MTH618 Wk 11 Tutorial

Sturm-Liouville Problems
Find the eigenvalues and eigenfunctions. Verify orthogonality. Show details of your
work.

1. y 00 + λy = 0, y(0) = 0, y(10) = 0.
2. y 00 + λy = 0, y(0) = 0, y(L) = 0.
3. y 00 + λy = 0, y(0) = 0, y 0 (L) = 0.

Fourier Integrals
1. Show that the integrals represents the indicated function. Hint: Use the Fourier
integral and the Fourier integral coefficients formulae; the integral tells you which
one, and its value tells you what function to consider.

´ ∞ cos xw+w sin xw 0,
 x<0
π
(a) 0 1+w2
dx = 2 , x=0
πe−x , x > 0


´ ∞ sin πw sin xw
(
π
2
sin x, 0 ≤ x ≤ π
(b) 0 1−w 2 dw =
0, x>π
´∞
(
π
, 0<x<π
(c) 0 1−cosw
πw
sin xw dw = 2
0, x > π

2. Fourier Cosine Integral Representations. Represent f (x) as an integral.

(
1, 0<x<1
(a) f (x) =
0, x > 1.
(
x2 , 0<x<1
(b) f (x) =
0, x > 1.

3. Fourier Sine Integral Representations. Represent f (x) as an integral.

(
x, 0 < x < a
(a) f (x) =
0, x > a.
(
1, 0 < x < 1
(b) f (x) =
0, x > 1.

1
 MTH618 Wk 11 Tutorial Solutions

Sturm-Liouville Problems
1. Note that p = 1, q = 0, r = 1 and a = 0, b = 10, and from the conditions, we
ascertain
y(0) = 0 =⇒ k1 = 1, k2 = 0
y(10) = 0 =⇒ l1 = 1, l2 = 0.
We have the characteristic equation
µ2 + λ = 0
from the ODE.
• For negative λ = −v 2 , the general solution is
y(x) = c1 evx + c2 e−vx .
From the boundary conditions, we obtain c1 = c2 = 0, so y(x) = 0 (useless
result).
• For λ = 0, the general solution is

y = c1 x + c2 .
From the boundary conditions, we obtain c1 = c2 = 0, so y(x) = 0 (useless
result).
• For positive λ = v 2 , the general solution is

y(x) = A cos vx + B sin vx.

Applying the initial conditions
y(0) = 0 =⇒ A=0
y(10) = 0 =⇒ 0 = y(10) = B sin 10v,
hence
10v = nπ
nπ
v = .
10
Taking B = 1, we obtain
nπx


y(x) = sin 10
, n = 1, 2, · · ·

1
 Therefore, the eigenvalues of the problem are
 nπ 2
λn = v 2 =
10
with corresponding eigenfunctions
 nπx 
yn (x) = sin
10
where n = 1, 2, · · · .
2. Generalising the procedure followed in No. 1 above, we will obtain the eigenvalues
 nπ 2
2
λn = v =
L
with corresponding eigenfunctions
 nπx 
yn (x) = sin
L
where n = 1, 2, · · · .
3. Note that p = 1, q = 0, r = 1 and a = 0, b = L, and from the conditions, we
ascertain
y(0) = 0 =⇒ k1 = 1, k2 = 0
y 0 (L) = 0 =⇒ l1 = 0, l2 = 1.
We have the characteristic equation
µ2 + λ = 0
from the ODE.
• For negative λ = −v 2 , the general solution is
y(x) = c1 evx + c2 e−vx .
From the boundary conditions, we obtain c1 = c2 = 0, so y(x) = 0 (useless
result).
• For λ = 0, the general solution is

y = c1 x + c2 .
From the boundary conditions, we obtain c1 = c2 = 0, so y(x) = 0 (useless
result).
• For positive λ = v 2 , the general solution is

y(x) = A cos vx + B sin vx.

Applying the initial conditions
y(0) = 0 =⇒ A=0

2
and
y 0 = Bv cos vx
y 0 (L) = 0 : 0 = Bv cos vL.

Note that B 6= 0 and v 6= 0, otherwise y = 0. Hence

cos vL = 0.
Property of cosines:  
2n − 1
cos π =0
2
so
2n − 1
vL = π
2
(2n − 1)π
v =
2L
for n = 1, 2, 3, . . ..
Taking B = 1, we obtain
 
(2n−1)πx
y(x) = sin 2L
, n = 1, 2, · · ·

Therefore, the eigenvalues of the problem are

 2
2 (2n − 1)π
λn = v =
2L
with corresponding eigenfunctions
 
(2n − 1)πx
y(x) = sin
2L
where n = 1, 2, · · · .

3
Fourier Integrals
1. For 
ˆ ∞ 0,
 x<0
cos xw + w sin xw π
dw = 2 , x=0,
0 1 + w2 πe−x , x > 0


π
bringing 2
from RHS to LHS of the equal sign, we obtain

ˆ ∞ 0,
 x<0
2 cos xw + w sin xw
dw = 1, x=0.
π 0 1 + w2 2e−x , x > 0


Notice that the RHS is f (x) defined over the complete x-axis (on the interval
−∞ < x < ∞): 
0,
 x<0
f (x) = 1, x=0
2e−x , x > 0.


Also, note that f (x) is neither odd nor even. Rearranging the terms to obtain the
coefficients of cos wx and sin wx, we obtain

ˆ ∞    0,
 x<0
2 1 2 w
cos xwdw + sin xw dw = 1, x=0,
0 π 1 + w2 π 1 + w2 2e−x ,

x>0

then comparing with the general form of Fourier Integral, we see that
2 1
A(w) =
π 1 + w2
2 w
B(w) = .
π 1 + w2
We are left to verify that
ˆ ∞
1 2 1
A(w) = f (v) cos wv dv =
π −∞ π 1 + w2
and ˆ ∞
1 2 w
B(w) = f (v) sin wv dv = .
π −∞ π 1 + w2

4
Integrating,
ˆ
1 ∞
A(w) = f (v) cos wv dv
π −∞
ˆ
1 +∞ −v
= (2e ) cos wv dv
π 0
ˆ
2 −∞ −v
= e cos wv dv
π 0
∞
2 e−v (w sin wv − cos wv)

=
π w2 + 1 0
2 1
= .
π w2 + 1
and
ˆ
1 ∞
B(w) = f (v) sin wv dv
π −∞
ˆ
1 +∞ −v
= (2e ) sin wv dv
π 0
ˆ ∞
2
= e−v sin wv dv
π 0
∞
2 e−v (sin wv + w cos wv)

=
π w2 + 1 0
2 w
= .
π w2 + 1

5
2. For ˆ (
∞ π
sin πw sin xw 2
sin x, 0 ≤ x ≤ π
dw = ,
0 1 − w2 0, x>π
π
bringing 2
from RHS to LHS of the equal sign, we obtain
ˆ (
2 ∞ sin πw sin xw sin x, 0 ≤ x ≤ π
2
dw = .
π 0 1−w 0, x>π

Notice that the RHS shows f (x) defined just over the interval 0 ≤ x < ∞ (this
is just half of the function since f (x) on the interval −∞ < x < 0 is missing and
this is determined by even or odd function property):
(
sin x, 0 ≤ x ≤ π
f (x) =
0, x > π.

Notice that cos wx is missing from the Fourier Integral which means A(w) = 0,
reducing this to a Fourier Sine Integral. This, in turn, implies that f (x) is an odd
function, hence 
0,
 x < −π
f (x) = sin x, −π ≤ x ≤ π .

0, x > π.
Rearranging the terms to obtain the coefficients of sin wx, we obtain
ˆ ∞  (
2 sin πw sin x, 0 ≤ x ≤ π
2
sin wx dw =
0 π1−w 0, x>π

which means
2 sin πw
B(w) = .
π 1 − w2
We are left to verify that
ˆ ∞
2
A(w) = f (v) cos wv dv = 0
π 0

and ˆ
2 ∞ 2 sin πw
B(w) = f (v) sin wv dv = .
π 0 π 1 − w2
The verification is left as an exercise to complete.

6
3. For ˆ (
∞ π
1 − cos πw 2
, 0<x<π
sin xw dw = ,
0 w 0, x>π
π
bringing 2
from RHS to LHS of the equal sign, we obtain

ˆ ∞ 1,
 0<x<π
2 1 − cos πw
sin xw dw = 1/2, x = π .
π 0 w 
0, x>π

Notice that the RHS is f (x) defined on the interval 0 < x < π:
(
1, 0 < x < π
f (x) =
0, x > π

Notice that cos xw is missing from the Fourier Integral which means A(w) = 0,
reducing this to a Fourier Sine Integral, hence f (x) is an odd function. So


 0, x < −π

−1, −π < x < 0
f (x) = .

 1, 0<x<π

0, x>π


Rearranging the terms to obtain the coefficients of sin xw, we obtain


ˆ ∞  1,
 0<x<π
2 1 − cos πw
sin xw dw = 1/2, x = π
0 π w 
0, x>π

then comparing with the general form of Fourier Sine Integral, we see that
2 1 − cos πw
B(w) = .
π w
We are left to verify that
ˆ ∞
2
A(w) = f (v) cos wv dv = 0
π 0

and ˆ
2 ∞ 2 1 − cos πw
B(w) = f (v) sin wv dv = .
π 0 π w
The verification is left as an exercise to complete.

7
Fourier Cosine Integral Representations
1. Note that (
1, 0 < x < 1
f (x) =
0, x > 1.
is viewed as an even function, so B(w) = 0 and we are left to find A(w).
ˆ
2 ∞
A(w) = f (v) cos wv dv
π 0
ˆ
2 1
= (1) cos wv dv
π 0
 1
2 1
= sin wv
π w 0
2 sin w
= .
π w
Then the Fourier Cosine Integral of f (x) is
ˆ ∞
A(w) cos wxdw = f (x)
0 
ˆ ∞  1,
 0<x<1
2 sin w 1
cos wx dw = 2
, x=1
0 π w 
0, x > 1.

ˆ ∞  π
2
, 0<x<1
sin w cos wx 
π
dw = 4
, x=1
0 w 
0, x > 1.

8
Fourier Sine Integral Representations
1. Note that (
x, 0 < x < a
f (x) =
0, x > a.
is viewed as an odd function, so A(w) = 0 and we are left to find B(w).
ˆ
2 ∞
B(w) = f (v) sin wv dv
π 0
ˆ
2 a
= (v) sin wv dv
π 0
 a
2 v 1
= − cos wv + 2 sin wv
π w w
 0
2 a 1
= − cos wa + 2 sin wa .
π w w
Then the Fourier Sine Integral of f (x) is
ˆ ∞
B(w) sin wxdw = f (x)
0 
ˆ ∞   x,
 0<x<a
2 a 1 1
− cos wa + 2 sin wa sin wx dw = 2
a, x=a
0 π w w 
0, x > a.

ˆ ∞   π
 2 x, 0<x<a
a 1 π
− cos wa + 2 sin wa sin wx dw = 4
a, x=a
0 w w 
0, x > a.