10958-020-05045-z
[23] Pskhu AV. The first boundary-value problem for a fractional diffusion-wave equation in a non-cylindrical domain.
Izvestiya: Mathematics 2017; 81 (6): 1212-1233. https://doi.org/10.1070/IM8520
[24] Pskhu A, Rekhviashvili S. Fractional diffusion-wave equation with application in electrodynamics. Mathematics
2020; 8 (11): 2086. doi: https://doi.org/10.3390/math8112086
[25] Ramazanov MI, Kosmakova MT, Kasymova LZh. On a problem of heat equation with fractional load. Lobachevskii
Journal of Mathematics 2020; 41 (9): 1873-1885. https://doi.org/10.1134/S199508022009022X
[26] Stanković B. On the function of E.M. Wright. Publications de l’Institut Mathématique 1970; 10 (24): 113-124.
[27] Wright E. Maitland. On the coefficients of power series having exponential singularities. Journal of the London
Mathematical Society 1933; 8 (1): 71-79. https://doi.org/10.1112/jlms/s1-8.1.71

1593
 PSKHU et al./Turk J Math

[28] Wright EM. The generalized Bessel function of order greater than one. The Quarterly Journal of Mathematics 1940;
11 (1): 36-48. https://doi.org/10.1093/qmath/os-11.1.36
[29] Wyss W. The fractional diffusion equation. Journal of Mathematical Physics 1986; 27 (11): 2782-2785.
https://doi.org/10.1063/1.527251

1594