ATOMIC SOLUTION OF PARTIAL DIFFERENTIAL

EQUATIONS VIA A TENSOR PRODUCT

THEORY OF BANACH SPACES

A DISSERTATION SUBMITTED TO THE BHARATHIAR UNIVERSITY,

COIMBATORE, IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE AWARD OF THE DEGREE OF

MASTER OF SCIENCE IN MATHEMATICS

SUBMITTED BY

SWETHA. K

REG.NO.: 22MATA32

UNDER THE GUIDANCE OF

Dr. N. ANNAPOORANI

DEPARTMENT OF MATHEMATICS

BHARATHIAR UNIVERSITY

COIMBATORE - 641 046

APRIL 2024
DECLARATION
Declaration
I, Mrs. Swetha. K, hereby declare that the dissertation entitled “Atomic
solution for Partial differential equations via a tensor product theory
of Banach spaces”, submitted to Department of Mathematics in Bharathiar
University, Coimbatore, in partial fulfillment of the requirements for the award of
the Degree of Master of Science in Mathematics, is a record of original research
work done by me under the supervision and guidance of Dr. N. Annapoorani,
Assistant Professor, Department of Mathematics, Bharathiar University, and that
it has not formed the basis for the award of any Degree/ Diploma/ Associateship/
Fellowship or other similar title to any candidate of any University.

Signature of the Candidate

i
CERTIFICATE
Certificate
This is to certify that the dissertation entitled “Atomic solution for Par-
tial differential equations via a tensor product theory of Banach spaces”,
submitted to the Bharathiar University, Coimbatore, in partial fulfillment of the
requirements for the award of the Degree of Master of Science in Mathemat-
ics, is a record of original research work done by Mrs. Swetha. K, during the
period of her study in the Department of Mathematics at Bharathiar University,
under my supervision and guidance and the dissertation has not formed the basis
for the award of any Degree/ Diploma/ Associateship/ Fellowship or other similar
title to any candidate of any University.

Signature of the Guide

Countersigned

Head of the Department

ii
ACKNOWLEDGEMENT
Acknowledgement

Words are a poor substitute when it comes to express my sincere thanks

and profound gratitude to my guide Dr. N. Annapoorani, Assistant professor,
Department of Mathematics, Bharathiar University, Coimbatore, for her inspiring
guidance, care and kind motivation.

I would like to convey my sincere thanks to Dr. S. Saravanan, Professor

and Head of the Department of Mathematics and all other faculty members of
Department of Mathematics, Bharathiar University, Coimbatore, for their valuable
guidance and encouragement to complete my work.

I am grateful to my parents for their moral support in pursuing my project

activities. I wish to express my heartful thanks to my friends and research scholars
for their valuable suggestions and timely help. I also render my sincere thanks to
all those who helped me to complete my dissertation.

- SWETHA. K

iii
ABSTRACT
Abstract
In this thesis, a novel analytical method for obtaining atomic solutions of par-
tial differential equations and fractional partial differential equations discussed. A
theory of tensor products in Banach spaces is utilized and some properties of atom
operators are studied to achieve the goals. The proposed scheme is demonstrated
through illustrative examples of non-homogeneous partial differential equations
and fractional partial differential equations. The approach is indeed quite direct
and effective in exactly solving partial differential equations. All possible solutions
for some fractional partial differential Burger’s type equations and the behavior of
these solutions are also obtained.

iv
CONTENTS
Contents

1 Introduction 1

1.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.2 Key theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2 Atomic solution for partial differential equations 11

2.1 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3 Atomic solution for fractional partial differential equations 27

3.1 The first Burger’s type equation . . . . . . . . . . . . . . . . . . . . 27

3.2 The second Burger’s type equation . . . . . . . . . . . . . . . . . . 36

CONCLUSION 41

BIBILIOGRAPHY 43

v
CHAPTER 1
Chapter 1

Introduction

It is well known that differential equations (DEs), particularly in the field of

mathematics, find numerous applications in solving real-world problems. They are
extensively utilized across various disciplines such as economics, physics, chem-
istry, and biology. Ordinary differential equations (ODEs) deal with derivatives of
functions involving a single variable, whereas partial differential equations (PDEs)
are employed to tackle problems involving functions of more than one variable.
Consequently, one-dimensional dynamical systems are often modeled using ODEs,
while multi-dimensional systems are typically described using PDEs. However, it’s
important to note that solving PDEs analytically is generally much more challeng-
ing compared to ODEs. There is no universal method applicable to solve all PDEs
of a given order, even numerically. Therefore, researchers often focus on studying
specific classes of PDEs individually. The separation of variables method is widely
favored for its relative simplicity and significant physical relevance in addressing
various challenges posed by certain types of PDEs. This approach typically in-
volves breaking down the linear PDE into a set of ODEs, where the principle of
superposition can be applied by combining specific linear combinations of funda-
mental solutions to obtain additional solutions. However, it’s important to note
that the separation of variables method comes with its limitations. It is primar-
ily effective in special cases where equations exhibit a high degree of symmetry.

1
Furthermore, the application of this method often requires imposing several con-
straints on coefficient expressions and initial or boundary conditions. The method
is based on the notion of decomposing the solution of the equation into a prod-
uct of functions, each dependent on a single variable. Indeed, the separation of
variables method is a powerful technique for reducing linear PDEs into a series of
ODEs. By constructing specific linear combinations of fundamental solutions, the
superposition principle allows for the generation of additional solutions. However,
it’s important to acknowledge the limitations of this method. It is most effective in
scenarios where equations exhibit a high degree of symmetry, limiting its applica-
bility to special cases. Additionally, the successful application of this method often
requires imposing various constraints on coefficient expressions, initial conditions,
and boundary conditions. This method operates on the principle of decomposing
the equations solution into a product of functions, each dependent on a single vari-
able. In other words, the final solution can be represented as the product of several
functions, each involving only one independent variable. In the realm of solving
PDEs, various analytical and numerical techniques are employed. These include
methods such as Fourier transform, Laplace transform, Green’s functions, finite
difference method, finite volume method, and finite element method, each offering
unique advantages and applications in different problem domains. The proposed
technique builds upon the tensor product theory of the Banach spaces, offering
a novel approach to obtain atomic solutions of DEs. Solving PDEs is fundamen-
tal for understanding various phenomena across scientific disciplines. Traditional
methods often face challenges when dealing with complex PDEs efficiently. How-
ever, leveraging the tensor product theory of the Banach spaces presents a promis-
ing and innovative strategy, providing a fresh analytical framework for addressing
these intricate problems. This thesis aims to explore and utilize the power of ten-
sor product theory within the domain of Banach Spaces to offer a new analytical
perspective for solving PDEs. By leveraging the inherent properties and struc-
ture of the Banach spaces alongside tensor product methodologies, this research

2
endeavors to provide a robust, comprehensive, and efficient strategy for tackling a
wide range of PDEs encountered in diverse scientific fields.
In this introduction, we will delve into the fundamental concepts of the Banach
spaces and tensor products, highlighting their significance in the context of solving
PDEs. Subsequently, we will discuss the proposed analytical approach, its theo-
retical foundations, and its potential implications for revolutionizing the way we
approach and solve PDEs. Through this exploration, we aim to present not only a
theoretical framework but also practical applications and insights that could lead
to more efficient and effective solutions for a myriad of PDEs.
The general scheme for the atomic solution method of PDEs involves several key
steps [8]:
Problem Formulation: We begin by defining PDEs problem along with any
boundary or initial conditions that govern the behavior of the system under con-
sideration. This step involves specifying the mathematical equation that describes
the physical phenomenon and any additional conditions that constrain the solu-
tion.
Atomic Basis Selection: Next, we select an appropriate atomic basis that can
effectively represent the solution of the PDE. Atoms are fundamental building
blocks that capture localized features of the solution. Choosing the right atomic
basis is crucial for accurately representing the solution and efficiently solving the
PDE. This selection process involves considering the specific properties of the prob-
lem and identifying basis functions that align with those properties.
Atomic Decomposition: Once the atomic basis functions are selected, we ex-
press the solution of the PDE as a linear combination of these basis functions.
This decomposition breaks down the complex solution into simpler components
represented by the chosen atoms.
Variational Formulation: We formulate a variational problem by defining a
functional that is minimized over the space of admissible solutions. This involves
setting up an optimization problem that seeks to minimize a certain energy or

3
error functional with respect to the coefficients of the atomic decomposition.
Solution Reconstruction: Next, we solve the variational problem to obtain
optimized coefficients for the atomic decomposition. With these coefficients, re-
construct the solution by joining the basis functions according to their weighted
contributions. This step yields an solution to the fractional PDE.
Convergence Analysis: We assess the convergence of the method by studying
the behavior of the error as the number of basis functions or the order of the ap-
proximation increases. This analysis helps determine the accuracy of the numerical
approximation.
Numerical Implementation: Now, we implement the method numerically, using
appropriate computational techniques to solve the variational problem and recon-
struct the solution. This involves employing numerical algorithms and software
tools to efficiently compute the optimized coefficients and generate the approxi-
mate solution of the fractional PDE.
Validation and Verification: Finally, we validate and verify the obtained atomic
solution. This general scheme of the atomic solution method of PDEs sets the stage
by highlighting the method’s systematic nature, efficiency, and versatility, which
can capture the attention of readers and convey the significance of the method in
solving complex PDEs. Additionally, it can serve as a foundation for discussing the
specific applications, implementations, and advancements of the atomic solution
method in the context of the given research or project. In this thesis, we also look
into solving fractional partial differential equations by using tensor product of the
Banach spaces. Classical methods often encounter dead ends when attempting to
solve some fractional partial differential equations, such as those requiring sepa-
ration of variables. Hence, in this thesis, we use the theory of the tensor product
of the Banach spaces to find exact solutions, called atomic solution which will
give us more than one solution to the proposed problems. As an application, we
will find all possible solutions for some fractional partial differential Burger’s type
equations and see the behavior of these solutions. New definitions of fractional

4
derivative and fractional integral real-valued functions, which were introduced in
2014 by Khalil et al., have revolutionised the theory of fractional calculus. In
this thesis, we present new exact solutions for solving fractional partial differential
equations by using the tensor product theory of Banach spaces called atomic solu-
tion (see [11], [9]). We explore the properties and advantages of this method, which
provides multiple solutions to the problems under consideration. We focus on frac-
tional partial differential Burger’s type equations, fundamental in various areas of
applied mathematics such as fluid mechanics, nonlinear acoustics, gas dynamics,
and traffic flow. Burger’s type was first introduced by Harry Bateman in 1915 and
later studied by Johannes Martinus Burgers in 1948. Therefore, we will discuss
the behaviors of the generated solutions since the subject of fractional partial dif-
ferential equations has been employed as an effective and qualitative tool used in
the formulation and modeling of complex dynamical systems and phenomena with
incomprehensible in engineering, physics, applied mathematics, and many more.
In recent research, atomic exact solutions have emerged as a powerful approach
for solving fractional partial differential equations within the context of the Ba-
nach spaces. This innovative method leveraging the mathematical framework of
the Banach spaces to provide accurate solutions to complex problems involving
fractional derivatives. The use of atomic functions enhances the precision of these
solutions, making them applicable to a diverse range of scenarios, from physics
to engineering, where fractional partial differential equations play a crucial role
in modeling intricate phenomena. This introduction sets the stage for exploring
the nuanced applications and theoretical underpinnings of atomic exact solutions
in the realm of fractional partial differential equations within the Banach spaces.
This method offers a systematic framework that combines analytical and numeri-
cal ideas to approximate solutions to fractional PDEs. By following this scheme,
we can develop a comprehensive understanding of the underlying principles and
techniques involved in solving fractional PDEs. This systematic approach not only
aids in the implementation and validation of the method but also contributes sig-

5
nificantly to advancing knowledge in this field. The introduction to the general
scheme of the atomic solution method for fractional PDEs offers a concise overview
of its systematic approach, highlighting its effectiveness in handling the complexi-
ties of fractional derivatives and non-local operators. It establishes the context for
discussing specific challenges and solutions related to fractional PDEs, showcasing
the methods accuracy and efficiency in providing solutions. Moreover, it serves as
a foundation for highlighting the potential applications and advancements of the
atomic solution method in the field of fractional calculus and applied mathemat-
ics. This general scheme provides a structured framework for solving fractional
PDEs, offering an organized approach to address these challenging mathematical
problems. Before presenting the main result regarding atomic solutions of par-
tial differential equations and fractional partial differential equations, we lay the
groundwork with some essential definitions and theorems.

1.1 Preliminaries

In this section, we present some preliminary concepts pertinent to the main

result of the thesis. To begin, we introduce the tensor(or direct) product of linear
spaces, which can be approached in various ways. One method is outlined below:

Definition 1.1. A complete normed linear space is called a Banach space.

Definition 1.2. Let N be a normed linear space over the field F . A continuous
linear transformation T : N → F is called a functional.

Definition 1.3. Let X be a Banach space over a field F which is real or complex,
then the dual space X ′ of X is the space of all continuous linear functionals on a
real or complex Banach space.

Definition 1.4. A mapping A from the cartesian product X × Y of vector spaces

into the vector space Z is bilinear if it is linear in each variable, that is,

A(α1 x1 + α2 x2 , y) = α1 A(x1 , y) + α2 A(x2 , y)

6
 A(x, β1 y1 + β2 y2 ) = β1 A(x, y1 ) + β2 A(x, y2 )

for all xi , x ∈ X, yi , y ∈ Y and scalars αi , βi , for i = 1, 2. We write B(X ×Y, Z)

for the vector space of bilinear mappings from the product X × Y into Z; when
Z is the scalar field we denote the corresponding space of bilinear forms simply by
B(X × Y ).

Definition 1.5. The tensor product X ⊗ Y , of the vector spaces X, Y can be

constructed as a space of linear functionals on B(X × Y ), for x ∈ X, y ∈ Y , we
denote by x ⊗ y the functional given by evaluation at the point (x, y). In other
words,
(x ⊗ y)(A) = ⟨A, x ⊗ y⟩ = A(x, y)

for each bilinear form A on X × Y . The tensor product X ⊗ Y is the subspace

of the dual space B(X × Y ) spanned by these elements. Thus, a typical tensor in
X ⊗ Y has the form
n
X
u= λi (xi ⊗ yi ),
i=1

where n is a natural number, λi ∈ K, xi ∈ X and yi ∈ Y .

Definition 1.6. A Hilbert space is a vector space H with an inner product ⟨f, g⟩
such that the norm defined by
p
|f | = ⟨f, f ⟩ .

Definition 1.7. A finite rank operator T : U → V between the Banach spaces is

a bounded operator such that its range is finite dimensional.
Just as in the Hilbert space H case, it can be written in the form
n
X
Th = ⟨ui , h⟩vi
i=1

for all h ∈ U , where now vi ∈ V and ui ∈ U ′ are bounded linear functionals on

the space H.

7
Definition 1.8. Let X and Y be any two Banach spaces and X ′ is the dual space
of X. For x ∈ X and y ∈ Y , the operator T : X ′ → Y , defined by

T (x′ ) = x′ (x)y = ⟨x, x′ ⟩y,

is bounded one rank linear operator. We write x ⊗ y for T and such operators are
called atoms.

Definition 1.9. If the solution of the differential equation can be represented by

an atom, then we say that it has an atomic solution.

Definition 1.10. A differential equation is called an (n, α) fractional differential

equation if it is of the form

f (t) = an y (n) + an − 1 y (n − α) + ....... + an − k y

where α is a factor of the natural number n. If n = 0, we take n to be the smallest n

for which α is a factor. This differential equation is called a fractional differential
equation of order n and factor α.

1.2 Key theorems

Atoms are commonly employed to establish a more effective methodology within

the Banach spaces and are regarded as fundamental components in tensor prod-
uct theory. A crucial result necessary for this work is presented in the following
theorem, which ensures that the sum of two atoms remains an atom:

Theorem 1.1. [1] Let x1 ⊗ y1 and x2 ⊗ y2 be two nonzero atoms in X ⊗ Y such

that x1 ⊗ y1 + x2 ⊗ y2 = x3 ⊗ y3 . Then either x1 = x2 = x3 or y1 = y2 = y3 .

This leads us to the following interesting theorem that lies at the heart of func-
tional analysis as well as approximation theory and guarantees that any continuous
function of several variables can be expressed as a sum of products of continuous
separated functions.

8
Theorem 1.2. [2] Let I, J be two compact intervals, and C(I), C(J) and C(I × J)
be the spaces of continuous functions on I, J and I, respectively. Then every
P
f ∈ C(I ×J) can be written in the form f (x, y) = ui (x)vi (y) where ui (x) ∈ C(I)
and vi (y) ∈ C(J).

Theorem 1.3. [2] The following properties are for X ⊗ Y :

1. x ⊗ (y1 + y2 ) = x ⊗ y1 + x ⊗ y2

2. (y1 + y2 ) ⊗ x = y1 ⊗ x + y2 ⊗ x

3. α(x ⊗ y) = αx ⊗ y = x ⊗ αy, for any scalar α

4. x ⊗ 0 = 0 ⊗ y = 0 ⊗ 0.

Theorem 1.4. [5] Let X and Y be the Banach spaces. Let a, b ∈ X and c, d ∈ Y
be nonzero vectors. If a ⊗ b = c ⊗ d then there exists a nonzero scalar β such that
1
a = βc and b = d.
β

Theorem 1.5. [5] Let x1 ⊗ y1 = x2 ⊗ y2 be two nonzero atoms such that

x1 ⊗ y1 + x2 ⊗ y2 = x3 ⊗ y3 .

Then either {x1 , x2 } or {y1 , y2 } are linearly dependent.

9
CHAPTER 2
Chapter 2

Atomic solution for partial

differential equations

Consider the following general first order, inseparable, non-homogeneous, and

linear partial differential equation:

uy (x, y) + f1 (x)g1 (y)ux (x, y) = f2 (x)g2 (y), (2.1)

where u is an unknown function and fi , gi (i = 1, 2) are given. Clearly equation

(2.1) is inseparable as we cannot move the x-terms to one side and the y-terms to
the other. Hence, the method of separation of variables does not work. According
to Theorem 1, we start our approach with assuming that

u(x, y) = P (x)Q(y). (2.2)

Now, we substitute equation (2.2) into the main partial differential equation (2.1).
Hence,
P (x)Q′ (y) + f1 (x)g1 (y)P ′ (x)Q(y) = f2 (x)g2 (y). (2.3)

Clearly, each term of equation(2.3) is just a product of two functions one of them
is pure in x and the other is pure in y. Therefore, in tensor product form,
equation(2.3) can be presented as

P (x) ⊗ Q′ (y) + f1 (x)P ′ (x) ⊗ g1 (y)Q(y) = f2 (x) ⊗ g2 (y). (2.4)

11
This implies that the sum of two atoms is an atom. By Theorem 1, we have one
of the following two cases:

(i) P (x) = f1 (x)P ′ (x) = f2 (x),

(2.5)
(ii) Q′ (y) = g1 (y)Q(y) = g2 (y).
Case (i): This case has the following three situations:
(a) P (x) = f1 (x)P ′ (x)
(b) P (x) = f2 (x)
(c) f1 (x)P ′ (x) = f2 (x).
From (a), we have Z
1
dx
P (x) = c1 e f1 (x) , (2.6)

where c1 is a constant. Also, by considering both situations (b) and (c) together
with (2.6), we have
Z
1
dx
P (x) = c1 e f1 (x) = f (x) and P ′ (x) = f2 (x) . (2.7)
2
f1 (x)
Thus, an atomic solution can be obtained for case (i) provided that the arguments
(2.6) and (2.7) are satisfied simultaneously otherwise no available atomic solution.
Now, our next step is to substitute (2.6) and (2.7) into (2.3) which yields

Q′ (y) + g1 (y)Q(y) = g2 (y). (2.8)

Clearly, (2.8) is a first order linear ordinary differential equation and its general
solution has the form
Z  Z 
− g1 (y)dy  Z g1 (y)dy
Q(y) = e g2 (y)e dy  . (2.9)



Hence, the first atomic solution with respect to case (i) can be obtained by
considering (2.2), (2.6) and (2.9) as follows:
Z
1
 Z  Z 
dx  − g1 (y)dy  Z g1 (y)dy 
u1 (x, y) = c1 e f1 (x) e  g 2 (y)e dy  . (2.10)

12
Case (ii): This case has, also, three situations, namely,
(a∗ ) Q′ (y) = g1 (y)Q(y),
(b∗ ) Q′ (y) = g2 (y),
(c∗ ) g1 (y)Q(y) = g2 (y) .
From (a∗ ), we have Z
g1 (y)dy
Q(y) = c2 e , (2.11)

where c2 is a constant. A second atomic solution can be obtained from case (ii)
by considering, together, the three cases (a∗ ) , (b∗ ), and (c∗ ) otherwise no available
atomic solution for this specific case.

Now, by substituting (b∗ ) , and (c∗ ) into (2.3) we get the following linear ordinary
differential equation in terms of x

1 f2 (x)
P ′ (x) + P (x) = , (2.12)
f1 (x) f1 (x)

which, clearly, has the solution

Z Z
1 1
 
− dx Z
f2 (x) dx 
P (x) = e f1 (x)  e f1 (x) dx . (2.13)
f1 (x)


Therefore, the second atomic solution with respect to case (ii) can be obtained
by substituting (2.11) and (2.13) into (2.2) as follows:
 Z Z
1 1
Z  
g1 (y)dy  − dx Z
f2 (x) dx 
u2 (x, y) = c2 e e f1 (x)  e f1 (x) dx . (2.14)
f1 (x)


2.1 Applications

In this section, we utilize the new method of atomic solutions to drive three
examples for handling non-homogeneous, linear and nonlinear partial differential
equations for which separation of variables method does not work.

13
 Example 1. Consider the following partial differential equation

ux + uy = ex+y , (2.15)

where u(x, y) is the unknown function and subjected to the following condition

u(0, 0) = 1. (2.16)

By substituting u(x, y) = P (x)Q(y) into (2.15) we get,

P ′ (x)Q(y) + P (x)Q′ (y) = ex+y . (2.17)

Therefore, in the tensor product form, (2.17) becomes

P ′ (x) ⊗ Q(y) + P (x) ⊗ Q′ (y) = ex ⊗ ey . (2.18)

By Theorem 1, we have one of the following two cases:

(i) P ′ (x) = P (x) = ex ,

(2.19)
′ y
(ii) Q(y) = Q (y) = e .

For case (i), we have the following three situations:

(a) P ′ (x) = P (x),
(b) P ′ (x) = ex ,
(c) P (x) = ex .
From (2.16), we can assume, without loss of generality, that

P (0) = Q(0) = 1. (2.20)

Now, from situation (a) and (2.20), we get

P (x) = ex . (2.21)

Also, by considering (2.20), both situations (b) and (c) give the same result in
(2.21). Therefore, an atomic solution exists with respect to case (i) and can be

14
obtained by substituting (2.21) into (2.17) to get Q′ (y) + Q(y) = ey that is a linear
ordinary differential equation which can be solved by taking into account (2.20) as

ey + e−y
Q(y) = . (2.22)
2

Hence, referring to (2.21) and (2.22), the first atomic solution with respect to case
(i) is
ey + e−y
 
x
u1 (x, y) = e . (2.23)
2
Similarly, we can handle case (ii) in order to get the second atomic solution with
respect to case (ii) that is

ex + e−x
 
y
u2 (x, y) = e . (2.24)
2

The two atomic solutions u1 (x, y) in (2.23) and u2 (x, y) in (2.24) of problem
(2.15) are displayed respectively in Figure 2.1 and 2.2

Figure 2.1: The first atomic solution u1 (x, y) in (2.23) of problem (2.15).

15
 Figure 2.2: The second atomic solution u2 (x, y) in (2.24) of problem (2.15).

Example 2. Consider the following partial differential equation

uy − (tan x tan y)ux = sin x sin y, (2.25)

where u(x, y) is the unknown function and subjected to the following condition
π π  1
u , = . (2.26)
4 4 2
By substituting u(x, y) = P (x)Q(y) into (2.25) we get,

P (x)Q′ (y) − (tan x tan y)P ′ (x)Q(y) = sin x sin y. (2.27)

Therefore, in the tensor product form, (2.27) becomes

P (x) ⊗ Q′ (y) + (tan x)P ′ (x) ⊗ (− tan y)Q(y) = sin x ⊗ sin y. (2.28)

Using Theorem 1, we have one of the following two cases:

(i) P (x) = (tan x)P ′ (x) = sin x,

(2.29)
(ii) Q′ (y) = (− tan y)Q(y) = sin y.

16
For case (i), we have the following three situations:
(a) P ′ (x) = (cot x)P (x),
(b) P (x) = sin x,
(c) P ′ (x) = cos x.
From (2.26), we can assume, without loss of generality, that
π  π 
1
P =Q =√ . (2.30)
4 4 2
Clearly, when considering the condition P π4 = √12 that is assumed in (2.30), all

three situations (a), (b), and (c) yield the same result that is

P (x) = sin x. (2.31)

Therefore, an atomic solution exists with respect to case (i) and can be obtained
by substituting (2.31) into (2.27) to get Q′ (y) − (tan y)Q(y) = sin y. This linear
ordinary differential equation can be solved by taking into account the condition
Q π4 = √12 assumed in (2.30) as

2 sin2 y + 1
 
Q(y) = sec y . (2.32)
4
Hence, referring to (2.30) and (2.32), the first atomic solution with respect to case
(i) is
2 sin2 y + 1
 
u1 (x, y) = sin x sec y . (2.33)
4
For case (ii), we have three situations, namely,
(a∗ ) Q′ (y) = (− tan y)Q (y),
(b∗ ) Q′ (y) = sin y and
(c∗ ) (− tan y)Q(y) = sin y.

Now, by assuming situations (a∗ ) , (b∗ ), and (c∗ ) together with the condition
Q π4 = √12 that is assumed in (2.30), we get, respectively, Q(y) = cos y, Q(y) =


√
− cos y + 2, and Q(y) = − cos y. Therefore, no available atomic solution for
case (ii) and hence, this case does not admit an atomic solution. The only atomic
solution u1 (x, y) in 2.33) of problem (2.25) is displayed in Figure 2.3.

17
 Figure 2.3: The first atomic solution u1 (x, y) in (2.33) of problem (2).

Example 3. Consider the following initial-value problem

u(x, y)ux (x, y) + u(x, y)uyyy (x, y) = uxx (x, y)uy (x, y), (2.34)

where u(x, y) is an unknown function. Clearly, (2.34) is inseparable as we cannot

move the x-terms to one side and the y-terms to the other. Hence, the method of
separation of variables does not work. The following conditions are imposed on u
:
u(0, 0) = ux (0, 0) = uy (0, 0) = uxx (0, 0) = uyy (0, 0) = 1. (2.35)

We start our approach with assuming that

u(x, y) = P (x)Q(y). (2.36)

Now, we substitute (2.36) into (2.34) to get

P ′ (x)Q(y) + P (x)Q′′′ (y) = P ′′ (x)Q′ (y). (2.37)

18
Clearly, each term of (2.37) is just a product of two functions one of them is pure
in x and the other is pure in y. Therefore, in the tensor product form, (2.37) can
be presented as

P ′ (x) ⊗ Q(y) + P (x) ⊗ Q′′′ (y) = P ′′ (x) ⊗ Q′ (y). (2.38)

This implies that the sum of two atoms is an atom. By Theorem 1, we have one
of the following two cases:

(i) P ′ (x) = P (x) = P ′′ (x),

(2.39)
′′′ ′
(ii) Q(y) = Q (y) = Q (y).

Case (i): This case has the following three situations, namely,
(a) P ′ (x) = P (x),
(b) P ′′ (x) = P ′ (x),
(c) P (x) = P ′′ (x).
Now, from (2.35), we can assume, without loss of generality, that

P (0) = Q(0) = P ′ (0) = Q′ (0) = P ′′ (0) = Q′′ (0) = 1. (2.40)

Thus, by (2.40), the three situations of case (i) can be solved for P (x) as

P (x) = ex . (2.41)

Now, our next step is to substitute (2.41) into (2.37) taking into account that

P (x) = P ′ (x) = P ′′ (x) = ex ,

which implies
Q′′′ (y) − Q′ (y) + Q(y) = 0. (2.42)

Therefore, by (2.40) and (2.42), we have

√ ! 1 √ ! 1 

3 y 1 3 y
u1 (x, y) = ex cos y e 2 + √ sin y e2  . (2.43)
2 3 2

19
Case (ii): This case has the following three situations, namely,
(a) Q′′′ (y) = Q(y),
(b) Q′ (y) = Q(y),
(c) Q′′′ (y) = Q′ (y).
Now, by (2.40), the three situations (a), (b), and (c) yield the same result that is

Q(y) = ey . (2.44)

By substituting (2.44) into (2.37) we get P ′′ (x) − P ′ (x) − P (x) = 0, which is a

second order homogeneous ordinary differential equation with constant coefficients
and can be solved as
√ ! 1 √ ! 1
5 x 1 5 x
P (x) = cosh x e 2 + √ sinh x e2 . (2.45)
2 5 2

Hence, referring to (2.36), (2.44) and (2.45), the second atomic solution with re-
spect to case (ii) is
√ ! 1 √ ! 1 

5 x 1 5 x
u2 (x, y) = ey cosh x e 2 + √ sinh x e2  . (2.46)
2 5 2

To visualize the solutions u1 (x, y) in (2.45) and u2 (x, y) in (2.46) of problem (2.34),
we plot these two solutions in Figure 2.4 and Figure 2.5, respectively.

20
 Figure 2.4: The first atomic solution u1 (x, y) in (2.43) of problem (3).

Figure 2.5: The second atomic solution u1 (x, y) in (2.43) of problem (3).

21
 Example 4 Consider the following non-linear 2-D PDE

uxx uy + uyy ux = uy u, (2.47)

where u(x, y) is the unknown function and subjected to the following conditions:

u(0, 0) = 1, ux (0, 0) = 1, and uy (0, 0) = 1. (2.48)

By substituting u(x, y) = P (x)Q(y) into (2.47) we get,

P ′′ (x)Q′ (y) + P ′ (x)Q′′ (y) = P (x)Q′ (y). (2.49)

Therefore, in the form of tensor product, (2.49) becomes

P ′′ (x) ⊗ Q′ (y) + P ′ (x) ⊗ Q′′ (y) = P (x) ⊗ Q′ (y). (2.50)

By Theorem 1, we have one of the following two situations:

(i) P ′′ (x) = P ′ (x) = P (x),

(2.51)
′′ ′
(ii) Q (y) = Q (y).

Hence, for case (i), we have the following three situations:

(a) P ′′ (x) = P ′ (x),
(b) P ′′ (x) = P (x),
(c) P ′ (x) = P (x).
Without loss of generality and from (2.48), we can suppose that

P (0) = Q(0) = P ′ (0) = Q′ (0) = 1. (2.52)

Now, from situation (a) and the two conditions P (0) = 1 and P ′ (0) = 1, we get

P (x) = ex . (2.53)

Also, by considering (2.52), both situations (b) and (c) give the same result in
(2.53). Therefore, an atomic solution exists with respect to case(i). Now, we

22
proceed by substituting (2.53) into (2.49) which implies Q′′ (y) = 0. But, by
(2.52), we have Q(0) = 1 and Q′ (0) = 1 so,

Q(y) = y + 1. (2.54)

Hence, referring to (2.53) and (2.54), the first atomic solution with respect to case
(i) is
u1 = (x, y) = ex (y + 1). (2.55)

Case (ii). Clearly, for this particular case, we have only one situation that is
Q′′ (y) = Q′ (y). Hence, having Q(0) = 1 and Q′ (0) = 1 from (2.52) yields

Q(x) = ey . (2.56)

Our next step is to substitute (2.56) into (2.49) which yields

P ′′ (x) + P ′ (x) − P (x) = 0. (2.57)

Assuming the conditions in (2.52), the second-order ODE (2.57) has the following
solution
 √ √ 
−1 + 5 −1 − 5
1 √ x √ x
P (x) = √ ( 5 + 3)e 2 + ( 5 − 3)e 2 . (2.58)
2 5

Hence, referring to (2.56) and (2.58) , the second atomic solution with respect to
case (ii) is
 √ √ 
−1 + 5 −1 − 5
y
e √ x √ x
u2 (x, y) = √ ( 5 + 3)e 2 + ( 5 − 3)e 2 . (2.59)
2 5

The two atomic solutions u1 (x, y) and u2 (x, y) of problem (2.61) are exhibited
respectively in Figures 2.6 and 2.7.

23
 Figure 2.6: The first atomic solution u1 (x, y) in (2.55) of problem (4).

Figure 2.7: The second atomic solution u2 (x, y) in (2.59) of problem (4).

Example 5 Consider the following non-linear 2-D PDE

uxx uy + ux uyy = ux uy , (2.60)

24
where u(x, y) is the unknown function and subjected to the following conditions:

u(0, 0) = 1, ux (0, 0) = 1, and uy (0, 0) = 1. (2.61)

By substituting u(x, y) = P (x)Q(y) into (2.60) we get,

P ′′ (x)Q′ (y) + P ′ (x)Q′′ (y) = P ′ (x)Q′ (y). (2.62)

Therefore, in tensor product form, (2.62) becomes

P ′′ (x) ⊗ Q′ (y) + P ′ (x) ⊗ Q′′ (y) = P ′ (x) ⊗ Q′ (y), (2.63)

By Theorem 1, the two resultant situations are

(i) P ′′ (x) = P ′ (x),

(2.64)
(ii) Q′′ (y) = Q′ (y).
Without loss of generality and from (2.61), we can suppose that

P (0) = Q(0) = P ′ (0) = Q′ (0) = 1. (2.65)

First, form case (i) and assuming (2.65), we have

P (x) = ex . (2.66)

Now, substituting (2.66) into the main equation (2.62) yields Q′′ (y) = 0. Therefore,
we get
Q(y) = y, (2.67)

where the conditions Q(0) = 1 and Q′ (0) = 1 are given by (2.65). Hence, the first
atomic solution corresponding to case (i) can be obtained by considering (2.66)
and (2.67) as follows:
u1 (x, y) = yex . (2.68)

Similarly, from case (ii), one can find out the second atomic solution that is

u2 (x, y) = xey . (2.69)

In order to visualize the solutions of problem (2.60), we plot its solutions u1 (x, y)
and u2 (x, y) in Figure 2.8 and Figure 2.9, respectively.

25
 Figure 2.8: The first atomic solution u1 (x, y) in (2.68) of problem (5).

Figure 2.9: The second atomic solution u2 (x, y) in (2.69) of problem (5).
Chapter 3

Atomic solution for fractional

partial differential equations

In this chapter, we will illustrate the atomic solution method using the tensor
product of the Banach space to find all possible solutions for two fractional partial
differential Burger’s type equations [10].

3.1 The first Burger’s type equation

In this subsection, we consider the following fractional partial differential Burger’s

type equations (see [3, 7]):

β α U = Dα Dβ Dβ U + U,
2Dy U + 2Dx x y y (3.1)

x > 0, y > 0.
β ∂β α β β
α U = ∂ U (x, y) and Dβ Dβ U = ∂ ∂ U (x, y)
Here, Dy U = U (x, y), Dx y y
∂y ∂x ∂y ∂y
subject to the following conditions:

U (0, y) = Q(y),

β
U (x, 0) = Dy U (x, 0) = P (x).

27
To deal with the problem (3.1), we list below the proposed procedure as follows:
Step (i): We first pose the solution of problem (3.1) as

U (x, y) = P (x) ⊗ Q(y). (3.2)

Now, substituting (3.2) in (3.1) yields

P (x) ⊗ 2Qβ (y) + 2P α (x) ⊗ Q(y)

(3.3)
= P α (x) ⊗ Q2β (y) + P (x) ⊗ Q(y),

here, P (x) and Q(y) are unknown functions.

∂β ∂β ∂α
Also, Q2β (y) = Q(y) and P α (x) = P (x). Consequently, one can
∂y ∂y ∂x
deduce
P (0) = 1 and Q(0) = Qβ (0) = 1. (3.4)

Indeed, we have
U (x, y) = P (x) ⊗ Q(y) = P (x)Q(y).

Then, from the conditions we get

U (0, y) = P (0) ⊗ Q(y) = P (0)Q(y) = Q(y),

U (x, 0) = P (x) ⊗ Q(0) = P (x)Q(0) = P (x),

β
Dy U (x, 0) = P (x) ⊗ Qβ (0) = P (x)Qβ (0) = P (x).

By identification, we obtain the result as required.

Step (ii): Collecting terms in (3.3) using Theorem 1 we obtain

P (x) ⊗ 2Qβ (y) + 2P α (x) − P (x) ⊗ Q(y) = P α (x) ⊗ Q2β (y).

(3.5)

In (3.5), we have the sum of two atoms is an atom. We have two cases to consider:
Case (i):
P (x) = 2P α (x) − P (x) = P α (x).

Case (ii):
2Qβ (y) = Q(y) = Q2β (y).

28
Discussion of Case (i): This case has the three following situations:

(1) P (x) = 2P α (x) − P (x),

(2) P (x) = P α (x), (3.6)

(3) 2P α (x) − P (x) = P α (x).

Let us first solve the first situation (1) P (x) = 2P α (x) − P (x). This implies

P α (x) = P (x), (3.7)

which is a linear fractional differential equation. Therefore, we can use the result
to find
xα
P (x) = ce α ,

for some constant c. Now, using conditions (3.4) implies that c = 1. Hence, we
have
xα
P (x) = e α . (3.8)

We should notice here that the other remaining situations of case (i), (i.e. situa-
tions (2) and (3) reported in (3.6), can provide the same solution as situation (1).
Therefore, we consider them as solutions. Now, we go back to the equation (3.3)
to obtain
xα xα xα xα
2e α Qβ (y) + 2e α Q(y) = e α Q2β (y) + e α Q(y).

Simplifying the above equality implies

2Qβ (y) + 2Q(y) = Q2β (y) + Q(y), (3.9)

(3.9) has the form

Q2β (y) − 2Qβ (y) − Q(y) = 0. (3.10)

Again, the above equation represents a linear fractional differential equation. So,
we find
yβ yβ
r1 r2
Q(y) = c1 e β + c2 e β , (3.11)

29
in which
√ √
r1 = 1 + 2 and r2 = 1 − 2,

where c1 and c2 are constants to be determined. Consequently, conditions (3.4)

lead to the following assertions:

c1 + c2 = 1 and r1 c1 + r2 c2 = 1,

from which we get c1 = c2 = 21 . Hence, we have

β β
 
√ y √ y
1  (1− 2) β (1+ 2)
β ,

Q(y) =  e + e (3.12)
2 

combining (3.8) and (3.12), we get the first atomic solution for the fractional partial
differential (3.1), which would be as follows
β β
 
xα √ y √ y
1  (1− 2) (1+ 2)
β +e β ,

U (x, y) = P (x)Q(y) = e α  e (3.13)
2  

and this ends case (i).

Discussion of Case (ii): In this case, we have

2Qβ (y) = Q(y) = Q2β (y).

Thus, there are also three situations that should be considered here. Such situa-
tions are given as
(1) 2Qβ (y) = Q(y),

(2) Q(y) = Q2β (y), (3.14)

(3) 2Qβ (y) = Q2β (y).

For the case to be valid, the three situations must give the same solution for
Q(y). So, we start with situation (1) reported in (3.14), which represents a linear
fractional differential equation. Thus, we can obtain

1 yβ
Q(y) = ce 2 β ,

30
for some constant c. As a result, condition (3.4) gives Q(0) = c = 1, and hence we
have
1 yβ
Q(y) = e 2 β .

In the same regard, one might observe that the second situation mentioned in
(3.14) is also a linear fractional differential equation. Then, we get

yβ yβ
−
Q(y) = c1 e β + c2 e β ,

for some constants c1 and c2 . By using conditions (3.4), we get c1 = 1 and c2 = 0.

Consequently, we can have
yβ
Q(y) = e β .

Thus, the solutions of situations (1) and (2) reported in (3.14) do not match with
each other, (i.e. do not give the same solution). Thus, case (ii) does not give a
second atomic solution for the fractional partial differential equation (3.1), and
this ends case (ii).

Remark 3.1. Many researchers think that the procedure is finished here, and this
is the only possible solution for the fractional partial differential equation (3.1).
Well, this reasoning is wrong. To obtain all possible solutions for our problem, we
must extract all the possible cases by recollecting terms in (3.3).

In light of Remark 3.1 and after recollecting terms in (3.3) with the use of the
properties of tensor product, we obtain

2P (x) ⊗ Qβ (y) + 2P α (x) − P (x) ⊗ Q(y) = P α (x) ⊗ Q2β (y).

(3.15)

Now, due to the sum of two atoms being an atom in (3.16), then it should to be
considered a further two cases that can be reported as follows:
Case (iii):
2P (x) = 2P α (x) − P (x) = P α (x).

31
Case (iv):
Qβ (y) = Q(y) = Q2β (y).

Discussion of Case (iii): As it can be seen here, this case has the following three
situations:
(1) 2P (x) = 2P α (x) − P (x),

(2) 2P (x) = P α (x), (3.16)

(3) 2P α (x) − P (x) = P α (x).

The first situation of (3.16) implies

3
P (x) = P α (x).
2

So, we have
3 xα
P (x) = ce 2 α ,

for some constant c, and so with the help of conditions (3.4), we get c = 1. Hence,
we have
3 xα
P (x) = e 2 α .

Similarly, the solution to the second situation of (3.16) can be obtained to be as

xα
2
P (x) = e α .

Thus, one can clearly observe that the solutions of situations (1) and (2) of (3.16)
do not match with each other. So, this case does not give the desired second atomic
solution for problem (3.1), and this ends case (iii).
Discussion of Case (iv): This case has also three situations reported in the following
manner:
(1) Qβ (y) = Q(y),

(2) Q(y) = Q2β (y), (3.17)

(3) Qβ (y) = Q2β (y).

32
Here, we will find that all these situations can give the same solution, which would
be as
yβ
Q(y) = ce β ,

for some constant c.

Now, by using conditions (3.4), we get c = 1. Consequently, we can have

yβ
Q(y) = ce β . (3.18)

Now, going back to (3.16) yields

yβ yβ yβ
2P (x)e β + (2P α (x) − P (x))e β = P α (x)e β .

Therefore, we have
P α (x) = −P (x).

Thus, we obtain
xα
−
P (x) = ce α ,

for some constant c, and hence conditions (3.4) implies c = 1. Then, we have

xα
−
P (x) = e α . (3.19)

Combining (3.18) and (3.1) provides the second atomic solution for problem (3.1),
which would be as
β
xα y
−
U (x, y) = P (x)Q(y) = e α e β . (3.20)

This ends case (iv). It is worth mentioning that more cases can be obtained by
reformulating (3.3). We list them below for completeness:
Case (v):
P (x) = 2P α (x) − P (x) = P α (x).

33
Case (vi):
2Qβ (y) − Q(y) = Q(y) = Q2β (y).

Case (vii):
P (x) = P α (x) = P α (x).

Case (viii):
2Qβ (y) − Q(y) = 2Q(y) = Q2β (y).

In a similar manner to the previous discussion, one might deduce that case (v),
case (vi) and case (viii) do not give another atomic solution for problem (3.1).
However, by the same reasoning, case (viii) can give the same first atomic solution
reported in (3.13). To see how the generated solutions of problem (3.1) look like
according to different fractional-order values of α and β, we plot the first atomic
solution (3.13) in Figure 3.1, Figure 3.2 and the second atomic solution (3.20) in
Figure 3.3, Figure 3.4.

1 1
Figure 3.1: The first atomic solution u1 (x, y) in (3.13) of (3.1) for α = 2 and β = 2

34
 1
Figure 3.2: The second atomic solution u2 (x, y) in (3.20) of (3.1) for α = 2 and β = 21 .

1
Figure 3.3: The first atomic solution u1 (x, y) in (3.13) of (3.1) for α = 4 and β = 34 .

35
 1
Figure 3.4: The second atomic solution u2 (x, y) in (3.20) of (3.1) for α = 4 and β = 43 .

3.2 The second Burger’s type equation

In this part, we aim to present the second fractional partial differential Burger’s
type equation, which has the form:

β α U = D2α D2β U,
Dy U + Dx x y (3.21)

x > 0, y > 0. Equation (3.21) is subject to the following conditions:

α U (0, y) = Q(y),
U (0, y) = Dx

β
U (x, 0) = Dy U (x, 0) = P (x).

Although (3.21) is a linear fractional partial differential equation, it is very diffi-

cult to solve using separation of variables method. Hence, we will use the tensor
product technique to try to find the atomic solution. For this purpose, we list
below the proposed procedure as follows:
Step (i): We assume that the solution of problem (3.21) has the form

U (x, y) = P (x) ⊗ Q(y). (3.22)

Step (ii): We substitute (3.22) into (3.21) to obtain

P (x) ⊗ Qβ (y) + P α (x) ⊗ Q(y) = P 2α (x) + Q2β (y). (3.23)

36
Consequently, we can deduce the following conditions:

P (0) = P α (0) = 1, and Q(0) = Qβ (0) = 1. (3.24)

It should be noted here that through (3.23), we can observe that the sum of two
atoms is an atom. Hence, we have the two following cases:
Case (i):
P (x) = P α (x) = P 2α (x).

Case (ii) :
Q(y) = Qα (y) = Q2β (y).

Discussion of case (i): This case has clearly the following three situations:

(1) P (x) = P α (x),

(2) P (x) = P 2α (x), (3.25)

(3) P α (x) = P 2α (x).

Thus, to solve the first situation (1) of (3.24), we obtain

xα
P (x) = ce α ,

for some constant c. Consequently, conditions (3.24) imply that c = 1, and hence
we have
xα
P (x) = e α . (3.26)

We should notice here that the other two situations (2) and (3) of (3.25) can give
the same solution. Now, substituting yields

xα xα xα
e α ⊗ Qβ (y) + e α ⊗ Q(y) = e α ⊗ Q2β (y).

Simplifying the above equality give

Qβ (y) + Q(y) = Q2β (y),

37
which represents a linear fractional partial differential equation. Then, we get
yβ yβ
r1 r2
Q(y) = c1 e β + c2 e β ,
√ √
1− 5 1+ 5
where r1 = and r2 = , and where c1 and c2 are two constants to be
2 2
determined. In this regard, using conditions (3.24) implies
√ √
−1 + 5 1+ 5
c1 = √ , and c2 = √ .
2 5 2 5
Hence, we have

√ ! yβ √ ! yβ
−1 + 5 r1 1+ 5 1 2 β r
Q(y) = √ e β + √ e , (3.27)
2 5 2 5 2
√ √
1− 5 1+ 5
where r1 = and r2 = . Combining (3.26) and (3.27) leads us to the
2 2
first atomic solution for the fractional partial differential equation (3.21), which
would be as follows
U (x, y) = P (x)Q(y)
yβ yβ
 
x α √ ! √ !
 −1 + 5 r1 1 + 5 1 r2 β  (3.28)
=e α  √ e β + √ e ,
 2 5 2 5 2 

√ √
1− 5 1+ 5
where r1 = and r2 = , and this ends case (i).
2 2
Discussion of Case (ii): The symmetry of the equation makes the discussion of
case (ii) is similar to case (i). In other words, case (ii) can give a second atomic
solution to the problem (3.21), which would be in the form
U (x, y) = P (x)Q(y)
 β
xα xα y

√ ! √ ! (3.29)
−1 + 5 r1 1 + 5 1 r2
= √ e α + √ e α e β ,
 
2 5 2 5 2
√ √
1− 5 1+ 5
where r1 = and r2 = . This ends case (ii), as well. To observe the
2 2
behavior of the solutions generated for problem (3.21) across different fractional-
order values of α and β, we depict the first atomic solution (3.28) in Figure 3.5

38
and Figure 3.6, and the second atomic solution (3.29) in Figure 3.7, Figure 3.8.

1
Figure 3.5: The first atomic solution u1 (x, y) in (3.28) of (3.21) for α = 2 and β = 12 .

1
Figure 3.6: The second atomic solution u2 (x, y) in (3.29) of (3.21) for α = 2 and β = 12 .

39
 1
Figure 3.7: The first atomic solution u1 (x, y) in (3.28) of (3.21) for α = 4 and β = 34 .

1
Figure 3.8: The second atomic solution u2 (x, y) in (3.29) of (3.1) for α = 4 and β = 43 .

40
Conclusion

In this thesis, an innovative analytical method has been employed to derive

atomic solutions for both linear and nonlinear ordinary and fractional partial dif-
ferential equations. Leveraging the theory of tensor products of the Banach spaces
along with properties of atomic operators, the proposed method demonstrates effi-
cacy in obtaining atomic solutions for various non- homogeneous linear and nonlin-
ear partial differential equations as well as fractional partial differential equations
in most cases. This thesis has demonstrated the effectiveness of the atomic so-
lution metod in solving both partial and fractional partial differential equations.
Several examples have been explored to illustrate the application of the atomic
solution method and the graphs of atomic solutions are also plotted. In addition
to this, atomic solutions of a general first order, inseparable, non-homogeneous,
and nonlinear partial differential equation are also obtained.

41
BIBLIOGRAPHY
Bibliography

[1] W. G. Alshanti, I. M. Batiha and M. Abu Hammad, R. Khalil, A novel an-

alytical approach for solving partial differential equations via a tensor product
theory of Banach spaces, Partial Differential Equations in Applied mathematics,
(2023).

[2] W. G. Alshanti, I. Batiha and A. Alshanty, Atomic solutions of partial dif-

ferential equations via tensor product theory of Banach spaces, Contemporary
Mathematics, (2023) 286–295.

[3] A. Bushnaque, M. Al Horani and R. Khalil, Tensor product technique and

atomic solution of fractional Bate Man burger equation, Journal of mathematical
and computational sciences, 11, (2021) 330–336.

[4] R. Khalil, Best approximation in Lp (I, X), In Mathematical Proceedings of the

Cambridge Philosophical Society, 94, (1983) 277–279.

[5] R. Khalil, Isometries of (Lp ⊗Lp ), Tamkang Journal of Mathematics, 16, (1985)
177–185.

[6] R. Khalil, L. Abdullah, Atomic solution of certain inverse problems, European

Journal of Pure and Applied Mathematics, 3, (2010) 725–729.

[7] R. Khalil, M. Al Horani, A. Yousef and M. Sababheh, A new definition of frac-

tional derivative, Journal of Computing Science and Mathematics, 264, (2014)
65–67.

43
[8] W. Light and E. Cheny, Approximation Theory in Tensor Product Spaces,
Springer, London (1985).

[9] W. A. Light and E. W. Cheny, Approximation Theory in Tensor Product

Spaces, London, Springer, (2006).

[10] K. S. Miller, An Introduction to Fractional Calculus and Fractional Differen-

tial Equations, (1993).

[11] R. A. Ryan and R. Ryan, An Introduction to Tensor Products of Banach

Spaces, (2002).

44