0% found this document useful (0 votes)
208 views7 pages

Section 3.3: Fredholm Integral Equations

The document discusses Fredholm integral equations (FIEs). It defines an FIE as an equation of the form f(x) = g(x) + integral from a to b of k(x,y)f(y) dy, where k is the integral kernel. It presents a theorem stating that if k and g are continuous and the supremum of the integral of |k(x,y)| is less than 1, there exists a unique continuous solution f. It then gives two examples of solving FIEs.

Uploaded by

farhan musanna
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
208 views7 pages

Section 3.3: Fredholm Integral Equations

The document discusses Fredholm integral equations (FIEs). It defines an FIE as an equation of the form f(x) = g(x) + integral from a to b of k(x,y)f(y) dy, where k is the integral kernel. It presents a theorem stating that if k and g are continuous and the supremum of the integral of |k(x,y)| is less than 1, there exists a unique continuous solution f. It then gives two examples of solving FIEs.

Uploaded by

farhan musanna
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

Section 3.

3: Fredholm Integral Equations


Suppose that 𝓀 : [a, b] × [a, b] → R and g : [a, b] → R are given functions and that we wish
to find an f : [a, b] → R that satisfies
Z b
f (x) = g(x) + 𝓀(x, y) f (y) dy. (1)
a

Equation (1) is known as a Fredholm Integral Equation (F.I.E.) or a Fredholm Integral


Equation “of the second kind”. (F.I.E.’s of the “first kind” have g(x) = 0.) The function 𝓀
is referred to as the “integral kernel”.

The F.I.E. may be written as a fixed point equation


Tf = f
where the operator T is defined by
Z b
T f (x) = g(x) + 𝓀(x, y) f (y) dy.
a

Theorem: If 𝓀 : [a, b] × [a, b] → R and g : [a, b] → R are continuous and if


Z b
sup |𝓀(x, y)| dy < 1,
a≤x≤b a

there exists a unique continuous f : [a, b] → R that satisfies the Fredholm integral equation.

Proof:

• We will show that the sup condition implies that T is a contraction mapping in C([a, b])
(equipped with the usual uniform/sup norm). Then, since C([a, b]) is complete, we can
use the Contraction Mapping Theorem to show that there exists a unique fixed point
f ∈ C([a, b]).
• Note that
||T f1 − T f2 || = sup |T f1 (x) − T f2 (x)|
a≤x≤b

Z
b
= sup 𝓀(x, y)(f1 (y) − f2 (y)) dy

a≤x≤b a

Z b
≤ sup |𝓀(x, y)| · |(f1 (y) − f2 (y))| dy
a≤x≤b a

Z b
≤ ||f1 − f2 || sup |𝓀(x, y)| dy
a≤x≤b a
| {z }
<1

So, T is a contraction mapping.


• By the Contraction Mapping Theorem, the equation T f = f , and therefore the F.I.E.,
has a unique solution in C([a, b]). t
u

We now know that, if the conditions of the previous theorem are satisfied, we may solve (??)
by choosing any f0 = C([a, b]) and computing

f = lim T n f0 .
n→∞

The Fredholm Integral Operator, denoted by K, is defined as on functions f ∈ C([a, b])


as Z b
Kf := 𝓀(x, y) f (y) dy
a
where 𝓀 is an F.I.E. kernel. Note that K is a linear operator.
The F.I.E. is then written
f = g + Kf
which can also be written
T f = g + Kf
using the fixed point equation T f = f .
Note that
T f0 = g + Kf0

T 2 f0 = T (T f0 ) = T (g + Kf0 ) = g + K(g + Kf0 ) = g + Kg + K 2 f0

T 3 f0 = T (T 2 f0 ) = g + Kg + K 2 g + K 3 f0

..
.
T n f0 = g + Kg + K 2 g 2 + · · · + K n−1 g n−1 + K n f0

lim K n f0 = 0.
On HW 6, we will see that n→∞
Thus ∞
f = lim T n f0 = K n g.
X
n→∞
n=0
Example: Solve the Fredholm Integral Equation
Z 1
f (x) = 1 + x f (y) dy.
0

Note that Z b Z 1
sup |𝓀(x, y)| dy = sup x dy = 1.
a≤x≤b a 0≤x≤1 0

We need this strictly less than 1 in order to use our Theorem from page 1. To this end, we
will “back off of 1” a little bit and consider solving.
Z α
f (x) = 1 + x f (y) dy (2)
0

for 0 < α < 1.


Note that now Z b Z α
sup |𝓀(x, y)| dy = sup x dy = α2 < 1.
a≤x≤b a 0≤x≤α 0

Furthermore g(x) = 1 and 𝓀(x, y) = x are continuous functions on [0, 1] so all of the
conditions of the Theorem on page 1 are satisfied.
So, we may start with any f0 ∈ C([0, 1]) and repeatedly apply T where T f (x) = 1 +

0 x f (y) dy.

Let f0 (x) = 1. Then


Z α Z α
f1 (x) = T f0 (x) = 1 + x f0 (y) dy = 1 + x dy = 1 + αx,
0 0

and Rα Rα
f2 (x) = T f1 (x) = 1 + 0 x f1 (y) dy = 1 + 0 x(1 + αy) dy
h i
= 1 + x α + 21 α3
and Rα Rα
f3 (x) = T f2 (x) = 1 + 0 x f2 (y) dy = 1 + 0 x[1 + y(α + 21 α3 )] dy
h i
= 1 + x α + 21 α3 + 1 5
22
α .
Continuing, we get
1 1 1
 
n−1
fn (x) = 1 + x α + α3 + 2 α5 + · · · + n−1 α2 +1 .
2 2 2
Therefore, P∞ 1 2n+1
f (x) = limn→∞ fn (x) = 1 + x n=0 2n α

 n
P∞ α2 1
= 1 + xα n=0 2
= 1 + xα · 1−α2 /2


= 1+ 2−α2
x
It is easy to check that this satisfies √
the √
given F.I.E. Note that the sum in that second to
last line is still√convergent for α ∈ (− 2, 2) and furthermore that the solution satisfies (2)
for any α 6= ± 2 !

Example: Solve the Fredholm Integral Equation


Z π/2
f (x) = sin x + sin x cos y f (y) dy.
0

Note first that Z b Z π/2


sup |𝓀(x, y)| dy = sup | sin x cos y| dy
a≤x≤b a 0≤x≤π/2 0
Z π/2
= sup sin x cos y dy = sup sin x = sin(π/2) = 1 6< 1
0≤x≤π/2 0 0≤x≤π/2

However, in light of the comments at the end of the previous example, we are going to try
to leave the π/2 in place.

Let f0 (x) = 1.
Then R π/2
f1 (x) = sin x + 0 sin x cos y · f0 (y) dy
R π/2
= sin x + 0 sin x cos y · 1 dy
R π/2
= sin x + sin x 0 cos y dy

= sin x + sin x · 1 = 2 sin x.

Now,
R π/2
f2 (x) = sin x + 0 sin x cos y · f1 (y) dy
R π/2
= sin x + 0 sin x cos y · 2 sin y dy
R π/2
= sin x + 2 sin x 0 sin y cos y dy
| {z } | {z }
u du

π/2
= sin x + 2 sin x · 1
sin2 y

2 0

1
= sin x + 2 sin x · 2
= 2 sin x.

So, we have already reached our fixed point! That is, fn (x) = 2 sin x for n = 1, 2, . . .. Thus,
we have
f (x) = n→∞
lim fn (x) = 2 sin x.
It is easy to see/verify that this satisfies the given F.I.E.

Section 3.4: Boundary Value Problems


In this section we wish to find solutions to the boundary value problem (BVP) given by

u00 (x) = q(x)u(x), 0 < x < 1

u(0) = u0 , u(1) = u1

When q(x) is constant, the solution is easy. Recall that for a second order differential
equation of the form
au00 (x) + bu0 (x) + cu(x) = 0
one first finds roots r1 and r2 for the auxiliary equation

au2 + bu + c = 0.

Then

• If r1 and r2 are real and distinct, the solution has the form

u(x) = c1 er1 x + c2er2 x .

• If the roots are real and repeated (r1 = r2 = r), the solution has the form

u(x) = c1 erx + c2 xerx .

• If the roots are complex (r1 = a + ib, r2 = a − ib), the solution has the form

u(x) = c1 eax cos(bx) + c2 eax sin(bx).

Non-constant q(x) is more difficult and is the point of this Section.

To solve
u00 (x) = q(x)u(x), 0 < x < 1

u(0) = u0 , u(1) = u1
we begin by zeroing out the boundary conditions and considering the function

v(x) := u(x) − u0 + (u0 − u1 )x.

Note that v 00 (x) = u00 (x) and that

q(x)u(x) = q(x)v(x) + q(x)[u0 + (u1 − u0 )x].

Our new BVP is given by

v 00 (x) = q(x)v(x) + q(x)[u0 + (u1 − u0 )x]

v(0) = 0, v(1) = 0

On our road to a solution, we first consider something simpler.

Theorem: Let f : [0, 1] → R be continuous.


The unique solution of the BVP

v 00 (x) = −f (x), v(0) = 0, v(1) = 0

is given by Z 1
v(x) = g(x, y) f (y) dy
0
where (
x(1 − y) , 0 ≤ x ≤ y ≤ 1
g(x, y) =
y(1 − x) , 0 ≤ y ≤ x ≤ 1.

Proof:

• Note that Z y Z y
v 00 (s) = −f (s) ⇒ v 00 (s) ds = − f (s) ds
1 1
Z y
⇒ v 0 (y) = − f (s) ds + c1
1

• So, Z x Z xZ y
0
v (y) dy = − f (s) ds dy + c1 x
0 0 1
RxRy
which ⇒ v(x) = − 0 1 f (s) ds dy + c1 x + c2 .
Ry
• Integrating by parts with “u” = 1 f (s) ds and “dv” = dy (so “du” = f (y) dy and
“v” = y) we get
n o
f (s) ds]y=x
Ry Rx
v(x) = − [y 1 y=0 − 0 yf (y) dy + c1 x + c2
Rx Rx
= − [x 1 f (s) ds − 0 yf (y) dy] + c1 x + c2
Rx Rx
= − [x 1 f (y) dy − 0 yf (y) dy] + c1 x + c2
h R1 Rx i
= − −x x f (y) dy − 0 yf (y) dy + c1 x + c2

for x ∈ [0, 1].

• Now the boundary conditions give

v(0) = c2 = 0
R1 R1
v(1) = 0 yf (y) dy + c1 = 0 ⇒ c1 = − 0 yf (y) dy

• Thus, we have that


R1 Rx R1
v(x) = x x f (y) dy + 0 yf (y) dy − 0 yf (y) dy
Rx R1
= 0 y(1 − x)f (y) dy + x x(1 − y)f (y) dy
R1 √
= 0 g(x, y)f (y) dy

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy