SOLUTION SET IX FOR 18.

075FALL 2004
5. Boundary-Value Problems and Characteristic-Function Representations
5.6. Orthogonality of Characteristic Functions.
22. Reduce each of the following dierential equations to the standard form
d
dx
(p
dy
dx
) +(q +
r)y = 0:
(a) x
d
2
y
dx
2
+ 2
dy
dx
+ (x + )y = 0,
(b)
d
2
y
dx
2
+
dy
dx
cot x + y = 0,
(c)
d
2
y
dx
2
+ a
dy
dx
+ (b + )y = 0 (a, b constants),
(d) x
d
2
y
dx
2
+ (c x)
dy
dx
ay + y = 0 (a, c constants).
Solution. (a)
x
d
2
y
dx
2
+ 2
dy
dx
+ (x + )y = 0 x
2
d
2
y
dx
2
+ 2x
dy
dx
+ (x
2
+ x)y = 0

d
dx
(x
2
dy
dx
) + (x
2
+ x)y = 0
(b)
d
2
y
dx
2
+
dy
dx
cot x + y = 0 sinx
d
2
y
dx
2
+ cos x
dy
dx
+ sin x y = 0

d
dx
(sin x
dy
dx
) + sin x y = 0
(c)
d
2
y
dx
2
+ a
dy
dx
+ (b + )y = 0 e
ax
d
2
y
dx
2
+ a e
ax
dy
dx
+ (b e
ax
+ e
ax
)y = 0

d
dx
(e
ax
dy
dx
) + (b ae
ax
+ e
ax
)y = 0
Date: December 1, 2002.
1
2 SOLUTION SET IX FOR 18.075FALL 2004
(d)
x
d
2
y
dx
2
+ (c x)
dy
dx
ay + y = 0
e
x
x
c
d
2
y
dx
2
+ e
x
x
c
(
c
x
1)
dy
dx
+ (
ae
x
x
c
x
+
e
x
x
c
x
)y = 0

d
dx
(e
x
x
c
dy
dx
) + (ae
x
x
c1
+ e
x
x
c1
)y = 0
23. By considering the characteristic functions of the problem
d
2
y
dx
2
+
2
y = 0, y(0) = 0, ly

(l) + y(l) = 0,
with > 0, and using the results of Section 5.6, show that
_
l
0
sin(
1
x) sin(
2
x) dx = 0
when
1
and
2
are distinct positive solutions of the equation tan(l) + l = 0.
Solution. The general solution of
d
2
y
dx
2
+
2
y = 0, where = 0, is y = c
1
cos x + c
2
sin x.

2
(= 0) is a characteristic value only if
y(0) = 0 c
1
cos 0 + c
2
sin0 = 0, and (1)
ly

(l) + y(l) = 0 l[c

|
x=l
+ [c
1
cos(l) + c
2
sin(l)] = 0 (2)
Equation (1) implies c
1
= 0, c
2
= 0 and (2) gives lcos(l) + sin(l) = 0 which is
equivalent to
tan(l) + l = 0 (3)
So the nonzero characteristic values of the given problem are the non-zero solutions of
(3). We can easily see that if is a solution of (3) then y = sin(x) is a solution of the
given boundary-value problem. Thus, we conclude the set of nonzero characteristic values
of the given problem is = {
2
real : > 0, tan(l) + l = 0}, and for a
2
in the
corresponding characteristic function is sin(x).
Now consider the integration
_
l
0
sin(
1
x) sin(
2
x) dx, where
1
=
2
,
1
,
2
0, tan(
1
l)+
a
1
l = tan(
2
l) + a
2
l = 0. Clearly, if one of
1
,
2
= 0, then
_
l
0
sin(
1
x) sin(
2
x) dx =
_
l
0
0 dx = 0. If both
1
,
2
= 0 then sin(
1
x) and sin(
2
x) are characteristic functions of
the given problem, corresponding to dierent characteristic values. Note that in our case,
we have r(x) 1, so
_
l
0
sin(
1
x) sin(
2
x) dx =
_
1
0
r(x) sin(
1
x) sin(
2
x) dx = 0.
24. By considering the characteristic functions of the problem
x
d
2
y
dx
2
+
dy
dx
+
2
xy = 0, y(0) nite, y(l) = 0,
SOLUTION SET IX FOR 18.075FALL 2004 3
and using the results of Section 5.6, show that
_
l
0
xJ
0
(
1
x)J
0
(
2
x) dx = 0
when
1
and
2
are distinct positive roots of the equation J
0
(l) = 0.
Solution. Let X = x, then X
2
d
2
y
dX
2
+X
dy
dx
+X
2
y = 0. So the general solution of the given
equation is y = Z
0
(X) = Z
0
(x) = c
1
J
0
(x) + c
2
Y
0
(x). Then
2
(= 0) is a characteristic
value of the given boundary-value problem only if
lim
x0
[c
1
J
0
(x) + c
2
Y
0
(x)] is nite and c
1
J
0
(l) + c
2
Y
0
(l) = 0.
Since lim
x0
J
0
(x) = 1 and lim
x0
Y
0
(x) = , the niteness of the limit
lim
x0
[c
1
J
0
(x) + c
2
Y
0
(x)] implies that c
2
= 0
Hence c
1
= 0, and the second boundary condition implies J
0
(l) = 0. On the other hand,
it is clear that if > 0, and J
0
(l) = 0 then J
0
(x) is a solution of the given boundary
problem.
Now we conclude that the set of non-zero characteristic values of the given problems
is = {
2
: J
0
(l) = 0}, and for
2
in the corresponding characteristic function is
J
0
(x). Note that, in this problem, r(x) x, so for any two distinctive characteristic
values
2
1
,
2
2
we have
_
l
0
xJ
0
(
1
x)J
0
(
2
x) dx = 0. That is, if
1
,
2
> 0,
1
=
2
and
J
0
(
1
l) = 0 = J
0
(
2
l), then
_
l
0
xJ
0
(
1
x)J
0
(
2
x) dx = 0.
5.7. Expansion of Arbitrary Functions in Series of Orthogonal Functions.
31. Determine the coecients in the representation f(x) =

n=1
A
n
sin(nx), (0 < x < ),
in the following cases:
(a) f(x) = 1, (b) f(x) = x, (c) f(x) =
_
_
_
1 (x < /2),
1
2
(x = /2),
0 (x > /2)
Solution. We know that A
n
_
b
a
r(x)[
n
(x)]
2
dx =
_
b
a
r(x)f(x)
n
(x) dx since the functions
sin(nx), n = 1, 2, 3, . . . are orthogonal for 0 < x < .
4 SOLUTION SET IX FOR 18.075FALL 2004
(a) Thus
A
n
_

0
[sin(nx)]
2
dx =
_

0
sin(nx) dx, and
_

0
[sin(nx)]
2
dx =
1
2
_

0
[1 cos(2nx)] dx
=
1
2
_
x
1
2n
sin(2nx)
_

0
=

2
, and
_

0
sin(nx) dx =
_

1
n
cos(nx)
_

0
=
1
n
[1 cos(n)] .
So A
n
=
1 cos(n)
n/2
= 2[1 cos(n)]/n.
(b)
A
n
_

0
[sin(nx)]
2
dx =
_

0
x sin(nx) dx, and
_

0
[sin(nx)]
2
dx =

2
, from (a), and
_

0
xsin(nx) dx =
_

1
n
x cos(nx) +
1
n
2
sin(nx)
_

0
=

n
cos(n).
So A
n
=

n
cos(n)

2
=
2
n
cos(n) = (1)
n+1

2
n
SOLUTION SET IX FOR 18.075FALL 2004 5
(c)
A
n
_

0
[sin(nx)]
2
dx =
_
2
0
sin(nx) dx, and
_

0
[sin(nx)]
2
dx =

2
, from (a), and
_
2
0
sin(nx) dx =
_

1
n
cos(nx)
_
2
0
=
1
n
_
1 cos
n
2
_
.
So A
n
=
1
n
_
1 cos
n
2
_
/

2
=
_
1 cos
n
2
_
/
n
2
.
32. (a-c) Show that the series obtained in Problem 31 cannot be dierentiated term by
term. (Investigate the convergence of the result in each case.)
Solution.
(a) f(x) = 1 and f(x) =

n=1
2
n
[1 cos(n)] sin(nx). Dierentiating formally term
by term, we get f

(x) = 0 and f

(x) =

n=1
2[1cos(n)] cos(nx). So were investigating
0
?
=

n=1
2[1 cos(n)] cos(nx) in (0, ).
Now for 0 < x < , as n , cos(nx) 0 and for n odd, 2[1 cos(n)] = 4 = 0.
Thus, as n , 2[1 cos(n)] cos(nx) 0, so

n=1
2[1 cos(n)] cos(nx) diverges
for all 0 < x < and so indeed the series of Problem 31(a) cannot be dierentiated term
by term.
(b) f(x) = x and f(x) =

n=1
(1)
n+1 2
n
sin(nx). Dierentiating formally term by
term, we get f

(x) = 1 and f

(x) =

n=1
(1)
n+1
2 cos(nx). So were investigating
1
?
=

n=1
(1)
n+1
2 cos(nx) in (0, ).
Now for 0 < x < , as n , cos(nx) 0 and certainly 2(1)
n+1
0. Thus, as
n , (1)
n+1
2 cos(nx) 0, so

n=1
(1)
n+1
2 cos(nx) diverges for all 0 < x <
and so the series of Problem 31(b) cannot be dierentiated term by term.
6 SOLUTION SET IX FOR 18.075FALL 2004
(c) Dierentiating the series formally term by term, we get f

(x) = 0 and f

(x) =

n=1
2
_
1 cos
n
2
_
cos(nx). So were investigating
0
?
=

n=1
2
_
1 cos
n
2
_
cos(nx).
Now for 0 < x < , as n , cos(nx) 0 and for n = 4k + 2, k = 0, 1, 2, . . ., 2(1
cos
n
2
) = 4 = 0. Thus, as n , 2
_
1 cos
n
2
_
cos(nx) 0, so

n=1
2
_
1 cos
n
2
_
cos(nx)
diverges for all 0 < x < and so, indeed, the series cannot be dierentiated term by term.
33. (a) If the expansion f(x) =

n=1
A
n
sin(nx) is valid in (0, ), show formally that
_

0
[f(x)]
2
dx =

2

n=1
A
2
n
.
(b) From this result, and from the results of Problems 31(a,b), deduce the following
relations:
1
1
2
+
1
3
2
+
1
5
2
+ =

2
8
,
1
1
2
+
1
2
2
+
1
3
2
+ =

2
6
.
Solution. (a)
[f(x)]
2
=
_

n=1
A
n
sin(nx)
__

m=1
A
m
sin(mx)
_
=

n=1,m=1
A
n
A
m
sin(nx) sin(mx).
So since f(x) =

n=1
A
n
sin(nx) is valid in (0, ),
_

0
[f(x)]
2
dx =
_

0

n=1,m=1
A
n
A
m
sin nxsin mx dx
=

n=1,m=1
_

0
A
n
A
m
sin(nx) sin(mx) dx
=

n=1,m=1
A
n
A
m
_

0
sin(nx) sin(mx) dx
But
_

0
sin(nx) sin(mx) dx = 0 if m = n. So
_

0
[f(x)]
2
dx =

n=1
A
n
A
n
_

0
[sin(nx)]
2
dx
=

2

n=1
A
2
n
.
SOLUTION SET IX FOR 18.075FALL 2004 7
(b) From Problem 31(a) 1 =

n=1
2
n
[1 cos(n)] sin(nx), 0 < x < . So
_

0
1
2
dx = =

2

n=1
_
2
n
[1 cos(n)]
_
2
=

2

n=1
4

2

1
n
2
_
1 2 cos(n) +
1
2
[1 + cos(2n)]
_
=
4

n=1
1
n
2
[1 cos(n)].
So

2
4
=

n=1
1
n
2
[1 cos(n)]
=

m=0
2
(2m + 1)
2
, (n = 2m + 1)
i.e.

2
8
=
1
1
2
+
1
3
2
+
1
5
2
+
Similarly, from problem 31(b) x =

n=1
(1)
n+1

2
n
sin(nx), 0 < x < . So
_

0
x
2
dx =

3
3
=

2

n=1
_
(1)
n+1

2
n
_
2
= 2

n=1
1
n
2
.
So

2
6
=

n=1
1
n
2
=
1
1
2
+
1
2
2
+
1
3
2
+
34. Expand the function f(x) = 1 in a series of the characteristic functions of the boundary-
value problem
d
2
y
dx
2
+ y = 0, y(0) = 0, ly

(l) + ky(l) = 0 (k 0), over the interval (0, l).

x) + C
2
cos(

x) and y(0) = 0 C
2
= 0
so y(x) = C
1
sin(

x) and y

(x) = C
1

cos(

x)
and ly

(l) + ky(l) = 0 l

C
1
cos(

l) + kC
1
sin(

l) = 0
l

cos(

l) + k sin(

l) = 0
tan(

l) =
l

k
i.e. tan(x) =
x
k
, x > 0.
So there are innitely many solutions, one each on
_
k

2
, k +

2

and let
k
be the kth
solution. Thus, the characteristic functions of this boundary-value problem are
m
(x) =
sin
_

m
x
_
, where

m
> 0.
This boundary-value problem is clearly Sturm-Liouville, and r(x) = 1 so the
m
s are
orthogonal. So
1 = f(x) =

n=1
A
m

m
(x) and
A
m

_
l
0
sin
2
(
_

m
x) dx = =
_
l
0
sin(
_

m
x) dx
So
_
l
0
sin
2
(
_

m
x) dx =
1
2
_
l
0
_
1 cos(2
_

m
x)
_
dx
=
1
2
_
l
1
2

m
sin(2l
_

m
)
_
.
and
_
l
0
sin(
_

m
x) dx =
1

m
_
1 cos(l
_

m
)
_
Thus,
A
m
=
1

m
_
1 cos(l

m
)
_
1
4

m
_
2l

m
sin(2l

m
)
_ =
2
_
1 cos(l

m
)
_
l

1
2
sin(2l

m
)
.