Partial Differential Equations (Math 3303)



A®‚Ë @ Õæ… Aë áÖß

@ . X . @

Chapter 6
Fourier Integrals and Fourier Transforms

In this chapter we discuss methods to solve partial differential equation in infinite domains.

6.1 Fourier Integrals

In earlier chapters, we have described Fourier series for periodic functions. However, functions which
are not periodic cannot be represented by Fourier series. In many problems of physical interest, it
is desirable to develop an integral representation for such a function that is analogous to a Fourier
series.

Definition. (Fourier integrals )

Let f be a function and let
1 ∞
Z
A(α) = f (x) cos(αx)dx,
π −∞
1 ∞
Z
B(α) = f (x) sin(αx)dx.
π −∞

Then the integral Z ∞

[A(α) cos(αx) + B(α) sin(αx)] dα
0

is called the Fourier integral formula for f (x).

Theorem 1.Z Assume that f a piecewise smooth function on every finite interval [a, b] ⊆ R and
∞
1
assume that |f (x)|dx converges. Then the Fourier integral of f converges to [f (x+) + f (x−)]
−∞ 2
for all x ∈ R; that is
Z ∞
1
[A(α) cos(αx) + B(α) sin(αx)] dα = [f (x+) + f (x−)], ∀x ∈ R.
0 2
Example 1. Consider the function

 0, x < 0,
f (x) = x, 0 < x < 1
0, x > 1.


1
 (a) Graph f (x).

(b) Find the Fourier integral formula for f (x).

(c) Determine the convergence of the Fourier integral in part (b) at x = 1.

Solution:

Remarks.

(1) If f (x) is an even function, then B(α) = 0 and

2 ∞
Z
A(α) = f (x) cos(αx)dx.
π 0

(2) If f (x) is an odd function, then A(α) = 0 and

2 ∞
Z
B(α) = f (x) sin(αx)dx.
π 0

Example 1. Consider the function



 0, x < −π,
−1, −π < x < 0

f (x) =

 1, 0 < x < π,
0, x > π.


(a) Graph f (x).

(b) Find the Fourier integral representation for f (x).

(c) Determine the convergence of the Fourier integral in part (b) at x = −π.

Solution:

Application of Fourier integrals to partial differential equa-

tions
Given an problem that is defined for x in an infinite interval, there are five basic steps in solving the
problem by the Fourier integrals:

(1) Use separation of variables to convert the partial differential equation into two ordinary differ-
ential equations.

(2) Solve the boundary value problem for X(x) and find the eigenvalues and eigenfunctions.

(3) Solve the other ordinary differential equation.

(4) Use superposition to write u as a Fourier integral.

(5) Find the constants in the Fourier integral.

2
Example 1. Use Fourier integral to solve the initial-boundary value problem

ut = uxx , 0 < x < ∞, t > 0,

u(0, t) = 0, t ≥ 0,
u(x, 0) = f (x),

 1, 0 < x < 1,
where f (x) =
0, 1 < x < ∞.


Solution:

Example 2. Use Fourier integral to solve the initial-boundary value problem

utt = 4uxx , 0 < x < ∞, t > 0,

ux (0, t) = 0, t ≥ 0,
u(x, 0) = e−x , ut (x, 0) = 0, 0 < x < ∞.

Solution:

Exercises
(1) Find the Fourier integral formula for each of the following functions:


 0, x < π,
x, −π < x < 0

(a) f (x) =

 1, 0 < x < π
0 x > π.


x + 1, |x| < π,
(b) f (x) =
0, |x| > π.

sin x, |x| < π,
(c) f (x) =
0, |x| > π.

−1, −a < x < 0,
(2) Let f (x) = Show that f (x) has the Fourier sine integral repre-
1, 0 ≤ xZ< a0, otherwise.
2 ∞1
sentation f (x) ∼ [1 − cos(aα)] sin(xα)dα.
π 0 α

0, x < 0,
(3) Let f (x) =
cos x, 0 ≤ x < π0, x > π.

(a) Show that f (x) has the Fourier integral representation

1 ∞
Z
α
f (x) ∼ [sin α(π − x) − sin(xα)]dα.
π 0 (1 − α2 )
Z ∞
α sin(xα) π
(b) When x = 0, show that 2
dα = .
0 (1 − α ) 2

3
(4) Use Fourier integral to solve the initial-boundary value problem

utt = uxx , 0 < x < ∞, t > 0,

u(0, t) = 0, t ≥ 0,
1
u(x, 0) = , ut (x, 0) = 0.
x2 +1

(5) Use Fourier integral to solve the initial-boundary value problem

ut = uxx , 0 < x < ∞, t > 0,

u(0, t) = 0, t ≥ 0,

 x, 0 < x < 2,
u(x, 0) =
0, 2 < x < ∞.


6.2 Fourier Sine and Cosine Transforms

The Fourier sine and cosine are particularly appropriate in solving boundary-value problems for
semi-infinite regions. We first give a formal definition of the Fourier sine and cosine transforms.

Definition. (Fourier sine and cosine Transforms) Z ∞

Let f (x) be a continuous and piecewise smooth function such that |f (x)|dx converges.
0

(1) The Fourier cosine transform of f is

r Z ∞
2
Fc {f (x)} = f (x) cos(xα)dx = Fc (α), α>0
π 0
and its inverse is
r Z ∞
2
f (x) = Fc−1 {Fc (α)} = Fc (α) cos(xα)dα, x > 0.
π 0

(2) The Fourier sine transform of f is

r Z ∞
2
Fs {f (x)} = f (x) sin(xα)dx = Fs (α), α>0
π 0
and its inverse is
r Z ∞
2
f (x) = Fs−1 {Fc (α)} = Fs (α) sin(xα)dα, x > 0.
π 0

Example 1. Find the Fourier sine and cosine transforms of f (x) = e−kx , k > 0.

Solution:

4
Theorem 2. ( Fourier sine and cosine transforms for derivatives)
Assume that f, f 0 and f 00 are continuous and
Z ∞ Z ∞ Z ∞
0
|f (x)|dx, |f (x)|dx, |f 00 (x)|dx
0 0 0

converge. Moreover assume that

lim f (x) = lim f 0 (x) = 0.

x→∞ x→∞

Then
r
2 0
(1) Fc {f 00 (x)} = − f (0) − α2 Fc {f (x)},
π
r
2
(2) Fs {f 00 (x)} = α f (0) − α2 Fs {f (x)}.
π
Proof. Integration by parts.
r
00 2 0
Example 1. Use Fc {f (x)} = − f (0) − α2 Fc {f (x)} to find Fc {e−x }.
π
Solution:

Example 2. Solve the integral equation


Z ∞  1, 0 < α < π,
f (x) cos(xα)dx =
0 
0, π < α < ∞.

Solution:

Example 3. Solve the integral equation


Z ∞  α, 0 < α < π,
f (x) sin(xα)dx =
0 
0, π < α < ∞.

Solution:

Properties of Fourier Sine and Cosine Transforms

Theorem 3. ( The Fourier sine and cosine transforms are linear)
Let f and g be functions and let a and b be real numbers. Then

(a) Fs {af + bg} = aFs {f } + bFs {g},

(b) Fc {af + bg} = aFc {f } + bFc {g}.

Proof.

5
Application to initial-boundary value problems
Given an problem that is defined for x ∈ (0, ∞) there are three basic steps in solving the problem
by the Fourier sine or cosine transform:
(1) Apply the transform to the equation and to the given conditions to transform the problem.

(2) Solve the transformed problem to find the transform of the solution u.

(3) Find the inverse of the transform obtained in step (2).

Example 1. Use Fourier sine transform to solve the initial-boundary value problem

ut = uxx , 0 < x < ∞, t > 0,

u(0, t) = 0, t ≥ 0,
u(x, 0) = f (x),

 1, 0 < x < 1,
where f (x) =
0, 1 < x < ∞.


Solution:
Example 2. Use Fourier cosine transform to solve the boundary value problem

uxx + uyy = 0, 0 < x < ∞, 0 < y < 2,

ux (0, y) = 0, 0 < y < 2,
u(x, 0) = 0, 0 < x < ∞,
u(x, 2) = f (x), 0 < x < ∞,

 x, 0 < x < 1,
where f (x) =
0, 1 < x < ∞.


Solution:

Exercises
(1) Find the Fourier sine and cosine transforms of the following functions:

(a) f (x) = xe−x .

(b) f (x) = e−x cos x.
(c) f (x) = e−x sin x.

(2) Solve the following integral equations

Z ∞
(a) f (x) sin(xα)dx = e−α .
0

Z ∞  1 − α, 0 < α < 1,
(b) f (x) sin(xα)dx =
0 
0, 1 < α < ∞.

6
 (3) Use Fourier cosine transform to solve the boundary value problems:
(a)
utt = uxx , 0 < x < ∞, t > 0,
u(0, t) = 0, t ≥ 0,
u(x, 0) = xe−x , ut (x, 0) = 0 0 < x < ∞.
(b)
ut = uxx , 0 < x < ∞, t > 0,
ux (0, t) = 1, t ≥ 0,
u(x, 0) = 0, 0 < x < ∞..
(3) Use Fourier sine transform to solve the boundary value problems:
(a)
ut = uxx , 0 < x < ∞, t > 0,
u(0, t) = 1, t ≥ 0,
u(x, 0) = 0, 0 < x < ∞.
(b)
uxx + uyy = 0, 0 < x < ∞, 0 < y < 1,
ux (0, y) = y(1 − y), 0 < y < 2,
u(x, 0) = u(x, 1) = 0, 0 < x < ∞.
(c)
ut − uxx + tu = 0, 0 < x < ∞, 0 < y < 1,
ux (0, t) = 0, t ≥ 0,
−x
u(x, 0) = e , 0 < x < ∞.

6.3 Fourier Transforms

Fourier Transform is an extremely powerful mathematical tool for the analysis of non-periodic func-
tions. The Fourier transform is of fundamental importance in a broad range of applications, including
both ordinary and partial differential equations, quantum mechanics, signal and image processing,
control theory, and probability, to name but a few. We first give a formal definition of the Fourier
transform by using the complex Fourier integral formula.

Definition. (Fourier Transform) Z

∞
Let f (x) be a function such that |f (x)|dx exists.
−∞

(1) The Fourier transform of f is

Z ∞
1
F{f (x)} = √ f (x)e−iαx dx = F (α), α ∈ R.
2π −∞

(2) The inverse Fourier transform of a Fourier transform F (α) is

Z ∞
−1 1
F {F (α)} = √ F (α)eiαx dα ∼ f (x), x ∈ R.
2π −∞

7
 
 1, |x| < k,
Example 1. Find the Fourier transform of f (x) =
0, |x| > k.


Solution:
∞
Z r
−ky 2 π 2
Example 2. Use e dy = to find the Fourier transform of f (x) = e−kx , k > 0.
−∞ k
Solution:

Properties of Fourier Transforms

Theorem 4. ( The Fourier transform and its inverse are linear)
Let f and g be functions with Fourier transforms F (α) and G(α) respectively. Then for any real
numbers a and b we have

(a) F{af (x) + bg(x)} = aF{f (x)} + bF{g(x)},

(b) F −1 {aF (α) + bG(a)} = aF −1 {F (α)} + bF −1 {G(α)}.

Proof. By linearity of integration.

Theorem 5. (Scaling )
Let F{f (x)} be the Fourier transform of a function f (x). Then

F{f (x − c)} = e−iαc F{f (x)},

where c is a real constant.

Proof.

Theorem 6. ( Differentiation )
Let f, f 0 , . . . , f (n−1) be continuous and let f (n) be piecewise continuous on R. Assume that
f (x), f 0 (x), . . . , f (n−1) (x) approach zero as |x| → ∞. If f, f 0 , . . . , f (n) are absolutely integrable, then

F{f (n) (x)} = (iα)n F{f (x)}, n ∈ N.

Proof. Integration by parts and induction.

Solving problems by Fourier transforms

Given an problem that is defined for x ∈ R there are three basic steps in solving the problem by the
Fourier transform:

(1) Apply the Fourier transform to the equation and to the given conditions to transform the
problem.

(2) Solve the transformed problem to find the Fourier transform of the solution u.

(3) Find the inverse of the Fourier transform obtained in step (2).

8
Example 1. Use Fourier transform to solve the initial-boundary value problem
ut = uxx , −∞ < x < ∞, t > 0,
u(x, 0) = f (x),

 1, |x| < 1,
where f (x) =
0, |x| > 1.


Solution:
Example 2. Use Fourier transform to solve the boundary value problem
uxx + uyy = 0, −∞ < x < ∞, 0 < y < ∞,
u(x, 0) = f (x), −∞ < x < ∞,
u is bounded as y → ∞.
Solution:

Exercises
(1) Find the Fourier transforms of the following functions:
(a) f (x) = e−|x| .
2
(b) f (x) = e−ax .
1
(c) f (x) = .
|x|
(d) f (x) = cos x2 .
1
(e) f (x) = 2 .
x + a2

 1 − |x|, |x| < 1,
(f) f (x) =
0, |x| > 1.


(2) Use Fourier transform to solve the initial-boundary value problem

utt − uxx = 0, −∞ < x < ∞, t > 0,

u(x, 0) = f (x), −∞ < x < ∞,
ut (x, 0) = g(x) −∞ < x < ∞.

(3) Use Fourier transform to solve the initial-boundary value problem

ut − uxx = 0, −∞ < x < ∞, t > 0,

2
u(x, 0) = e−x , −∞ < x < ∞.

(4) Use Fourier transform to solve the initial-boundary value problem

ut − uxx − u = 0, −∞ < x < ∞, t > 0,

u(x, 0) = f (x), −∞ < x < ∞.

9
(5) Use Fourier transform to solve the boundary value problem

uxx + uyy = 0, −∞ < x < ∞, 0 < y < 1,

−2|x|
u(x, 0) = e , u(x, 1) = 0, −∞ < x < ∞.

(6) Use Fourier transform to solve the boundary value problem

uxx + uyy = u, −∞ < x < ∞, 0 < y < 1,

−x2
uy (x, 0) = 0, u(x, 1) = e , −∞ < x < ∞.

(7) Use Fourier transform to solve the boundary value problem

2
uxx + uyy = e−x , −∞ < x < ∞, 0 < y < 1,
u(x, 0) = u(x, 1) = 0, −∞ < x < ∞.

10