COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.

Dubovski

METHODS OF INTEGRAL TRANSFORMS

V. I. Agoshkov and P. B. Dubovski

Institute of Numerical Mathematics, Russian Academy of Sciences, Moscow, Russia

Keywords: Integral transform, Fourier transform, Laplace transform, Mellin transform,

Hankel transform, Meyer transform, Kontorovich-Lebedev transform, Mehler-Foque
transform, Hilbert transform, Laguerre transform, Legendre transform, convolution
transform, Bochner transform, chain transform, wavelet, wavelet transform, problems of
the oscillation theory, heat conductivity problems, problems of the theory of neutron
slowing-down, problems of hydrodynamics, problems of the elasticity theory,
Boussinesq problem, coagulation equation, physical kinetics.

Contents

S
TE S
1. Introduction

R
AP LS
2. Basic integral transforms
2.1. The Fourier Transform
2.1.1. Basic Properties of the Fourier Transform
C EO
2.1.2. The Multiple Fourier Transforms
2.2. The Laplace Transform
2.2.1. The Laplace Integral
2.2.2. The Inversion Formula for the Laplace Transform
E –

2.2.3. Limit Theorems

H

2.3. The Mellin Transform

PL O

2.4. The Hankel Transform

2.5. The Meyer Transform
M SC

2.6. The Kontorovich-Lebedev Transform

2.7. The Mehler-Foque Transform
2.8. The Hilbert Transform
2.9. The Laguerre and Legendre Transforms
SA NE

2.10. The Bochner Transform, the Convolution Transform, the Wavelet and Chain
Transforms
3. The application of integral transforms to problems of the oscillation theory
U

3.1. Electric Oscillations

3.2. Transverse Oscillations of a String
3.3. Transverse Oscillations of an Infinite Round Membrane
4. The application of integral transforms to heat conductivity problems
4.1. The Solution of the Heat Conductivity Problem by the use of the Laplace
Transform
4.2. The Solution of the Heat Conductivity Problem by the use of the Fourier Transform
4.3. The Problem of Temperature Condition of a Sphere
5. The application of integral transforms in the theory of neutron slow-down and
diffusion
5.1. The Solution of the Equation of Neutron Slow-down for Moderator of Infinite Size
5.2. The Problem of Diffusion of Thermal Neutrons
6. The application of integral transforms to problems of hydrodynamics
6.1. Two-dimensional Irrotational Flow of an Ideal Liquid

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

6.2. Flow of an Ideal Liquid through a Gap

6.3. Outflow of an Ideal Liquid through a Round Aperture
7. The application of integral transforms in the elasticity theory
7.1. Axially Symmetric Stresses in a Cylinder
7.2. The Boussinesq Problem in a Half-space
7.3. Determination of Stresses in a Wedge
8. The application of integral transforms in the coagulation kinetics
8.1. The exact Solution of the Coagulation Equation
8.2. The Violation of the Law of Conservation of Mass
9. Brief instructions for the application of integral transforms
Glossary
Bibliography
Biographical Sketches

Summary

S
TE S
R
AP LS
The methods of integral transforms are very efficient to solve and research differential
and integral equations of mathematical physics. These methods consist in the integration
of an equation with some weight function of two arguments that often results in the
C EO
simplification of a given initial problem. The main condition for the application of an
integral transform is the validity of the inversion theorem which allows one to find an
unknown function knowing its image. The Fourier, Laplace, Mellin, Hankel, Meyer,
Hilbert, and other transforms are used depending on a weight function and an
integration domain. With the help of these transforms many problems of the oscillation
E –

theory, heat conductivity, neutron diffusion and slowing-down, hydrodynamics, the

H

elasticity theory, and physical kinetics can be solved.

PL O

1. Introduction
M SC

An integral transform is a functional transform of the form

SA NE

F ( x) = ∫ K ( x, t ) f (t )dt ,
Γ
where Γ is a finite or infinite contour in the complex plane, K ( x, t ) is the kernel of a
U

transform. The most frequently considered integral transforms are those for which
K ( x, t ) = K ( xt ) and Γ is the real axis or its part (a, b) . If −∞ < a, b < ∞ , then a transform
is referred to as a finite integral transform. Formulae that allow us to reconstruct a
function f (t ) by a given one F ( x) are said to be inversion formulae of integral
transforms.

If x, t ∈ R n and Γ is a domain in the n-dimensional Euclidean space, then multiple

(multidimensional) integral transforms are considered.

Integral transforms are frequently applied when solving differential and integral
equations and their choice depends on the type of a considered equation. The main
condition when choosing an integral transform is the possibility to reduce a differential
or integral expression to a more simple differential equation (or, what is better, to an

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

algebraic ratio) with respect to a function F ( x) implying that an inversion formula is

known. If a contour Γ is finite (for example, a segment), then a transform F ( x) is
called a finite transform of f (t ) . It is obvious that the number of integral transforms can
be considerably increased by introducing new kernels.

In the following we mainly consider transforms where a contour Γ is the real axis or
real semiaxis supposing that all integrals are finite.

2. Basic Integral Transforms

2.1. The Fourier Transform

The expression

S
TE S
∞
1 −ixτ
F ( x) ≡ F [ f ] = ∫e f (τ )dτ

R
2π

AP LS
−∞

is called the Fourier transform of a function f (t ) .

C EO
A function F ( x) is called the Fourier image of a function f . The inverse Fourier
transform has the form
E –

∞
H

−1 1
f (t ) ≡ F [ F ( x)] = ∫e
itx
F ( x)dx .
PL O

2π −∞
M SC

Combining these expressions, we come to the exponential Fourier formula

∞ ∞
1 −itτ −ixτ
f (t ) = ∫e ∫e f (τ )dτ dx ,
SA NE

(1)
2π −∞ −∞
which is equivalent to the integral Fourier formula
U

∞ ∞
1
f (t ) =
π ∫ dx ∫ f (τ ) cos( x(τ − t ))dτ . (2)
0 −∞

Rearranging the cosine of difference, we obtain the identity

∞
f (t ) = ∫ [a ( x) cos(tx) + b( x) sin(tx)]dx , (3)
0

where

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

∞ ∞
1 1
a ( x) =
π ∫ f (t ) cos xt dt , b( x ) =
π ∫ f (t ) sin xt dt .
−∞ −∞

If f (t ) is an even function, then the formula (3) takes the form

∞ ∞
2
f (t ) =
π ∫ cos tx du ∫ f (τ ) cos xτ dτ . (4)
0 0

Similarly, if f (t ) is an odd function, then

∞ ∞
2
f (t ) =
π ∫ sin tx dx ∫ f (τ ) sin xτ dτ . (5)

S
TE S
0 0

R
AP LS
Conditions on a function f , whereby the formulae (1), (2) are valid, and the direct and
inverse Fourier transforms are determined in the following theorem, where L(−∞, +∞)
denotes a space of functions intergrable in the sense of Lebesgue over (−∞, +∞) .
C EO
Theorem 1. Assume that f ∈ L(−∞, +∞) is a function with a bounded variation on any
finite interval. Then the formulae (1), (2) are valid if we replace their left-hand sides by
[ f (t + 0) + f (t − 0)] / 2 at the points of discontinuity of f (t ) .
E –
H
PL O

2.1.1. Basic Properties of the Fourier Transform

If a function f (t ) is integrable over the interval (−∞, ∞) , then a function F ( x) exists for
M SC

all t . The functions F ( x) and f (t ) , the former being the Fourier transform of the latter,
are together called a couple of the Fourier transforms (or a Fourier transform pair).
SA NE

Assume that

∞
2
Fc ( x) =
π ∫ f (t ) cos xt dt , (6)
U

then from the formula (4) it follows that

∞
2
f (t ) =
π ∫ Fc ( x) cos tx dx . (7)
0

The functions connected in such a way are called a couple of the Fourier cosine-
transforms. Similarly, from the formula (5) we can obtain a couple of the Fourier sine-
transforms:

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

∞
2
Fs ( x) =
π ∫ f (t ) sin xt dt , (8)
0

∞
2
f (t ) =
π ∫ Fs ( x) sin tx dx . (9)
0

If f (t ) is an even function, then

F ( x) = Fc ( x) ;

if f (t ) is an odd function, then

F ( x) = iFs ( x) .

S
TE S
R
AP LS
Let the functions F ( x) and G ( x) be the Fourier transforms of functions f (t ) and g (t )
determined by the formulae (6), (7), respectively.
C EO
The functions

∞
1
F (u )G (u ) and h(t ) = ∫ g (τ ) f (t − τ )dτ
E –

2π −∞
H
PL O

are a couple of the Fourier transforms. The function h(t ) is called a convolution of
functions f (t ) and g (t ) and is denoted by as h = f ∗ g = g ∗ f .
M SC

Theorem 2 (about convolution). Let f , g ∈ L(−∞, ∞) . Then h(t ) = f ∗ g (t ) belongs

to L(−∞, ∞) , and the function 2π F ( x)G ( x) is its Fourier transform. Conversely, the
SA NE

product 2π F ( x)G ( x) belongs to L(−∞, ∞) , and its Fourier transform is f ∗ g (t ) .

The Parseval formulae. Assume that f (t ) ∈ L(−∞, ∞) , is intergrable over any finite
U

interval, and
l
1 − xτ
G ( x) =
2π l →∞
∫ g (τ )e
lim dτ
−l

for all x , moreover, G ( x) is finite everywhere and belongs to L(−∞, ∞) . Then the
Parseval equality

∞ ∞

∫ F ( x)G( x)dx = ∫ f (t ) g (−t )dt . (10)

−∞ −∞

holds. In particular, for f = g we have the Parseval formula

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

∞ ∞

∫ | F ( x) | dx = ∫ | f (t ) |
2 2
dt . (11)
−∞ −∞

2.1.2. The Multiple Fourier Transforms

By definition we have

∞ ∞
1 −ixt
F ( x) ≡ F [ f (t )] =
2π ∫ ∫e f (t )dt ,
−∞ −∞

where x = ( x1 , x2 ), t = (t1 , t2 ), xt = x1t1 + x2 t2 , dt = dt1dt2 . The function F ( x) is called

the Fourier transform of a function f (t ) of two variables. For functions f (t ) and F ( x) ,

S
TE S
that belongs to L( R 2 ) , the following inversion formula holds:

R
AP LS
f (t ) ≡ F −1
[ F ( x)] =
1
2π
∞ ∞

∫ ∫e
ixt
F ( x)dx .
C EO
−∞ −∞

If x, t ∈ R n , then
E –

1 1
e−ixt f (t )dt , f (t ) =
H

F ( x) =
(2π ) n / 2 R
∫ n ∫
(2π ) n / 2 R
n eixt F ( x)dx .
PL O
M SC

2.2. The Laplace Transform

2.2.1. The Laplace Integral

SA NE

Let f (t ) be a function of real variable t , 0 ≤ t < +∞ , intergrable in the sense of

Lebesgue over any finite interval (0, A) . Let p be a complex number. The function
∞
U

F ( p) ≡ L [ f (t )] = ∫ e− pt f (t ) dt (12)
0

is called the Laplace transform of a function f (t ) .

2.2.2. The Inversion Formula for the Laplace Transform

Using the definition of the Laplace transform and assuming that p = γ + iy , from (1),
(12) we obtain:

γ +iw +w ∞
γt
∫ e F ( p)dp = ie
pt
∫e
ity
dy ∫ e −γτ [e −γτ f (τ )]dτ . (13)
γ −iw −w 0

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

But according to (1) for w → ∞ the double integral in the right-hand side of the
equation (13) is equal to 2π e −γ t f (t ) for t > 0 and to zero for t < 0 . Therefore the
equation (13) gives

γ +iw
1
2π i w→∞ γ −∫iw
f (t ) = lim e pt F ( p) dp (14)

for t > 0 and zero for t < 0 . The expression (14) is the inversion formula for the Laplace
transform. A function f (t ) must satisfy conditions providing the existence of the
Laplace transform (12) and γ must be greater than a real part of any singular point of
the Laplace image F ( p) .

2.2.3. Limit Theorems

S
TE S
R
AP LS
The following statements (limit theorems) are valid

(1) if F ( p) is the Laplace transform and L ( f ′(t )) exists, then

C EO
lim pF ( p ) = f (0 + 0) ;
p →∞

(2) if, besides, there exist a limit of f (t ) as t → ∞ , then

E –
H

lim pF ( p ) = lim f (t ) .
PL O

p →0 t →∞
M SC

2.3. The Mellin Transform

The Mellin transform

SA NE

∞
s −1
F ( s ) ≡ M [ f (t )] = ∫ f (t )t dt , s = σ + iτ (15)
0
U

is closely connected with the Fourier and Laplace transforms.

The Mellin transform can be successfully applied when solving a certain class of plain
harmonic problems in a sectorial domain, problems of the elasticity theory, and also
when studying special functions, summing series, and calculating integrals. The
theorems concerning the Mellin transform can be obtained from the corresponding
theorems for the Fourier and Laplace transforms by the change of variables.

Laplace transforms by the change of variables.

Theorem 4. Let tσ −1 f (t ) ∈ L(0, +∞) . Then the following inversion formula holds:

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

σ +iλ
f (t + 0) + f (t − 0) 1
= lim ∫ F ( s )t − s ds , (16)
2 2π i λ →∞ σ −iλ

where the Mellin image F ( s ) is defined in (15).

Theorem 5. Let F = M [ f ], G = M [ g ] Let

either t k −1 f (t ) ∈ L(0, +∞), G (1 − k − ix) ∈ L(−∞, +∞ ) ,

or F (k + ix) ∈ L(−∞, +∞), t k g (t ) ∈ L(0, +∞) . Then

k +i∞ ∞
1
2π i k −∫i∞ ∫ f (t ) g (t )dt .
F ( s)G (1 − s)ds =

S
(17)

TE S
0

R
AP LS
Besides, the following relation holds:
C EO
k + i∞ ∞
1 ⎛ 1 ⎞ dt
∫
2π i k −i∞
F ( s )G ( s )ds = ∫ g (t ) f ⎜ ⎟ .
⎝t⎠ t
(18)
0
E –

Theorem 6 (about convolution). Assume that t k f (t ) and t k g (t ) belong to L(0, +∞)

H

and
PL O

∞
⎛ t ⎞ dτ
M SC

h(t ) = ∫ f (τ ) g ⎜⎝ τ ⎟⎠ τ .
0
SA NE

Then t k h(t ) ∈ L(0, +∞) and its Mellin transform is F ( s )G ( s ) .

-
U

-
-

TO ACCESS ALL THE 42 PAGES OF THIS CHAPTER,

Visit: http://www.eolss.net/Eolss-sampleAllChapter.aspx

Bibliography

Ditkin V.A., Prudnikov A.P. (1974). Integral Transforms and Operational Calculus. –M.: Nauka. [The
foundations of the theory of integral transforms are presented (without proofs); extensive reference tables
of integral transforms of most frequently meeting concrete functions are also cited.]
Knyazev P.N. (1969). Integral Transforms. – Minsk: Vysheishaya shkola. [This book is primarily

©Encyclopedia of Life Support Systems (EOLSS)

COMPUTATIONAL METHODS AND ALGORITHMS – Vol. I - Methods of Integral Transforms - V. I. Agoshkov, P. B.
Dubovski

concerned with the theory of integral transforms for boundary value problems of the theory of analytical
functions.]
Manzhirov A.V., Polyanin A.D. (2000). Reference Book on Integral Equations: Methods of Solution. –
M.: Factorial Press. [Along with brief results on the theory and methods for solution of integral equations,
definitions and basic properties are presented as well as the applications of some integral transforms to
the solution of integral equations; tables of integral transforms are cited.
Sneddon Ian N. (1951). Fourier Transforms. NY: McGraw Hill Book company. [The basic theorems
about the Fourier, Laplace, Hankel transforms are presented; their applications to various applied
problems of mathematical physics (to problems of the oscillation theory, heat conductivity, neutron
diffusion, hydrodynamics, atomic and nuclear physics, the elasticity theory) are considered.]
Tranter C.J. (1951). Integral Transforms in Mathematical Physics, London: Methuen IX, 118 pp. [The
basic theorems and methods of integral transforms with applications to applied problems are presented. In
each chapter exercises for a reader are given. In an appendix approximate methods for calculation of
integral transforms are considered.]
Uflyand J.S. (1963). Integral Transforms in Problems of the Elasticity Theory. – M.-L.: USSR Academy
of Sciences Publisher. [A great number of concrete problems of the elasticity theory which are solved by

S
TE S
the methods of the Fourier, Mellin, Hankel, Mehler-Foque and Kontorovich Lebedev transforms is

R
considered.]

AP LS
Voloshchuk V.M. (1984). Kinetic Theory of Coagulation. – L.: Hidrometeoizdat. [Mathematical models
of kinetics of disperse systems are described. A number of problems for the Smolukhovsky coagulation
equation and its approximations are formulated; some methods for their solution are presented.]
C EO
Zemanyan A.H. (1974). Integral Transforms of Generalized Functions. –M.: Nauka. [The theory of
integral transforms of generalized functions with applications to problems of mathematical physics is
presented.]
E –

Biographical Sketches
H
PL O

Agoshkov Valery Ivanovich is a Doctor of Physical and Mathematical Sciences, professor of Institute of
Numerical Mathematics of Russian Academy of Sciences (Moscow). He is the expert in the field of
computational and applied mathematics, the theory of boundary problems for the partial differential
M SC

equations and transport equation, the theory of the conjugate operators and their applications. He is also
the author of more than 160 research works, including 9 monographs. His basic research works are
devoted to:
- the development of the effective methods of numerical mathematics;
SA NE

- the theory of Poincaré-Steklov operators and methods of the domain decomposition;

- the development of methods of the optimal control theory and the theory of conjugate equations and
their applications in the inverse problems of mathematical physics;
- the development and justification of new iterative algorithms of the inverse problems solution;
U

- the development of the theory of functional spaces used in the theory of boundary problems for the
transport equation;
- the determination of new qualitative properties of the conjugate equations solution.

Dubovski Pavel Borisovich is a Doctor of Physical and Mathematical Sciences, professor of Institute of
Numerical Mathematics of Russian Academy of Sciences (Moscow). He is the expert in the field of
differential and integral equations, the theory of Smoluchovsky equations, mathematical modeling. He is
the author of more than 40 research works, including one monograph. His basic works are devoted to the
development of the mathematical theory of coagulation and fragmentation kinetics, including the
revelation of new kinetic models and transition to a hydrodynamic limit, the development of the theory of
integral equations and nonlinear equations in partial derivatives, and the research of some problems of
hydrodynamics.

©Encyclopedia of Life Support Systems (EOLSS)