AEMC13

13 Boundary-Value Problems in
Rectangular Coordinates
Exercises 13.1
1. If u = XY then
u
x
= X
Y,
u
y
= XY
,
X
Y = XY
,
and
X
X
=
Y
Y
=
2
.
Then
X
2
X = 0 and Y
2
Y = 0
so that
X = A
1
e
2
x
,
Y = A
2
e
2
y
,
and
u = XY = c
1
e
c2(x+y)
.
2. If u = XY then
u
x
= X
Y,
u
y
= XY
,
X
Y = 3XY
,
and
X
3X
=
Y
Y
=
2
.
Then
X
3
2
X = 0 and Y
2
Y = 0
so that
X = A
1
e
3
2
x
,
Y = A
2
e
2
y
,
and
u = XY = c
1
e
c2(y3x)
.
3. If u = XY then
u
x
= X
Y,
u
y
= XY
,
X
Y = X(Y Y
),
and
605
Exercises 13.1
X
X
=
Y Y
Y
=
2
.
Then
X
2
X = 0 and Y
(1
2
)Y = 0
so that
X = A
1
e
2
x
,
Y = A
2
e
(1
2
)y
,
and
u = XY = c
1
e
y+c2(xy)
.
4. If u = XY then
u
x
= X
Y,
u
y
= XY
,
X
Y = X(Y +Y
),
and
X
X
=
Y +Y
Y
=
2
.
Then
X
2
X = 0 and Y
(1
2
)Y = 0
so that
X = A
1
e
2
x
,
Y = A
2
e
(1
2
)y
,
and
u = XY = c
1
e
y+c2(x+y)
.
5. If u = XY then
u
x
= X
Y,
u
y
= XY
,
xX
Y = yXY
,
and
xX
X
=
yY
Y
=
2
.
Then
X
1
x
2
X = 0 and Y
1
y
2
Y = 0
so that
X = A
1
x
2
,
Y = A
2
y
2
,
and
u = XY = c
1
(xy)
c2
.
606
Exercises 13.1
6. If u = XY then
u
x
= X
Y,
u
y
= XY
,
yX
Y = xXY
,
and
X
xX
=
Y
yY
=
2
.
Then
X
2
xX = 0 and Y
2
yY = 0
so that
X = A
1
e
2
x
2
/2
,
Y = A
2
e
2
y
2
/2
,
and
u = XY = c
1
e
c2(x
2
y
2
)
.
7. If u = XY then
u
xx
= X
Y, u
yy
= XY
, u
yx
= X
,
and
X
Y +X
+XY
= 0,
which is not separable.
8. If u = XY then
u
yx
= X
,
yX
+XY = 0,
and
X
X
=
Y
yY
=
2
.
Then
X
2
X = 0 and
2
yY
Y = 0
so that
X = A
1
e
2
x
,
Y = A
2
y
1/
2
,
and
u = XY = c
1
e
c2x
y
1/c2
.
9. If u = XT then
u
t
= XT
,
u
xx
= X
T,
kX
T XT = XT
,
and we choose
T
T
=
kX
X
X
= 1 k
2
607
Exercises 13.1
so that
T
(1 k
2
)T = 0 and X
(
2
)X = 0.
For
2
> 0 we obtain
X = A
1
cosh x +A
2
sinhx and T = A
3
e
(1+k
2
)t
so that
u = XT = e
(1+k
2
)t
(c
1
cosh x +c
2
sinhx) .
For
2
< 0 we obtain
X = A
1
cos x +A
2
sinx and T = A
3
e
(1k
2
)t
so that
u = XT = e
(1k
2
)t
(c
3
cos x +c
4
sinx).
If
2
= 0 then
X
= 0 and T
+T = 0,
and we obtain
X = A
1
x +A
2
and T = A
3
e
t
.
In this case
u = XT = e
t
(c
5
x +c
6
)
10. If u = XT then
u
t
= XT
,
u
xx
= X
T,
kX
T = XT
,
and
X
X
=
T
kT
=
2
so that
X
2
X = 0 and T
2
kT = 0.
For
2
> 0 we obtain
X = A
1
cos x +A
2
sinx and T = A
3
e
2
kt
,
so that
u = XT = e
c
2
3
kt
(c
1
cos c
3
x +c
2
sinc
3
x).
For
2
< 0 we obtain
X = A
1
e
x
+A
2
e
x
,
T = A
3
e
2
kt
,
and
u = XT = e
c
2
3
kt
_
c
1
e
c3x
+c
2
e
c3x
_
.
For
2
= 0 we obtain
T = A
3
, X = A
1
x +A
2
, and u = XT = c
1
x +c
2
.
608
Exercises 13.1
11. If u = XT then
u
xx
= X
T,
u
tt
= XT
,
a
2
X
T = XT
,
and
X
X
=
T
a
2
T
=
2
so that
X
2
X = 0 and T
a
2
2
T = 0.
For
2
> 0 we obtain
X = A
1
cos x +A
2
sinx,
T = A
3
cos at +A
4
sinat,
and
u = XT = (A
1
cos x +A
2
sinx)(A
3
cos at +A
4
sinat).
For
2
< 0 we obtain
X = A
1
e
x
+A
2
e
x
,
T = A
3
e
at
+A
4
e
at
,
and
u = XT =
_
A
1
e
x
+A
2
e
x
_ _
A
3
e
at
+A
4
e
at
_
.
For
2
= 0 we obtain
X = A
1
x +A
2
,
T = A
3
t +A
4
,
and
u = XT = (A
1
x +A
2
)(A
3
t +A
4
).
12. If u = XT then
u
t
= XT
,
u
tt
= XT
,
u
xx
= X
T,
a
2
X
T = XT
+ 2kXT
,
and
X
X
=
T
+ 2kT
a
2
T
=
2
so that
X
2
X = 0 and T
+ 2kT
a
2
2
T = 0.
For
2
> 0 we obtain
X = A
1
e
x
+A
2
e
x
,
T = A
3
e
(k+
k
2
+a
2
2
)t
+A
4
e
(k
k
2
+a
2
2
)t
,
and
609
Exercises 13.1
u = XT =
_
A
1
e
x
+A
2
e
x
_
_
A
3
e
(k+
k
2
+a
2
2
)t
+A
4
e
(k
k
2
+a
2
2
)t
_
.
For
2
< 0 we obtain
X = A
1
cos x +A
2
sinx.
If k
2
a
2
2
> 0 then
T = A
3
e
(k+
k
2
a
2
2
)t
+A
4
e
(k
k
2
a
2
2
)t
.
If k
2
a
2
2
< 0 then
T = e
kt
_
A
3
cos
_
a
2
2
k
2
t +A
4
sin
_
a
2
2
k
2
t
_
.
If k
2
a
2
2
= 0 then
T = A
3
e
kt
+A
4
te
kt
so that
u = XT = (A
1
cos x +A
2
sinx)
_
A
3
e
(k+
k
2
a
2
2
)t
+A
4
e
(k
k
2
a
2
2
)t
_
= (A
1
cos x +A
2
sinx)e
kt
_
A
3
cos
_
a
2
2
k
2
t +A
4
sin
_
a
2
2
k
2
t
_
=
_
A
1
cos
k
a
x +A
2
sin
k
a
x
_
_
A
3
e
kt
+A
4
te
kt
_
.
For
2
= 0 we obtain
XA
1
x +A
2
,
T = A
3
+A
4
e
2kt
,
and
u = XT = (A
1
x +A
2
)(A
3
+A
4
e
2kt
).
13. If u = XY then
u
xx
= X
Y,
u
yy
= XY
,
X
Y +xY
= 0,
and
X
X
=
Y
Y
=
2
so that
X
2
X = 0 and Y
2
Y = 0.
For
2
> 0 we obtain
X = A
1
cos x +A
2
sinx,
Y = A
3
e
y
+A
4
e
y
,
and
u = XY = (A
1
cos x +A
2
sinx)(A
3
e
y
+A
4
e
y
).
For
2
< 0 we obtain
X = A
1
e
x
+A
2
e
x
,
Y = A
3
cos y +A
4
siny,
and
610
Exercises 13.1
u = XY = (A
1
e
x
+A
2
e
x
)(A
3
cos y +A
4
siny).
For
2
= 0 we obtain
u = XY = (A
1
x +A
2
)(A
3
y +A
4
).
14. If u = XY then
u
xx
= X
Y,
u
yy
= XY
,
x
2
X
Y +xY
= 0,
and
x
2
X
X
=
Y
Y
=
2
so that
x
2
X
2
X = 0 and Y
2
Y = 0.
For
2
> 0 we obtain
Y = A
3
e
y
+A
4
e
y
.
If 1 4
2
> 0 then
X = A
1
x
(1/2+
14
2
/2)
+A
2
x
(1/2
14
2
/2)
.
If 1 4
2
< 0 then
X = x
1/2
_
A
1
cos
1
2
_
4
2
1 lnx +A
2
sin
1
2
_
4
2
1 lnx
_
.
If 1 4
2
= 0 then
X = A
1
x
1/2
+A
2
x
1/2
lnx
so that
u = XY =
_
A
1
x
(1/2+
14
2
/2)
+A
2
x
(1/2
14
2
/2)
_
_
A
3
e
y
+A
4
e
y
_
= x
1/2
_
A
1
cos
1
2
_
4
2
1 lnx +A
2
sin
1
2
_
4
2
1 lnx
_
_
A
3
e
y/2
+A
4
e
y/2
_
=
_
A
1
x
1/2
+A
2
x
1/2
lnx
__
A
3
e
y/2
+A
4
e
y/2
_
.
For
2
< 0 we obtain
X = A
1
e
(1/2+
1+4
2
/2)x
+A
2
e
(1/2
1+4
2
/2)x
,
Y = A
3
cos y +A
4
siny,
and
u = XY =
_
A
1
e
(1/2+
1+4
2
/2)x
+A
2
e
(1/2
1+4
2
/2)x
_
(A
3
cos y +A
4
siny) .
For
2
= 0 we obtain
X = A
1
x +A
2
,
Y = A
3
y +A
4
,
and
u = XY = (A
1
x +A
2
)(A
3
y +A
4
).
611
Exercises 13.1
15. If u = XY then
u
xx
= X
Y,
u
yy
= XY
,
X
Y +XY
= XY,
and
X
X
=
Y Y
Y
=
2
so that
X
2
X = 0 and Y
+ (
2
1)Y = 0.
For
2
> 0 we obtain
X = A
1
e
x
+A
2
e
x
.
If
2
1 > 0 then
Y = A
3
cos
_
2
1 y +A
4
sin
_
2
1 y.
If
2
1 < 0 then
Y = A
3
e
1
2
y
+A
4
e
1
2
y
.
If
2
1 = 0 then Y = A
3
y +A
4
so that
u = XY =
_
A
1
e
x
+A
2
e
x
_
_
A
3
cos
_
2
1 y +A
4
sin
_
2
1 y
_
,
=
_
A
1
e
x
+A
2
e
x
_
_
A
3
e
1
2
y
+A
4
e
1
2
y
_
= (A
1
e
x
+A
2
e
x
)(A
3
y +A
4
).
For
2
< 0 we obtain
X = A
1
cos x +A
2
sinx,
Y = A
3
e
1+
2
y
+A
4
e
1+
2
y
,
and
u = XY = (A
1
cos x +A
2
sinx)
_
A
3
e
1+
2
y
+A
4
e
1+
2
y
_
.
For
2
= 0 we obtain
X = A
1
x +A
2
,
Y = A
3
e
y
+A
4
e
y
,
and
u = XY = (A
1
x +A
2
)(A
3
e
y
+A
4
e
y
).
16. If u = XT then
u
tt
= XT
, u
xx
= X
T, and a
2
X
T g = XT
,
which is not separable.
17. Identifying A = B = C = 1, we compute B
2
4AC = 3 < 0. The equation is elliptic.
18. Identifying A = 3, B = 5, and C = 1, we compute B
2
4AC = 13 > 0. The equation is hyperbolic.
2
4AC = 0. The equation is parabolic.
2
612
Exercises 13.1
2
2
2
4AC = 0. The equation is parabolic.
2
4AC = 4 < 0. The equation is elliptic.
25. Identifying A = a
2
, B = 0, and C = 1, we compute B
2
4AC = 4a
2
> 0. The equation is hyperbolic.
26. Identifying A = k > 0, B = 0, and C = 0, we compute B
2
4AC = 4k < 0. The equation is elliptic.
27. If u = RT then
u
r
= R
T,
u
rr
= R
T,
u
t
= RT
,
RT
= k
_
R
T +
1
r
R
T
_
,
and
r
2
R
+rR
r
2
R
=
T
kT
=
2
.
If we use
2
< 0 then
r
2
R
+rR
+
2
r
2
R = 0 and T
2
kT = 0
so that
R = A
2
J
0
(r) +A
3
Y
0
(r),
T = A
1
e
k
2
t
,
and
u = RT = e
k
2
t
[c
1
J
0
(r) +c
2
Y
0
(r)]
28. If u = R then
u
r
= R
,
u
rr
= R
,
u
t
= R
,
u
tt
= R
,
R
+
1
r
R
+
1
r
2
R
= 0,
and
r
2
R
+rR
R
=
=
2
.
Then
r
2
R
+rR
2
R = 0 and
+
2
= 0
so that
R = c
3
r
+c
4
r
= c
1
cos +c
2
sin,
and
u = R = (c
1
cos +c
2
sin)(c
3
r
+c
4
r
).
613
Exercises 13.1
29. For u = A
1
e
2
y
cosh2x +B
1
e
2
y
sinh2x we compute
2
u
x
2
= 4
2
A
1
e
2
y
cosh2x + 4
2
B
1
e
2
y
sinh2x
and
u
y
=
2
A
1
e
2
y
cosh2x +
2
B
1
e
2
y
sinh2x.
Then
2
u/x
2
= 4u/y.
For u = A
2
e
2
y
cos 2x +B
2
e
2
y
sin2x we compute
2
u
x
2
= 4
2
A
2
e
2
y
cos 2x 4
2
B
2
e
2
y
sin2x
and
u
y
=
2
A
2
cos 2x
2
B
2
sin2x.
Then
2
u/x
2
= 4u/y.
For u = A
3
x +B
3
we compute
2
u/x
2
= u/y = 0. Then
2
u/x
2
= 4u/y.
30. We identify A = xy + 1, B = x + 2y, and C = 1. Then B
2
4AC = x
2
+ 4y
2
4. The equation x
2
+ 4y
2
= 4
denes an ellipse. The partial dierential equation is hyperbolic outside the ellipse, parabolic on the ellipse,
and elliptic inside the ellipse.
Exercises 13.2
1. k
2
u
x
2
=
u
t
, 0 < x < L, t > 0
u(0, t) = 0,
u
x
x=L
= 0, t > 0
u(x, 0) = f(x), 0 < x < L
2. k
2
u
x
2
=
u
t
, 0 < x < L, t > 0
u(0, t) = u
0
, u(L, t) = u
1
, t > 0
u(x, 0) = 0, 0 < x < L
3. k
2
u
x
2
=
u
t
, 0 < x < L, t > 0
u(0, t) = 100,
u
x
x=L
= hu(L, t), t > 0
u(x, 0) = f(x), 0 < x < L
4. k
2
u
x
2
+h(u 50) =
u
t
, 0 < x < L, t > 0
u
x
x=0
= 0,
u
x
x=L
t > 0
u(x, 0) = 100, 0 < x < L
5. a
2
2
u
x
2
=

2
u
t
2
, 0 < x < L, t > 0
u(0, t) = 0, u(L, t) = 0, t > 0
u(x, 0) = x(L x),
u
x
t=0
= 0, 0 < x < L
614
Exercises 13.3
6. a
2
2
u
x
2
=

2
u
t
2
, 0 < x < L, t > 0
u(0, t) = 0, u(L, t) = 0, t > 0
u(x, 0) = 0,
u
x
t=0
= sin
x
L
, 0 < x < L
7. a
2
2
u
x
2
2
u
t
=

2
u
t
2
, 0 < x < L, t > 0
u(0, t) = 0, u(L, t) = sint, t > 0
u(x, 0) = f(x),
u
t
t=0
= 0, 0 < x < L
8. a
2
2
u
x
2
+Ax =

2
u
t
2
, 0 < x < L, t > 0, A a constant
u(0, t) = 0, u(L, t) = 0, t > 0
u(x, 0) = 0,
u
x
t=0
= 0, 0 < x < L
9.

2
u
x
2
+

2
u
y
2
= 0, 0 < x < 4, 0 < y < 2
u
x
x=0
= 0, u(4, y) = f(y), 0 < y < 2
u
y
y=0
= 0, u(x, 2) = 0, 0 < x < 4
10.

2
u
x
2
+

2
u
y
2
= 0, 0 < x < , y > 0
u(0, y) = e
y
, u(, y) =
_
100, 0 < y 1
0, y > 1
u(x, 0) = f(x), 0 < x <
Exercises 13.3
1. Using u = XT and
2
as a separation constant leads to
X
+
2
X = 0,
X(0) = 0,
X(L) = 0,
and
T
+k
2
T = 0.
Then
X = c
1
sin
n
L
x and T = c
2
e
kn
2
2
L
2
t
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
e
kn
2
2
L
2
t
sin
n
L
x.
615
Exercises 13.3
Imposing
u(x, 0) =
n=1
A
n
sin
n
L
x
gives
A
n
=
2
L
_
L/2
0
sin
n
L
xdx =
2
n
_
1 cos
n
2
_
for n = 1, 2, 3, . . . so that
u(x, t) =
2
n=1
1 cos
n
2
n
e
kn
2
2
L
2
t
sin
n
L
x.
2. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X(L) = 0,
and
T
+k
2
T = 0.
Then
X = c
1
sin
n
L
x and T = c
2
e
kn
2
2
L
2
t
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
e
kn
2
2
L
2
t
sin
n
L
x.
Imposing
u(x, 0) =
n=1
A
n
sin
n
L
x
gives
A
n
=
2
L
_
L
0
x(L x) sin
n
L
xdx =
4L
2
n
3
3
[1 (1)
n
]
for n = 1, 2, 3, . . . so that
u(x, t) =
4L
2
n=1
1 (1)
n
n
3
e
kn
2
2
L
2
t
sin
n
L
x.
3. Using u = XT and
2
X
+
2
X = 0,
X
(0) = 0,
X
(L) = 0,
and
T
+k
2
T = 0.
Then
X = c
1
cos
n
L
x and T = c
2
e
kn
2
2
L
2
t
for n = 0, 1, 2, . . . so that
u =
n=0
A
n
e
kn
2
2
L
2
t
cos
n
L
x.
616
Exercises 13.3
Imposing
u(x, 0) = f(x) =
n=0
A
n
cos
n
L
x
gives
u(x, t) =
1
L
_
L
0
f(x) dx +
2
L
n=1
_
_
L
0
f(x) cos
n
L
xdx
_
e
kn
2
2
L
2
t
cos
n
L
x.
4. If L = 2 and f(x) is x for 0 < x < 1 and f(x) is 0 for 1 < x < 2 then
u(x, t) =
1
4
+ 4
n=1
_
1
2n
sin
n
2
+
1
n
2
2
_
cos
n
2
1
_
_
e
kn
2
2
4
t
cos
n
2
x.
5. Using u = XT and
2
X
+
2
X = 0,
X
(0) = 0,
X
(L) = 0,
and
T
+ (h +k
2
)T = 0.
Then
X = c
1
cos
n
L
x and T = c
2
e
_
h+
kn
2
2
L
2
_
t
for n = 0, 1, 2, . . . so that
u =
n=0
A
n
e
_
h+
kn
2
2
L
2
_
t
cos
n
L
x.
Imposing
u(x, 0) = f(x) =
n=0
A
n
cos
n
L
x
gives
u(x, t) =
e
ht
L
_
L
0
f(x) dx +
2
L
n=1
_
_
L
0
f(x) cos
n
L
xdx
_
e
_
h+
kn
2
2
L
2
_
t
cos
n
L
x.
6. In Problem 5 we instead nd that X(0) = 0 and X(L) = 0 so that
X = c
1
sin
n
L
x
and
u =
2
L
n=1
_
_
L
0
f(x) sin
n
L
xdx
_
e
_
h+
kn
2
2
L
2
_
t
sin
n
L
x.
617
1 2 3 4 5 6
t
20
40
60
80
100
u
x=3p 4
x=5p 6
x=11p 12
x=p
0
20
40
60
80
100
x
0
50
100
150
200
t
0
10
20
30
40
uHx,tL
0
20
40
60
Exercises 13.3
7.
8. (a) The solution is
u(x, t) =
n=1
A
n
e
k(n
2
2
/100
2
)t
sin
n
100
x,
where
A
n
=
2
100
__
50
0
0.8xsin
n
100
xdx +
_
100
50
0.8(100 x) sin
n
100
xdx
_
=
320
n
2
2
sin
n
x
.
Thus,
u(x, t) =
320
n=1
1
n
2
_
sin
n
2
_
e
k(n
2
2
/100
2
)t
sin
n
100
x.
(b) Since A
n
= 0 for n even, the rst ve nonzero terms correspond to n = 1, 3, 5, 7, 9. In this case
sin(n/2) = sin(2p 1)/2 = (1)
p+1
for p = 1, 2, 3, 4, 5, and
u(x, t) =
320
p=1
(1)
p+1
(2p 1)
2
e
(1.6352(2p1)
2
2
/100
2
)t
sin
(2p 1)
100
x.
Exercises 13.4
1. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X(L) = 0,
and
T
+
2
a
2
T = 0.
618
Exercises 13.4
Then
X = c
1
sin
n
L
x and T = c
2
cos
na
L
t +c
3
sin
na
L
t
for n = 1, 2, 3, . . . so that
u =
n=1
_
A
n
cos
na
L
t +B
n
sin
na
L
t
_
sin
n
L
x.
Imposing
u(x, 0) =
1
4
x(L x) =
n=1
A
n
sin
n
L
x
and
u
t
(x, 0) = 0 =
n=1
B
n
na
L
sin
n
L
x
gives
A
n
=
L
n
3
3
[1 (1)
n
] and B
n
= 0
for n = 1, 2, 3, . . . so that
u(x, t) =
L
2
n=1
1 (1)
n
n
3
cos
na
L
t sin
n
L
x.
2. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X(L) = 0,
and
T
+
2
a
2
T = 0.
Then
X = c
1
sin
n
L
x and T = c
2
cos
na
L
t +c
3
sin
na
L
t
for n = 1, 2, 3, . . . so that
u =
n=1
_
A
n
cos
na
L
t +B
n
sin
na
L
t
_
sin
n
L
x.
Imposing
u(x, 0) = 0 =
n=1
A
n
sin
n
L
x
and
u
t
(x, 0) = x(L x) =
n=1
B
n
na
L
sin
n
L
x
gives
A
n
= 0 and B
n
na
L
=
4L
2
n
3
3
[1 (1)
n
]
for n = 1, 2, 3, . . . so that
u(x, t) =
4L
3
a
4
n=1
1 (1)
n
n
4
sin
na
L
t sin
n
L
x.
619
Exercises 13.4
3. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X(L) = 0,
and
T
+
2
a
2
T = 0.
Then
X = c
1
sin
n
L
x and T = c
2
cos
na
L
t +c
3
sin
na
L
t
for n = 1, 2, 3, . . . so that
u =
n=1
_
A
n
cos
na
L
t +B
n
sin
na
L
t
_
sin
n
L
x.
Imposing
u(x, 0) =
n=1
A
n
sin
n
L
x
gives
A
n
=
2
L
_
_
L/3
0
3
L
xsin
n
L
xdx +
_
2L/3
L/3
sin
n
L
xdx +
_
L
2L/3
_
3
3
L
x
_
sin
n
L
xdx
_
so that
A
1
=
6
2
,
A
2
= A
3
= A
4
= 0,
A
5
=
6
3
5
2
2
,
A
6
= 0,
a
7
=
6
3
7
2
2
. . . .
Imposing
u
t
(x, 0) = 0 =
n=1
B
n
na
L
sin
n
L
x
gives B
n
= 0 for n = 1, 2, 3, . . . so that
u(x, t) =
6
2
_
cos
a
L
t sin

L
x
1
5
2
cos
5a
L
t sin
5
L
x +
1
7
2
cos
7a
L
t sin
7
L
x
_
.
4. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X() = 0,
and
T
+
2
a
2
T = 0.
Then
X = c
1
sinnx and T = c
2
cos nat +c
3
sinnat
620
Exercises 13.4
for n = 1, 2, 3, . . . so that
u =
n=1
(A
n
cos nat +B
n
sinnat) sinnx.
Imposing
u(x, 0) =
1
6
x(
2
x
2
) =
n=1
A
n
sinnx and u
t
(x, 0) = 0
gives
B
n
= 0 and A
n
=
2
n
3
(1)
n+1
for n = 1, 2, 3, . . . so that
u(x, t) = 2
n=1
(1)
n+1
n
3
cos nat sinnx.
5. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X() = 0,
and
T
+
2
a
2
T = 0.
Then
X = c
1
sinnx and T = c
2
cos nat +c
3
sinnat
for n = 1, 2, 3, . . . so that
u =
n=1
(A
n
cos nat +B
n
sinnat) sinnx.
Imposing
u(x, 0) = 0 =
n=1
A
n
sinnx and u
t
(x, 0) = sinx =
n=1
B
n
na sinnx
gives
A
n
= 0, B
1
=
1
a
2
, and B
n
= 0
for n = 2, 3, 4, . . . so that
u(x, t) =
1
a
sinat sinx.
6. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X(1) = 0,
and
T
+
2
a
2
T = 0.
Then
X = c
1
sinnx and T = c
2
cos nat +c
3
sinnat
621
Exercises 13.4
for n = 1, 2, 3, . . . so that
u =
n=1
(A
n
cos nat +B
n
sinnat) sinnx.
Imposing
u(x, 0) = 0.01 sin3x =
n=1
A
n
sinnx
and
u
t
(x, 0) = 0 =
n=1
B
n
na sinnx
gives B
n
= 0 for n = 1, 2, 3, . . ., A
3
= 0.01, and A
n
= 0 for n = 1, 2, 4, 5, 6, . . . so that
u(x, t) = 0.01 sin3x cos 3at.
7. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X(L) = 0,
and
T
+
2
a
2
T = 0.
Then
X = c
1
sin
n
L
x and T = c
2
cos
na
L
t +c
3
sin
na
L
t
for n = 1, 2, 3, . . . so that
u =
n=1
_
A
n
cos
na
L
t +B
n
sin
na
L
t
_
sin
n
L
x.
Imposing
u(x, 0) =
n=1
A
n
sin
n
L
x
and
u
t
(x, 0) = 0 =
n=1
B
n
na
L
sin
n
L
x
gives
B
n
= 0 and A
n
=
8h
n
2
2
sin
n
2
for n = 1, 2, 3, . . . so that
u(x, t) =
8h
n=1
1
n
2
sin
n
2
cos
na
L
t sin
n
L
x.
8. Using u = XT and
2
X
+
2
X = 0,
X
(0) = 0,
X
(L) = 0,
and
T
+a
2
2
T = 0.
622
Exercises 13.4
Then
X = c
1
cos
n
L
x and T = c
2
cos
na
L
t
for n = 1, 2, 3, . . . . Since = 0 is also an eigenvalue with eigenfunction X(x) = 1 we have
u = A
0
+
n=1
A
n
cos
na
L
t cos
n
L
x.
Imposing
u(x, 0) = x =
n=0
A
n
cos
n
L
x
gives
A
0
=
1
L
_
L
0
xdx =
L
2
and
A
n
=
2
L
_
L
0
xcos
n
L
xdx =
2L
n
2
2
[(1)
n
1]
for n = 1, 2, 3, . . . , so that
u(x, t) =
L
2
+
2L
n=1
(1)
n
1
n
2
cos
na
L
t cos
n
L
x.
9. Using u = XT and
2
X
+
2
X = 0,
X(0) = 0,
X() = 0,
and
T
+ 2T
+
2
T = 0.
Then
X = c
1
sinnx and T = e
t
_
c
2
cos
_
n
2
2
t +c
3
sin
_
n
2
2
t
_
so that
u =
n=1
e
t
_
A
n
cos
_
n
2
2
t +B
n
sin
_
n
2
2
t
_
sinnx.
Imposing
u(x, 0) = f(x) =
n=1
A
n
sinnx
and
u
t
(x, 0) = 0 =
n=1
_
B
n
_
n
2
2
A
n
_
sinnx
gives
u(x, t) = e
t
n=1
A
n
_
cos
_
n
2
2
t +

_
n
2
2
sin
_
n
2
2
t
_
sinnx,
where
A
n
=
2
_

0
f(x) sinnxdx.
10. Using u = XT and
2
as a separation constant leads to X
+
2
X = 0, X(0) = 0, X() = 0 and
623
Exercises 13.4
T
+ (1 +
2
)T = 0, T
(0) = 0. Then X = c
2
sinnx and T = c
3
cos
n
2
+ 1 t for n = 1, 2, 3, . . . so that
u =
n1
B
n
cos
_
n
2
+ 1 t sinnx.
Imposing u(x, 0) =
n=1
B
n
sinnx gives
B
n
=
2
_
/2
0
xsinnxdx +
2
_

/2
( x) sinnxdx =
4
n
2
sin
n
2
=
_
0, n even
4
n
2
(1)
(n+3)/2
, n = 2k 1, k = 1, 2, 3, . . ..
Thus with n = 2k 1,
u(x, t) =
4
n=1
sin
n
2
n
2
cos
_
n
2
+ 1 t sinnx =
4
k=1
(1)
k+1
(2k 1)
2
cos
_
(2k 1)
2
+ 1 t sin(2k 1)x.
11. Separating variables in the partial dierential equation gives
X
(4)
X
=
T
a
2
T
=
4
so that
X
(4)
4
X = 0
T
+a
2
4
T = 0
and
X = c
1
cosh x +c
2
sinhx +c
3
cos x +c
4
sinx
T = c
5
cos a
2
t +c
6
sina
2
t.
The boundary conditions translate into X(0) = X(L) = 0 and X
(0) = X
(L) = 0. From X(0) = X
(0) = 0
we nd c
1
= c
3
= 0. From
X(L) = c
2
sinhL +c
4
sinL = 0
X
(L) =
2
c
2
sinhL
2
c
4
sinL = 0
we see by subtraction that c
4
sinL = 0. This equation yields the eigenvalues = nL for n = 1, 2, 3, . . . .
The corresponding eigenfunctions are
X = c
4
sin
n
L
x.
Thus
u(x, t) =
n=1
_
A
n
cos
n
2
2
L
2
at +B
n
sin
n
2
2
L
2
at
_
sin
n
L
x.
From
u(x, 0) = f(x) =
n=1
A
n
sin
n
L
x
we obtain
A
n
=
2
L
_
L
0
f(x) sin
n
L
xdx.
From
u
t
=
n=1
_
A
n
n
2
2
a
L
2
sin
n
2
2
L
2
at +B
n
n
2
2
a
L
2
cos
n
2
2
L
2
at
_
sin
n
L
x
624
Exercises 13.4
and
u
t
t=0
= g(x) =
n=1
B
n
n
2
2
a
L
2
sin
n
L
x
we obtain
B
n
n
2
2
a
L
2
=
2
L
_
L
0
g(x) sin
n
L
xdx
and
B
n
=
2L
n
2
2
a
_
L
0
g(x) sin
n
L
xdx.
12. (a) Using X = c
1
cosh x +c
2
sinhx +c
3
cos x +c
4
sinx
and X(0) = 0, X
(0) = 0 we nd, in turn, c

3
= c
1
and c
4
= c
2
. The conditions X(L) = 0 and X
(L) = 0
then yield the system of equations for c
1
and c
2
:
c
1
(cosh L cos L) +c
2
(sinhL sinL) = 0
c
1
(sinhL +sinL) +c
2
(cosh L cos L) = 0.
In order that this system have nontrivial solutions the determinant of the coecients must be zero:
(cosh L cos L)
2
(sinh
2
L sin
2
L) = 0.
= 0 is not an eigenvalue since this leads to X = 0. Thus the last equation simplies to coshLcos L = 1
or cosh xcos x = 1, where x = L.
(b) The equation coshxcos x = 1 is the same as cos x = sech x. The gure
indicates that the equation has an innite number of roots.
(c) Using a CAS we nd the rst four positive roots to be x
1
= 4.7300, x
2
= 7.8532, x
3
= 10.9956, and
x
4
= 14.1372. Thus the rst four eigenvalues are
1
= x
1
/L = 4.7300/L,
2
= x
2
/L = 7.8532/L,
3
= x
3
/L = 10.9956/L, and
4
= 14.1372/L.
13. From (5) in the text we have
u(x, t) =
n=1
_
A
n
cos
na
L
t +B
n
sin
na
L
t
_
sin
n
L
x.
Since u
t
(x, 0) = g(x) = 0 we have B
n
= 0 and
u(x, t) =
n=1
A
n
cos
na
L
t sin
n
L
x
=
n=1
A
n
1
2
_
sin
_
n
L
x +
na
L
t
_
+ sin
_
n
L
x
na
L
t
__
=
1
2
n=1
A
n
_
sin
n
L
(x +at) + sin
n
L
(x at)
_
.
From
u(x, 0) = f(x) =
n=1
A
n
sin
n
L
x
625
Exercises 13.4
we identify
f(x +at) =
n=1
A
n
sin
n
L
(x +at)
and
f(x at) =
n=1
A
n
sin
n
L
(x at),
so that
u(x, t) =
1
2
[f(x +at) +f(x at)].
14. (a) We note that
x
=
x
= 1,
t
= a, and
t
= a. Then
u
x
=
u
x
+
u
x
= u
+u
and
2
u
x
2
=

x
(u
+u
) =
u
x
+
u
x
+
u
x
+
u
x
= u
+ 2u
+u
.
Similarly
2
u
t
2
= a
2
(u
2u
+u
).
Thus
a
2

2
u
x
2
=

2
u
t
2
becomes

2
u
= 0.
(b) Integrating
2
u
= 0
we obtain
_

d =
_
0 d
u
= f().
Integrating this result with respect to we obtain
_
u
d =
_
f() d
u = F() +G().
Since = x +at and = x at, we then have
u = F() +G() = F(x +at) +G(x at).
Next, we have
u(x, t) = F(x +at) +G(x at)
u(x, 0) = F(x) +G(x) = f(x)
u
t
(x, 0) = aF
(x) aG
(x) = g(x)
Integrating the last equation with respect to x gives
F(x) G(x) =
1
a
_
x
x0
g(s) ds +c
1
.
626
Exercises 13.4
Substituting G(x) = f(x) F(x) we obtain
F(x) =
1
2
f(x) +
1
2a
_
x
x0
g(s) ds +C
where c = c
1
/2. Thus
G(x) =
1
2
f(x)
1
2a
_
x
x0
g(s) ds c.
(c) From the expressions for F and G,
F(x +at) =
1
2
f(x +at) +
1
2a
_
x+at
x0
g(s) ds +c
G(x at) =
1
2
f(x at)
1
2a
_
xat
x0
g(s) ds c.
Thus,
u(x, t) = F(x +at) +G(x at) =
1
2
[f(x +at) +f(x at)] +
1
2a
_
x+at
xat
g(s) ds.
Here we have used
_
xat
x0
g(s) ds =
_
x0
xat
g(s) ds.
15. u(x, t) =
1
2
[sin(x +at) + sin(x at)] +
1
2a
_
x+at
xat
ds
=
1
2
[sinxcos at + cos xsinat + sinxcos at cos xsinat] +
1
2a
s
x+at
xat
= sinxcos at +t
16. u(x, t) =
1
2
sin(x +at) + sin(x at)] +
1
2a
_
x+at
xat
cos s ds
= sinxcos at +
1
2a
[sin(x +at) sin(x at)] = sinxcos at +
1
a
cos xsinat
17. u(x, t) = 0 +
1
2a
_
x+at
xat
sin2s ds =
1
2a
_
cos(2x + 2at) + cos(2x 2at)
2
_
=
1
4a
[cos 2xcos 2at + sin2xsin2at + cos 2xcos 2at + sin2xsin2at] =
1
2a
sin2xsin2at
18.
19. (a)
(b) Since g(x) = 0, dAlemberts solution with a = 1 is
u(x, t) =
1
2
[f(x +t) +f(x t)].
627
-6 -4 -2 0 2 4 6
1
t= 0
-6 -4 -2 0 2 4 6
1
t= 0.2
-6 -4 -2 0 2 4 6
1
t= 0.4
-6 -4 -2 0 2 4 6
1
t= 0.6
-6 -4 -2 0 2 4 6
1
t= 0.8
-6 -4 -2 0 2 4 6
1
t= 1.
-6 -4 -2 0 2 4 6
1
t= 2.
-6 -4 -2 0 2 4 6
1
t= 3.
-6 -4 -2 0 2 4 6
1
t= 4.
-6 -4 -2 0 2 4 6
1
t= 5.
-6 -4 -2 0 2 4 6
1
t= 0
-6 -4 -2 0 2 4 6
1
t= 0.2
-6 -4 -2 0 2 4 6
1
t= 0.4
-6 -4 -2 0 2 4 6
1
t= 0.6
-6 -4 -2 0 2 4 6
1
t= 0.8
-6 -4 -2 0 2 4 6
1
t= 1.
Exercises 13.4
Sample plots are shown below.
(c) The single peaked wave disolves into two peaks moving outward.
20. (a) With a = 1, dAlemberts solution is
u(x, t) =
1
2
_
x+t
xt
g(s) ds where g(s) =
_
1, |s| 0.1
0, |s| > 0.1
.
Sample plots are shown below.
628
-6 -4 -2 0 2 4 6
1
t= 2.
-6 -4 -2 0 2 4 6
1
t= 3.
-6 -4 -2 0 2 4 6
1
t= 4.
-6 -4 -2 0 2 4 6
1
t= 5.
0.5 1 1.5 2 2.5 3
x
-1
-0.5
0.5
1
u
0.5 1 1.5 2 2.5 3
x
-1
-0.5
0.5
1
u
0.5 1 1.5 2 2.5 3
x
-1
-0.5
0.5
1
u
0.5 1 1.5 2 2.5 3
x
-1
-0.5
0.5
1
u
Exercises 13.5
(b) Some frames of the movie are shown in part (a), The string has a roughly rectangular shape with the base
on the x-axis increasing in length.
21. (a) and (b) With the given parameters, the solution is
u(x, t) =
8
n=1
1
n
2
sin
n
2
cos nt sinnx.
For n even, sin(n/2) = 0, so the rst six nonzero terms correspond to n = 1, 3, 5, 7, 9, 11. In this case
sin(n/2) = sin(2p 1)/2 = (1)
p+1
for p = 1, 2, 3, 4, 5, 6, and
u(x, t) =
8
p=1
(1)
p+1
(2p 1)
2
cos(2p 1)t sin(2p 1)x.
Frames of the movie corresponding to t = 0.5, 1, 1.5, and 2 are shown.
Exercises 13.5
1. Using u = XY and
2
X
+
2
X = 0,
X(0) = 0,
X(a) = 0,
and
Y
2
Y = 0,
Y (0) = 0.
Then
X = c
1
sin
n
a
x and Y = c
2
sinh
n
a
y
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
sin
n
a
x sinh
n
a
y.
629
Exercises 13.5
Imposing
u(x, b) = f(x) =
n=1
A
n
sinh
nb
a
sin
n
a
x
gives
A
n
sinh
nb
a
=
2
a
_
a
0
f(x) sin
n
a
xdx
so that
u(x, y) =
n=1
A
n
sin
n
a
x sinh
n
a
y
where
A
n
=
2
a
csch
nb
a
_
a
0
f(x) sin
n
a
xdx.
2. Using u = XY and
2
X
+
2
X = 0,
X(0) = 0,
X(a) = 0,
and
Y
2
Y = 0,
Y
(0) = 0.
Then
X = c
1
sin
n
a
x and Y = c
2
cosh
n
a
y
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
sin
n
a
x cosh
n
a
y.
Imposing
u(x, b) = f(x) =
n=1
A
n
cosh
nb
a
sin
n
a
x
gives
A
n
cosh
nb
a
=
2
a
_
a
0
f(x) sin
n
a
xdx
so that
u(x, y) =
n=1
A
n
sin
n
a
x cosh
n
a
y
where
A
n
=
2
a
sech
nb
a
_
a
0
f(x) sin
n
a
xdx.
3. Using u = XY and
2
X
+
2
X = 0,
X(0) = 0,
X(a) = 0,
and
630
Exercises 13.5
Y
2
Y = 0,
Y (b) = 0.
Then
X = c
1
sin
n
a
x and Y = c
2
cosh
n
a
y c
2
cosh
nb
a
sinh
nb
a
sinh
n
a
y
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
sin
n
a
x
_
cosh
n
a
y
cosh
nb
a
sinh
nb
a
sinh
n
a
y
_
.
Imposing
u(x, 0) = f(x) =
n=1
A
n
sin
n
a
x
gives
A
n
=
2
a
_
a
0
f(x) sin
n
a
xdx
so that
u(x, y) =
2
a
n=1
__
a
0
f(x) sin
n
a
xdx
_
sin
n
a
x
_
cosh
n
a
y
cosh
nb
a
sinh
nb
a
sinh
n
a
y
_
.
4. Using u = XY and
2
X
+
2
X = 0,
X
(0) = 0,
X
(a) = 0,
and
Y
2
Y = 0,
Y (b) = 0.
Then
X = c
1
cos
n
a
x
for n = 0, 1, 2, . . . and
Y = c
2
(y b) or Y = c
2
cosh
n
a
y c
2
cosh
nb
a
sinh
nb
a
sinh
n
a
y
so that
u = A
0
(y b) +
n=1
A
n
cos
n
a
x
_
cosh
n
a
y
cosh
nb
a
sinh
nb
a
sinh
n
a
y
_
.
Imposing
u(x, 0) = x = A
0
b +
n=1
A
n
cos
n
a
x
gives
A
0
b =
1
a
_
a
0
xdx =
1
2
a
and
A
n
=
2
a
_
a
0
xcos
n
a
xdx =
2a
n
2
2
[(1)
n
1]
631
Exercises 13.5
so that
u(x, y) =
a
2b
(b y) +
2a
n=1
(1)
n
1
n
2
cos
n
a
x
_
cosh
n
a
y
cosh
nb
a
sinh
nb
a
sinh
n
a
y
_
.
5. Using u = XY and
2
X
+
2
X = 0,
X(0) = 0,
and
Y
+
2
Y = 0,
Y
(0) = 0,
Y
(1) = 0.
Then
Y = c
1
cos ny
for n = 0, 1, 2, . . . and
X = c
2
x or X = c
2
sinhnx
for n = 1, 2, 3, . . . so that
u = A
0
x +
n=1
A
n
sinhnx cos ny.
Imposing
u(1, y) = 1 y = A
0
+
n=1
A
n
sinhnx cos ny
gives
A
0
=
_
1
0
(1 y) dy
and
A
n
sinhn = 2
_
1
0
(1 y) cos ny =
2[1 (1)
n
]
n
2
2
sinhn
for n = 1, 2, 3, . . . so that
u(x, y) =
1
2
x +
2
n=1
1 (1)
n
n
2
sinhn
sinhnx cos ny.
6. Using u = XY and
2
X
2
X = 0,
X
(1) = 0,
and
Y
+
2
Y = 0,
Y
(0) = 0,
Y
() = 0.
Then
Y = c
1
cos ny
632
Exercises 13.5
for n = 0, 1, 2, . . . and
X = c
2
cosh nx c
2
sinhn
cosh n
sinhnx
for n = 0, 1, 2, . . . so that
u = A
0
+
n=1
A
n
_
cosh nx
sinhn
cosh n
sinhnx
_
cos ny.
Imposing
u(0, y) = g(y) = A
0
+
n=1
A
n
cos ny
gives
A
0
=
1
_

0
g(y) dy and A
n
=
2
_

0
g(y) cos ny dy
for n = 1, 2, 3, . . . so that
u(x, y) =
1
_

0
g(y) dy +
n=1
_
2
_

0
g(y) cos ny dy
__
cosh nx
sinhn
cosh n
sinhnx
_
cos ny.
7. Using u = XY and
2
X
2
X = 0,
X
(0) = X(0),
and
Y
+
2
Y = 0,
Y (0) = 0,
Y () = 0.
Then
Y = c
1
sinny and X = c
2
(ncosh nx + sinhnx)
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
(ncosh nx + sinhnx) sinny.
Imposing
u(, y) = 1 =
n=1
A
n
(ncosh n + sinhn) sinny
gives
A
n
(ncosh n + sinhn) =
2
_

0
sinny dy =
2[1 (1)
n
]
n
for n = 1, 2, 3, . . . so that
u(x, y) =
2
n=1
1 (1)
n
n
ncosh nx + sinhnx
ncosh n + sinhn
sinny.
8. Using u = XY and
2
X
+
2
X = 0,
X(0) = 0,
X(1) = 0
and
633
Exercises 13.5
Y
2
Y = 0,
Y
(0) = Y (0).
Then
X = c
1
sinnx and Y = c
2
(ncosh ny + sinhny)
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
(ncosh ny + sinhny) sinnx.
Imposing
u(x, 1) = f(x) =
n=1
A
n
(ncosh n + sinhn) sinnx
gives
A
n
(ncosh n + sinhn) =
2
_

0
f(x) sinnxdx
for n = 1, 2, 3, . . . so that
u(x, y) =
n=1
A
n
(ncosh ny + sinhny) sinnx
where
A
n
=
2
n cosh n + sinhn
_
1
0
f(x) sinnxdx.
9. This boundary-value problem has the form of Problem 1 in this section, with a = b = 1, f(x) = 100, and
g(x) = 200. The solution, then, is
u(x, y) =
n=1
(A
n
cosh ny +B
n
sinhny) sinnx,
where
A
n
= 2
_
1
0
100 sinnxdx = 200
_
1 (1)
n
n
_
and
B
n
=
1
sinhn
_
2
_
1
0
200 sinnxdx A
n
cosh n
_
=
1
sinhn
_
400
_
1 (1)
n
n
_
200
_
1 (1)
n
n
_
cosh n
_
= 200
_
1 (1)
n
n
_
[2 csch n cothn].
10. This boundary-value problem has the form of Problem 2 in this section, with a = 1 and b = 1. Thus, the
solution has the form
u(x, y) =
n=1
(A
n
cosh nx +B
n
sinhnx) sinny.
The boundary condition u(0, y) = 10y implies
10y =
n=1
A
n
sinny
634
Exercises 13.5
and
A
n
=
2
1
_
1
0
10y sinny dy =
20
n
(1)
n+1
.
The boundary condition u
x
(1, y) = 1 implies
1 =
n=1
(nA
n
sinhn +nB
n
cosh n) sinny
and
nA
n
sinhn +nB
n
cosh n =
2
1
_
1
0
(sinny)dy
A
n
sinhn +B
n
cos n =
2
n
_
1 (1)
n
B
n
=
2
n
_
(1)
n
1
sech n
20
n
(1)
n+1
tanhn.
11. Using u = XY and
2
X
+
2
X = 0,
X(0) = 0,
X() = 0,
and
Y
2
Y = 0.
Then the boundedness of u as y gives Y = c
1
e
ny
and X = c
2
sinnx for n = 1, 2, 3, . . . so that
u =
n=1
A
n
e
ny
sinnx.
Imposing
u(x, 0) = f(x) =
n=1
A
n
sinnx
gives
A
n
=
2
_

0
f(x) sinnxdx
so that
u(x, y) =
n=1
_
2
_

0
f(x) sinnxdx
_
e
ny
sinnx.
12. Using u = XY and
2
X
+
2
X = 0,
X
(0) = 0,
X
() = 0,
and
Y
2
Y = 0.
By the boundedness of u as y we obtain Y = c
1
e
ny
for n = 1, 2, 3, . . . or Y = c
1
and X = c
2
cos nx for
n = 0, 1, 2, . . . so that
u = A
0
+
n=1
A
n
e
ny
cos nx.
635
Exercises 13.5
Imposing
u(x, 0) = f(x) = A
0
+
n=1
A
n
cos nx
gives
A
0
=
1
_

0
f(x) dx and A
n
=
2
_

0
f(x) cos nxdx
so that
u(x, y) =
1
_

0
f(x) dx +
n=1
_
2
_

0
f(x) cos nxdx
_
e
ny
cos nx.
13. Since the boundary conditions at y = 0 and y = b are functions of x we choose to separate Laplaces equation
as
X
X
=
Y
Y
=
2
so that
X
+
2
X = 0
Y
2
Y = 0
and
X(x) = c
1
cos x +c
2
sinx
Y (y) = c
3
cosh y +c
2
sinhy.
Now X(0) = 0 gives c
1
= 0 and X(a) = 0 implies sina = 0 or = n/a for n = 1, 2, 3, . . . . Thus
u
n
(x, y) = XY =
_
A
n
cosh
n
a
y +B
n
sinh
n
a
y
_
sin
n
a
x
and
u(x, y) =
n=1
_
A
n
cosh
n
a
y +B
n
sinh
n
a
y
_
sin
n
a
x. (1)
At y = 0 we then have
f(x) =
n=1
A
n
sin
n
a
x
and consequently
A
n
=
2
a
_
a
0
f(x) sin
n
a
xdx. (2)
At y = b,
g(y) =
n=1
_
A
n
cosh
n
a
b +B
n
sinh
n
b
a
_
sin
n
a
x
indicates that the entire expression in the parentheses is given by
A
n
cosh
n
a
b +B
n
sinh
n
a
b =
2
a
_
a
0
g(x) sin
n
a
xdx.
We can now solve for B
n
:
B
n
sinh
n
a
b =
2
a
_
a
0
g(x) sin
n
a
xdx A
n
cosh
n
a
b
B
n
=
1
sinh
n
a
b
_
2
a
_
a
0
g(x) sin
n
a
xdx A
n
cosh
n
a
b
_
. (3)
A solution to the given boundary-value problem consists of the series (1) with coecients A
n
and B
n
given in
(2) and (3), respectively.
636
Exercises 13.5
14. Since the boundary conditions at x = 0 and x = a are functions of y we choose to separate Laplaces equation
as
X
X
=
Y
Y
=
2
so that
X
2
X = 0
Y
+
2
Y = 0
and
X(x) = c
1
cosh x +c
2
sinhx
Y (y) = c
3
cos y +c
2
siny.
Now Y (0) = 0 gives c
3
= 0 and Y (b) = 0 implies sinb = 0 or = n/b for n = 1, 2, 3, . . . . Thus
u
n
(x, y) = XY =
_
A
n
cosh
n
b
x +B
n
sinh
n
b
x
_
sin
n
b
y
and
u(x, y) =
n=1
_
A
n
cosh
n
b
x +B
n
sinh
n
b
x
_
sin
n
b
y. (4)
At x = 0 we then have
F(y) =
n=1
A
n
sin
n
b
y
and consequently
A
n
=
2
b
_
b
0
F(y) sin
n
b
y dy. (5)
At x = a,
G(y) =
n=1
_
A
n
cosh
n
b
a +B
n
sinh
n
b
a
_
sin
n
b
y
indicates that the entire expression in the parentheses is given by
A
n
cosh
n
b
a +B
n
sinh
n
b
a =
2
b
_
b
0
G(y) sin
n
b
y dy.
We can now solve for B
n
:
B
n
sinh
n
b
a =
2
b
_
b
0
G(y) sin
n
b
y dy A
n
cosh
n
b
a
B
n
=
1
sinh
n
b
a
_
2
b
_
b
0
G(y) sin
n
b
y dy A
n
cosh
n
b
a
_
. (6)
A solution to the given boundary-value problem consists of the series (4) with coecients A
n
and B
n
given in
(5) and (6), respectively.
637
Exercises 13.5
In Problems 15 and 16 we refer to the discussion in the text under the heading Superposition Principle.
15. We identify a = b = , f(x) = 0, g(x) = 1, F(y) = 1, and G(y) = 1. Then A
n
= 0 and
u
1
(x, y) =
n=1
B
n
sinhny sinnx
where
B
n
=
2
sinhn
_

0
sinnxdx =
2[1 (1)
n
]
n sinhn
.
Next
u
2
(x, y) =
n=1
(A
n
cosh nx +B
n
sinhnx) sinny
where
A
n
=
2
_

0
sinny dy =
2[1 (1)
n
]
n
and
B
n
=
1
sinhn
_
2
_

0
sinny dy A
n
cosh n
_
=
1
sinhn
_
2[1 (1)
n
]
n

2[1 (1)
n
]
n
cosh n
_
=
2[1 (1)
n
]
n sinhn
(1 cosh n).
Now
A
n
cosh nx +B
n
sinhnx =
2[1 (1)
n
]
n
_
cosh nx +
sinhnx
sinhn
(1 cosh n)
_
=
2[1 (1)
n
]
n sinhn
[cosh nxsinhn + sinhnx sinhnxcosh n]
=
2[1 (1)
n
]
n sinhn
[sinhnx + sinhn( x)]
and
u(x, y) = u
1
+u
2
=
2
n=1
1 (1)
n
nsinhn
sinhny sinnx
+
2
n=1
[1 (1)
n
][sinhnx + sinhn( x)]
nsinhn
sinny.
16. We identify a = b = 2, f(x) = 0, g(x) =
_
x, 0 < x < 1
2 x, 1 < x < 2
, F(y) = 0, and G(y) = y(2 y). Then A
n
= 0
and
u
1
(x, y) =
n=1
B
n
sinh
n
2
y sin
n
2
x
where
B
n
=
1
sinhn
_
2
0
g(x) sin
n
2
xdx
=
1
sinhn
__
1
0
xsin
n
2
xdx +
_
2
1
(2 x) sin
n
2
xdx
_
=
8 sin
n
2
n
2
2
sinhn
.
638
0.5 1 1.5 2 2.5 3
x
0.5
1
1.5
2
2.5
3
y
u=0 u=0 u=0
u=0
u=0
u=0
u=1
u=1
u=x u=2-x
u=yH2-yL
u=yH2-yL
0
1
2
3
x
0
1
2
3
y
0
100
200
uHx,yL
0
1
2
3
x
Exercises 13.5
Next, since A
n
= 0 in u
2
, we have
u
2
(x, y) =
n=1
B
n
sinh
n
2
xsin
n
2
where
B
n
=
1
sinhn
_
b
0
y(2 y) sin
n
2
y dy =
16[1 (1)
n
]
n
3
3
sinhn
.
Thus
u(x, y) = u
1
+u
2
=
8
n=1
sin
n
2
n
2
sinhn
sinh
n
2
y sin
n
2
x
+
16
n=1
[1 (1)
n
]
n
3
sinhn
sinh
n
2
xsin
n
2
y.
17. From the gure showing the boundary conditions we see that the maxi-
mum value of the temperature is 1 at (1, 2) and (2, 1).
18. (a)
(b) The maximum value occurs at (/2, ) and is f(/2) = 25
2
.
(c) The coecients are
A
n
=
2
csch n
_

0
100x( x) sinnxdx
=
200 csch n
_
200
n
3
_
1 (1)
n
_
_
=
400
n
3
_
1 (1)
n
_
csch n.
639
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
170
140
110
80
80
60 60
30
30
0
0.2
0.4
0.6
0.8
1
x
0
0.2
0.4
0.6
0.8
1
y
0
50
100
150
200
uHx,yL
0
0.2
0.4
0.6
0.8
x
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
2
1
0.5
0.2
0
-0.05
Exercises 13.5
19. (a)
(b)
20.
Exercises 13.6
1. Using v(x, t) = u(x, t) 100 we wish to solve kv
xx
= v
t
subject to v(0, t) = 0, v(1, t) = 0, and v(x, 0) = 100.
Let v = XT and use
2
as a separation constant so that
X
+
2
X = 0,
X(0) = 0,
X(1) = 0,
and
T
+
2
kT = 0.
640
Exercises 13.6
Then
X = c
1
sinnx and T = c
2
e
kn
2
2
t
for n = 1, 2, 3, . . . so that
v =
n=1
A
n
e
kn
2
2
t
sinnx.
Imposing
v(x, 0) = 100 =
n=1
A
n
sinnx
gives
A
n
= 2
_
1
0
(100) sinnxdx =
200
n
[1 (1)
n
]
so that
u(x, t) = v(x, t) + 100 = 100 +
200
n=1
(1)
n
1
n
e
kn
2
2
t
sinnx.
2. Letting u(x, t) = v(x, t) + (x) and proceeding as in Example 1 in the text we nd (x) = u
0
u
0
x. Then
v(x, t) = u(x, t) + u
0
x u
0
and we wish to solve kv
xx
= v
t
subject to v(0, t) = 0, v(1, t) = 0, and v(x, 0) =
f(x) +u
0
x u
0
. Let v = XT and use
2
as a separation constant so that
X
+
2
X = 0,
X(0) = 0,
X(1) = 0,
and
T
+
2
kT = 0.
Then
X = c
1
sinnx and T = c
2
e
kn
2
2
t
for n = 1, 2, 3, . . . so that
v =
n=1
A
n
e
kn
2
2
t
sinnx.
Imposing
v(x, 0) = f(x) +u
0
x u
0
=
n=1
A
n
sinnx
gives
A
n
= 2
_
1
0
_
f(x) +u
0
x u
0
_
sinnxdx
so that
u(x, t) = v(x, t) +u
0
u
0
x = u
0
u
0
x +
n=1
A
n
e
kn
2
2
t
sinnx.
3. If we let u(x, t) = v(x, t) +(x), then we obtain as in Example 1 in the text
k
+r = 0
or
(x) =
r
2k
x
2
+c
1
x +c
2
.
641
Exercises 13.6
The boundary conditions become
u(0, t) = v(0, t) +(0) = u
0
u(1, t) = v(1, t) +(1) = u
0
.
Letting (0) = (1) = u
0
we obtain homogeneous boundary conditions in v:
v(0, t) = 0 and v(1, t) = 0.
Now (0) = (1) = u
0
implies c
2
= u
0
and c
1
= r/2k. Thus
(x) =
r
2k
x
2
+
r
2k
x +u
0
= u
0
r
2k
x(x 1).
To determine v(x, t) we solve
k
2
v
x
2
=
v
dt
, 0 < x < 1, t > 0
v(0, t) = 0, v(1, t) = 0,
v(x, 0) =
r
2k
x(x 1) u
0
.
Separating variables, we nd
v(x, t) =
n=1
A
n
e
kn
2
2
t
sinnx,
where
A
n
= 2
_
1
0
_
r
2k
x(x 1) u
0
_
sinnxdx = 2
_
u
0
n
+
r
kn
3
3
_
[(1)
n
1]. (1)
Hence, a solution of the original problem is
u(x, t) = (x) +v(x, t)
= u
0
r
2k
x(x 1) +
n=1
A
n
e
kn
2
2
t
sinnx,
where A
n
is dened in (1).
4. If we let u(x, t) = v(x, t) +(x), then we obtain as in Example 1 in the text
k
+r = 0
or
(x) =
r
2k
x
2
+c
1
x +c
2
.
The boundary conditions become
u(0, t) = v(0, t) +(0) = u
0
u(1, t) = v(1, t) +(1) = u
1
.
Letting (0) = u
0
and (1) = u
1
we obtain homogeneous boundary conditions in v:
v(0, t) = 0 and v(1, t) = 0.
Now (0) = u
0
and (1) = u
1
imply c
2
= u
0
and c
1
= u
1
u
0
+r/2k. Thus
(x) =
r
2k
x
2
+
_
u
1
u
0
+
r
2k
_
x +u
0
.
642
Exercises 13.6
To determine v(x, t) we solve
k
2
v
x
2
=
v
dt
, 0 < x < 1, t > 0
v(0, t) = 0, v(1, t) = 0,
v(x, 0) = f(x) (x).
Separating variables, we nd
v(x, t) =
n=1
A
n
e
kn
2
2
t
sinnx,
where
A
n
= 2
_
1
0
[f(x) (x)] sinnxdx. (2)
Hence, a solution of the original problem is
u(x, t) = (x) +v(x, t)
=
r
2k
x
2
+
_
u
1
u
0
+
r
2k
_
x +u
0
+
n=1
A
n
e
kn
2
2
t
sinnx,
where A
n
is dened in (2).
5. Substituting u(x, t) = v(x, t) +(x) into the partial dierential equation gives
k

2
v
x
2
+k
+Ae
x
=
v
t
.
This equation will be homogeneous provided satises
k
+Ae
x
= 0.
The general solution of this dierential equation is
(x) =
A
2
k
e
x
+c
1
x +c
2
.
From (0) = 0 and (1) = 0 we nd
c
1
=
A
2
k
(e
1) and c
2
=
A
2
k
.
Hence
(x) =
A
2
k
e
x
+
A
2
k
(e
1)x +
A
2
k
=
A
2
k
_
1 e
x
+ (e
1)x
.
Now the new problem is
k
2
v
x
2
=
v
t
, 0 < x < 1, t > 0,
v(0, t) = 0, v(1, t) = 0, t > 0,
v(x, 0) = f(x) (x), 0 < x < 1.
Identifying this as the heat equation solved in Section 13.3 in the text with L = 1 we obtain
v(x, t) =
n=1
A
n
e
kn
2
2
t
sinnx
643
Exercises 13.6
where
A
n
= 2
_
1
0
[f(x) (x)] sinnxdx.
Thus
u(x, t) =
A
2
k
_
1 e
x
+ (e
1)x
n=1
A
n
e
kn
2
2
t
sinnx.
k
2
v
x
2
+k
hv h =
v
t
.
k
h = 0.
Since k and h are positive, the general solution of this latter equation is
(x) = c
1
cosh
_
h
k
x +c
2
sinh
_
h
k
x.
From (0) = 0 and () = u
0
we nd c
1
= 0 and c
2
= u
0
/ sinh
_
h/k . Hence
(x) = u
0
sinh
_
h/k x
sinh
_
h/k
.
k
2
v
x
2
hv =
v
t
, 0 < x < , t > 0
v(0, t) = 0, v(, t) = 0, t > 0
v(x, 0) = (x), 0 < x < .
If we let v = XT then
X
X
=
T
+hT
kT
=
2
gives the separated dierential equations
X
+
2
X = 0 and T
+
_
h +k
2
_
T = 0.
The respective solutions are
X(x) = c
3
cos x +c
4
sinx
T(t) = c
5
e
(h+k
2
)t
.
From X(0) = 0 we get c
3
= 0 and from X() = 0 we nd = n for n = 1, 2, 3, . . . . Consequently, it follows
that
v(x, t) =
n=1
A
n
e
(h+kn
2
)t
sinnx
where
A
n
=
2
_

0
(x) sinnxdx.
Hence a solution of the original problem is
u(x, t) = u
0
sinh
_
h/k x
sinh
_
h/k
+e
ht
n=1
A
n
e
kn
2
t
sinnx
644
Exercises 13.6
where
A
n
=
2
_

0
u
0
sinh
_
h/k x
sinh
_
h/k
sinnxdx.
Using the exponential denition of the hyperbolic sine and integration by parts we nd
A
n
=
2u
0
nk(1)
n
(h +kn
2
)
.
k
2
v
x
2
+k
hv h +hu
0
=
v
t
.
k
h +hu
0
= 0 or k
h = hu
0
.
This second-order, linear, non-homogeneous dierential equation has solution
(x) = c
1
cosh
_
h
k
x +c
2
sinh
_
h
k
x +u
0
,
where we assume h > 0 and k > 0. From (0) = u
0
and (1) = 0 we nd c
1
= 0 and
c
2
= u
0
/ sinh
_
h/k . Thus, the steady-state solution is
(x) =
u
0
sinh
_
h
k
sinh
_
h
k
x +u
0
= u
0
_
_
1
sinh
_
h
k
x
sinh
_
h
k
_
_
.
8. The partial dierential equation is
k
2
u
x
2
hu =
u
t
.
Substituting u(x, t) = v(x, t) +(x) gives
k
2
v
x
2
+k
hv h =
v
t
.
k
h = 0.
Assuming h > 0 and k > 0, we have
= c
1
e
h/k x
+c
2
e
h/k x
,
where we have used the exponential form of the solution since the rod is innite. Now, in order that the
steady-state temperature (x) be bounded as x , we require c
1
= 0. Then
(x) = c
2
e
h/k x
and (0) = u
0
implies c
2
= u
0
. Thus
(x) = u
0
e
h/k x
.
a
2

2
v
x
2
+a
2
+Ax =

2
v
t
2
.
a
2
+Ax = 0.
645
Exercises 13.6
The general solution of this dierential equation is
(x) =
A
6a
2
x
3
+c
1
x +c
2
.
From (0) = 0 we obtain c
2
= 0, and from (1) = 0 we obtain c
1
= A/6a
2
. Hence
(x) =
A
6a
2
(x x
3
).
a
2

2
v
x
2
=

2
v
t
2
v(0, t) = 0, v(1, t) = 0, t > 0,
v(x, 0) = (x), v
t
(x, 0) = 0, 0 < x < 1.
Identifying this as the wave equation solved in Section 13.4 in the text with L = 1, f(x) = (x), and g(x) = 0
we obtain
v(x, t) =
n=1
A
n
cos nat sinnx
where
A
n
= 2
_
1
0
[(x)] sinnxdx =
A
3a
2
_
1
0
(x
3
x) sinnxdx =
2A(1)
n
a
2
3
n
3
.
Thus
u(x, t) =
A
6a
2
(x x
3
) +
2A
a
2
n=1
(1)
n
n
3
cos nat sinnx.
10. We solve
a
2
2
u
x
2
g =

2
u
t
2
, 0 < x < 1, t > 0
u(0, t) = 0, u(1, t) = 0, t > 0
u(x, 0) = 0,
u
t
t=0
= 0, 0 < x < 1.
The partial dierential equation is nonhomogeneous. The substitution u(x, t) = v(x, t) +(x) yields a homoge-
neous partial dierential equation provided satises
a
2
g = 0.
By integrating twice we nd
(x) =
g
2a
2
x
2
+c
1
x +c
2
.
The imposed conditions (0) = 0 and (1) = 0 then lead to c
2
= 0 and c
1
= g/2a
2
. Hence
(x) =
g
2a
2
_
x
2
x
_
.
The new problem is now
a
2
2
v
x
2
=

2
v
t
2
, 0 < x < 1, t > 0
v(0, t) = 0, v(1, t) = 0
v(x, 0) =
g
2a
2
_
x x
2
_
,
v
t
t=0
= 0.
646
Exercises 13.6
Substituting v = XT we nd in the usual manner
X
+
2
X = 0
T
+a
2
2
T = 0
with solutions
X(x) = c
3
cos x +c
4
sinx
T(t) = c
5
cos at +c
6
sinat.
The conditions X(0) = 0 and X(1) = 0 imply in turn that c
3
= 0 and = n for n = 1, 2, 3, . . . . The condition
T
(0) = 0 implies c
6
= 0. Hence, by the superposition principle
v(x, t) =
n=1
A
n
cos(ant) sin(nx).
At t = 0,
g
2a
2
_
x x
2
_
=
n=1
A
n
sinnx
and so
A
n
=
g
a
2
_
1
0
_
x x
2
_
sinnxdx =
2g
a
2
n
3
3
[1 (1)
n
] .
Thus the solution to the original problem is
u(x, t) = (x) +v(x, t) =
g
2a
2
_
x
2
x
_
+
2g
a
2
n=1
1 (1)
n
n
3
cos(ant) sin(nx).
11. Substituting u(x, y) = v(x, y) +(y) into Laplaces equation we obtain
2
v
x
2
+

2
v
y
2
+
(y) = 0.
This equation will be homogeneous provided satises (y) = c
1
y +c
2
. Considering
u(x, 0) = v(x, 0) +(0) = u
1
u(x, 1) = v(x, 1) +(1) = u
0
u(0, y) = v(0, y) +(y) = 0
we require that (0) = u
1
,
1
= u
0
and v(0, y) = (y). Then c
1
= u
0
u
1
and c
2
= u
1
. The new
boundary-value problem is
2
v
x
2
+

2
v
y
2
= 0
v(x, 0) = 0, v(x, 1) = 0,
v(0, y) = (y), 0 < y < 1,
where v(x, y) is bounded at x . This problem is similar to Problem 11 in Section 13.5. The solution is
v(x, y) =
n=1
_
2
_
1
0
[(y) sinny] dy
_
e
nx
sinny
= 2
n=1
_
(u
1
u
0
)
_
1
0
y sinny dy u
1
_
1
0
sinny dy
_
e
nx
sinny
=
2
n=1
u
0
(1)
n
u
1
n
e
nx
sinny.
647
Exercises 13.6
Thus
u(x, y) = v(x, y) +(y)
= (u
0
u
1
)y +u
1
+
2
n=1
u
0
(1)
n
u
1
n
e
nx
sinny.
12. Substituting u(x, y) = v(x, y) +(x) into Poissons equation we obtain
2
v
x
2
+
(x) +h +

2
v
y
2
= 0.
The equation will be homogeneous provided satises
(x)+h = 0 or (x) =
h
2
x
2
+c
1
x+c
2
. From (0) = 0
we obtain c
2
= 0. From () = 1 we obtain
c
1
=
1
+
h
2
.
Then
(x) =
_
1
+
h
2
_
x
h
2
x
2
.
The new boundary-value problem is
2
v
x
2
+

2
v
y
2
= 0
v(0, y) = 0, v(, y) = 0,
v(x, 0) = (x), 0 < x < .
This is Problem 11 in Section 13.5. The solution is
v(x, y) =
n=1
A
n
e
ny
sinnx
where
A
n
=
2
_

0
[(x) sinnx] dx
=
2(1)
n
m
_
1
+
h
2
_
h(1)
n
_
n
+
2
n
2
_
.
Thus
u(x, y) = v(x, y) +(x) =
_
1
+
h
2
_
x
h
2
x
2
+
n=1
A
n
e
ny
sinnx.
Exercises 13.7
1. Referring to Example 1 in the text we have
X(x) = c
1
cos x +c
2
sinx
and
T(t) = c
3
e
k
2
t
.
From X
(0) = 0 (since the left end of the rod is insulated), we nd c

2
= 0. Then X(x) = c
1
cos x and the
other boundary condition X
(1) = hX(1) implies

sin +hcos = 0 or cot =

h
.
648
Exercises 13.7
Denoting the consecutive positive roots of this latter equation by
n
for n = 1, 2, 3, . . . , we have
u(x, t) =
n=1
A
n
e
k
2
n
t
cos
n
x.
From the initial condition u(x, 0) = 1 we obtain
1 =
n=1
A
n
cos
n
x
and
A
n
=
_
1
0
cos
n
xdx
_
1
0
cos
2
n
xdx
=
sin
n
/
n
1
2
_
1 +
1
2n
sin2
n
_
=
2 sin
n
n
_
1 +
1
n
sin
n
cos
n
_ =
2 sin
n
n
_
1 +
1
hn
sin
n
(
n
sin
n
)
_
=
2hsin
n
n
[h + sin
2
n
]
.
The solution is
u(x, t) = 2h
n=1
sin
n
n
(h + sin
2
n
)
e
k
2
n
t
cos
n
x.
k

2
v
x
2
+k
=
v
t
.
This equation will be homogeneous if
(x) = 0 or (x) = c
1
x+c
2
. The boundary condition u(0, t) = 0 implies
(0) = 0 which implies c
2
= 0. Thus (x) = c
1
x. Using the second boundary condition we obtain
_
v
x
+
x=1
= h[v(1, t) +(1) u
0
],
which will be homogeneous when
(1) = h(1) +hu

0
.
Since (1) =
(1) = c
1
we have c
1
= hc
1
+hu
0
and c
1
= hu
0
/(h 1). Thus
(x) =
hu
0
h 1
x.
k

2
v
x
2
=
v
t
, 0 < x < 1, t > 0
v(0, t) = 0,
v
x
x=1
= hv(1, t), h > 0, t > 0
v(x, 0) = f(x)
hu
0
h 1
x, 0 < x < 1.
Referring to Example 1 in the text we see that
v(x, t) =
n=1
A
n
e
k
2
n
t
sin
n
x
649
Exercises 13.7
and
u(x, t) = v(x, t) +(x) =
hu
0
h 1
x +
n=1
A
n
e
k
2
n
t
sin
n
x
where
f(x)
hu
0
h 1
x =
n=1
A
n
sin
n
x
and
n
is a solution of
n
cos
n
= hsin
n
. The coecients are
A
n
=
_
1
0
[f(x) hu
0
x/(h 1)] sin
n
xdx
_
1
0
sin
2
n
xdx
=
_
1
0
[f(x) hu
0
x/(h 1)] sin
n
xdx
1
2
_
1
1
2n
sin2
n
_
=
2
_
1
0
[f(x) hu
0
x/(h 1)] sin
n
xdx
1
1
n
sin
n
cos
n
=
2
_
1
0
[f(x) hu
0
x/(h 1)] sin
n
xdx
1
1
hn
(hsin
n
) cos
n
=
2
_
1
0
[f(x) hu
0
x/(h 1)] sin
n
xdx
1
1
hn
(
n
cos
n
) cos
n
=
2h
h + cos
2
n
_
1
0
_
f(x)
hu
0
h 1
x
_
sin
n
xdx.
3. Separating variables in Laplaces equation gives
X
+
2
X = 0
Y
2
Y = 0
and
X(x) = c
1
cos x +c
2
sinx
Y (y) = c
3
cosh y +c
4
sinhy.
From u(0, y) = 0 we obtain X(0) = 0 and c
1
= 0. From u
x
(a, y) = hu(a, y) we obtain X
(a) = hX(a) and

cos a = hsina or tana =
h
.
Let
n
, where n = 1, 2, 3, . . . , be the consecutive positive roots of this equation. From u(x, 0) = 0 we obtain
Y (0) = 0 and c
3
= 0. Thus
u(x, y) =
n=1
A
n
sinh
n
y sin
n
x.
Now
f(x) =
n=1
A
n
sinh
n
b sin
n
x
and
A
n
sinh
n
b =
_
a
0
f(x) sin
n
xdx
_
a
0
sin
2
n
xdx
.
650
Exercises 13.7
Since
_
a
0
sin
2
n
xdx =
1
2
_
a
1
2
n
sin2
n
a
_
=
1
2
_
a
1
n
sin
n
a cos
n
a
_
=
1
2
_
a
1
h
n
(hsin
n
a) cos
n
a
_
=
1
2
_
a
1
h
n
(
n
cos
n
a) cos
n
a
_
=
1
2h
_
ah + cos
2
n
a
,
we have
A
n
=
2h
sinh
n
b[ah + cos
2
n
a]
_
a
0
f(x) sin
n
xdx.
4. Letting u(x, y) = X(x)Y (y) and separating variables gives
X
Y +XY
= 0.
The boundary conditions
u
y
y=0
= 0 and
u
y
y=1
= hu(x, 1)
correspond to
X(x)Y
(0) = 0 and X(x)Y
(1) = hX(x)Y (1)

or
Y
(0) = 0 and Y
(1) = hY (1).
Since these homogeneous boundary conditions are in terms of Y , we separate the dierential equation as
X
X
=
Y
Y
=
2
.
Then
Y
+
2
Y = 0
and
X
2
X = 0
have solutions
Y (y) = c
1
cos y +c
2
siny
and
X(x) = c
3
e
x
+c
4
e
x
.
We use exponential functions in the solution of X(x) since cosh x and sinhx are both unbounded as x .
Now, Y
(0) = 0 implies c
2
= 0, so Y (y) = c
1
cos y. Since Y
(y) = c
1
siny, the boundary condition
Y
(1) = hY (1) implies

c
1
sin = hc
1
cos or cot =

h
.
Consideration of the graphs of f() = cot and g() = /h show that cos = h has an innite number of roots.
The consecutive positive roots
n
for n = 1, 2, 3, . . . , are the eigenvalues of the problem. The corresponding
eigenfunctions are Y
n
(y) = c
1
cos
n
y. The condition lim
x
u(x, y) = 0 is equivalent to lim
x
X(x) = 0. Thus
c
4
= 0 and X(x) = c
3
e
x
. Therefore
u
n
(x, y) = X
n
(x)Y
n
(x) = A
n
e
nx
cos
n
y
651
Exercises 13.7
and by the superposition principle
u(x, y) =
n=1
A
n
e
nx
cos
n
y.
[It is easily shown that there are no eigenvalues corresponding to = 0.] Finally, the condition u(0, y) = u
0
implies
u
0
=
n=1
A
n
cos
n
y.
This is not a Fourier cosine series since the coecients
n
of y are not integer multiples of /p, where p = 1 in
this problem. The functions cos
n
y are however orthogonal since they are eigenfunctions of the Sturm-Lionville
problem
Y
+
2
Y = 0,
Y
(0) = 0
Y
(1) +hY (1) = 0,

with weight function p(x) = 1. Thus we nd
A
n
=
_
1
0
u
0
cos
n
y dy
_
1
0
cos
2
n
y dy
.
Now
_
1
0
u
0
cos
n
y dy =
u
0
n
sin
n
y
1
0
=
u
0
n
sin
n
and
_
1
0
cos
2
n
y dy =
1
2
_
1
0
(1 + cos 2
n
y) dy =
1
2
_
y +
1
2
n
sin2
n
y
_
1
0
=
1
2
_
1 +
1
2
n
sin2
n
_
=
1
2
_
1 +
1
n
sin
n
cos
n
_
.
Since cot = /h,
cos
=
sin
h
and
_
1
0
cos
2
n
y dy =
1
2
_
1 +
sin
2
n
h
_
.
Then
A
n
=
u0
n
sin
n
1
2
_
1 +
1
h
sin
2
n
=
2hu
0
sin
n
n
_
h + sin
2
n
_
and
u(x, y) = 2hu
0
n=1
sin
n
n
_
1 + sin
2
n
_e
nx
cos
n
y
where
n
for n = 1, 2, 3, . . . are the consecutive positive roots of cot = /h.
5. The boundary-value problem is
k

2
u
x
2
=
u
t
, 0 < x < L, t > 0,
u(0, t) = 0,
u
x
x=L
= 0, t > 0,
652
Exercises 13.7
u(x, 0) = f(x), 0 < x < L.
Separation of variables leads to
X
+
2
X = 0
T
+k
2
T = 0
and
X(x) = c
1
cos x +c
2
sinx
T(t) = c
3
e
k
2
t
.
From X(0) = 0 we nd c
1
= 0. From X
(L) = 0 we obtain cos L = 0 and

=
(2n 1)
2L
, n = 1, 2, 3, . . . .
Thus
u(x, t) =
n=1
A
n
e
k(2n1)
2
2
t/4L
2
sin
_
2n 1
2L
_
x
where
A
n
=
_
L
0
f(x) sin
_
2n1
2L
_
xdx
_
L
0
sin
2
_
2n1
2L
_
xdx
=
2
L
_
L
0
f(x) sin
_
2n 1
2L
_
xdx.
a
2

2
v
x
2
+
(x) =

2
v
t
2
.
This equation will be homogeneous if
(x) = 0 or (x) = c
1
x+c
2
. The boundary condition u(0, t) = 0 implies
(0) = 0 which implies c
2
= 0. Thus (x) = c
1
x. Using the second boundary condition, we obtain
E
_
v
x
+
x=L
= F
0
,
E
(L) = F
0
.
Since
(x) = c
1
we conclude that c
1
= F
0
/E and
(x) =
F
0
E
x.
a
2

2
v
x
2
=

2
v
t
2
, 0 < x < L, t > 0
v(0, t) = 0,
v
x
x=L
= 0, t > 0,
v(x, 0) =
F
0
E
x,
v
t
t=0
= 0, 0 < x < L.
Referring to Example 2 in the text we see that
v(x, t) =
n=1
A
n
cos a
_
2n 1
2L
_
t sin
_
2n 1
2L
_
x
where
F
0
E
x =
n=1
A
n
sin
_
2n 1
2L
_
x
653
Exercises 13.7
and
A
n
=
F
0
_
L
0
xsin
_
2n1
2L
_
xdx
E
_
L
0
sin
2
_
2n1
2L
_
xdx
=
8F
0
L(1)
n
E
2
(2n 1)
2
.
Thus
u(x, t) = v(x, t) +(x)
=
F
0
E
x +
8F
0
L
E
2
n=1
(1)
n
(2n 1)
2
cos a
_
2n 1
2L
_
t sin
_
2n 1
2L
_
x.
7. Separation of variables leads to
Y
+
2
Y = 0
X
2
X = 0
and
Y (y) = c
1
cos y +c
2
siny
X(x) = c
3
cosh x +c
4
sinhx.
From Y (0) = 0 we nd c
1
= 0. From Y
(1) = 0 we obtain cos = 0 and

=
(2n 1)
2
, n = 1, 2, 3, . . . .
Thus
Y (y) = c
2
sin
_
2n 1
2
_
y.
From X
(0) = 0 we nd c
4
= 0. Then
u(x, y) =
n=1
A
n
cosh
_
2n 1
2
_
x sin
_
2n 1
2
_
y
where
u
0
= u(1, y) =
n=1
A
n
cosh
_
2n 1
2
_
sin
_
2n 1
2
_
y
and
A
n
cosh
_
2n 1
2
_
=
_
1
0
u
0
sin
_
2n1
2
_
y dy
_
1
0
sin
2
_
2n1
2
_
y dy
=
4u
0
(2n 1)
.
Thus
u(x, y) =
4u
0
n=1
1
(2n 1) cosh
_
2n1
2
_
cosh
_
2n 1
2
_
x sin
_
2n 1
2
_
y.
k

2
u
x
2
=
u
t
, 0 < x < 1, t > 0
u
x
x=0
= hu(0, t),
u
x
x=1
= hu(1, t), h > 0, t > 0,
u(x, 0) = f(x), 0 < x < 1.
Referring to Example 1 in the text we have
X(x) = c
1
cos x +c
2
sinx
and
T(t) = c
3
e
k
2
t
.
654
Exercises 13.7
Applying the boundary conditions, we obtain
X
(0) = hX(0)
X
(1) = hX(1)
or
c
2
= hc
1
c
1
sin +c
2
cos = hc
1
cos hc
2
sin.
Choosing c
1
= and c
2
= h (to satisfy the rst equation above) we obtain
2
sin +hcos = hcos h
2
sin
2hcos = (
2
h
2
) sin.
The eigenvalues
n
are the consecutive positive roots of
tan =
2h
2
h
2
.
Then
u(x, t) =
n=1
A
n
e
k
2
n
t
(
n
cos
n
x +hsin
n
x)
where
f(x) = u(x, 0) =
n=1
A
n
(
n
cos
n
x +hsin
n
x)
and
A
n
=
_
1
0
f(x)(
n
cos
n
x +hsin
n
x)dx
_
1
0
(
n
cos
n
x +hsin
n
x)
2
dx
=
2
2
n
+ 2h +h
2
_
1
0
f(x)(
n
cos
n
x +hsin
n
x)dx.
[Note: the evaluation and simplication of the integral in the denominator requires the use of the relationship
(
2
h
2
) sin = 2hcos .]
9. (a) Using u = XT and separation constant
4
we nd
X
(4)
4
X = 0
and
X(x) = c
1
cos x +c
2
sinx +c
3
cosh x +c
4
sinhx.
Since u = XT the boundary conditions become
X(0) = 0, X
(0) = 0, X
(1) = 0, X
(1) = 0.
Now X(0) = 0 implies c
1
+c
3
= 0, while X
(0) = 0 implies c
2
+c
4
= 0. Thus
X(x) = c
1
cos x +c
2
sinx c
1
cosh x c
2
sinhx.
The boundary condition X
(1) = 0 implies
c
1
cos c
2
sin c
1
cosh c
2
sinh = 0
while the boundary condition X
(1) = 0 implies
c
1
sin c
2
cos c
1
sinh c
2
cosh = 0.
655
0 1 2 3 4 5
10
20
30
40
50
Exercises 13.7
We then have the system of two equations in two unknowns
(cos + cosh )c
1
+ (sin + sinh)c
2
= 0
(sin sinh)c
1
(cos + cosh )c
2
= 0.
This homogeneous system will have nontrivial solutions for c
1
and c
2
provided
cos + cosh sin + sinh

sin sinh cos cosh
= 0
or
2 2 cos cosh = 0.
Thus, the eigenvalues are determined by the equation cos cosh = 1.
(b) Using a computer to graph cosh and 1/ cos = sec we
see that the rst two positive eigenvalues occur near 1.9 and 4.7.
Applying Newtons method with these initial values we nd that
the eigenvalues are
1
= 1.8751 and
2
= 4.6941.
10. (a) In this case the boundary conditions are
u(0, t) = 0,
u
x
x=0
= 0
u(1, t) = 0,
u
x
x=1
= 0.
Separating variables leads to
X(x) = c
1
cos x +c
2
sinx +c
3
cosh x +c
4
sinhx
subject to
X(0) = 0, X
(0) = 0, X(1) = 0, and X
(1) = 0.
Now X(0) = 0 implies c
1
+c
3
= 0 while X
(0) = 0 implies c
2
+c
4
= 0. Thus
X(x) = c
1
cos x +c
2
sinx c
1
cosh x c
2
sinhx.
The boundary condition X(1) = 0 implies
c
1
cos +c
2
sin c
1
cosh c
2
sinh = 0
while the boundary condition X
(1) = 0 implies
c
1
sin +c
2
cos c
1
sinh c
2
cosh = 0.
We then have the system of two equations in two unknowns
(cos cosh )c
1
+ (sin sinh)c
2
= 0
(sin + sinh)c
1
+ (cos cosh )c
2
= 0.
This homogeneous system will have nontrivial solutions for c
1
and c
2
provided
cos cosh sin sinh

sin sinh cos cosh
= 0
656
0 2 4 6 8 10
20
40
60
80
100
Exercises 13.8
or
2 2 cos cosh = 0.
Thus, the eigenvalues are determined by the equation cos cosh = 1.
(b) Using a computer to graph cosh and 1/ cos = sec we see that
the rst two positive eigenvalues occur near the vertical asymp-
totes of sec , at 3/2 and 5/2. Applying Newtons method with
these initial values we nd that the eigenvalues are
1
= 4.7300
and
2
= 7.8532.
Exercises 13.8
1. This boundary-value problem was solved in Example 1 in the text. Identifying b = c = and f(x, y) = u
0
we
have
u(x, y, t) =
m=1
n=1
A
mn
e
k(m
2
+n
2
)t
sinmxsinny
where
A
mn
=
4
2
_

0
_

0
u
0
sinmxsinny dxdy
=
4u
0
2
_

0
sinmxdx
_

0
sinny dy
=
4u
0
mn
2
[1 (1)
m
][1 (1)
n
].
2. This boundary-value problem was solved in Example 1 in the text. Identifying b = c = 1 and f(x, y) =
xy(x 1)(y 1) we have
u(x, y, t) =
m=1
n=1
A
mn
e
k(m
2
+n
2
)t
sinmxsinny
where
A
mn
= 4
_
1
0
_
1
0
xy(x 1)(y 1) sinmxsinny dxdy
= 4
__
1
0
x(x 1) sinmxdx
___
1
0
y(y 1) sinny dy
_
= 4
_
2
m
3
3
_
(1)
m
1
__
2
n
3
3
_
(1)
n
1
_
=
16
m
3
n
3
6
_
(1)
m
1
_
(1)
n
1
.
In Problems 3 and 4 we need to solve the partial dierential equation
a
2
_
2
u
x
2
+

2
u
y
2
_
=

2
u
t
2
.
657
Exercises 13.8
To separate this equation we try u(x, y, t) = X(x)Y (y)T(t):
a
2
(X
Y T +XY
T) = XY T
X
=
Y
Y
+
T
a
2
T
=
2
.
Then
X
+
2
X = 0 (1)
Y
Y
=
T
a
2
T
+
2
=
2
Y
+
2
Y = 0 (2)
T
+a
2
_
2
+
2
_
T = 0. (3)
The general solutions of equations (1), (2), and (3) are, respectively,
X(x) = c
1
cos x +c
2
sinx
Y (y) = c
3
cos y +c
4
siny
T(t) = c
5
cos a
_
2
+
2
t +c
6
sina
_
2
+
2
t.
3. The conditions X(0) = 0 and Y (0) = 0 give c
1
= 0 and c
3
= 0. The conditions X() = 0 and Y () = 0 yield
two sets of eigenvalues:
= m, m = 1, 2, 3, . . . and = n, n = 1, 2, 3, . . . .
A product solution of the partial dierential equation that satises the boundary conditions is
u
mn
(x, y, t) =
_
A
mn
cos a
_
m
2
+n
2
t +B
mn
sina
_
m
2
+n
2
t
_
sinmxsinny.
To satisfy the initial conditions we use the superposition principle:
u(x, y, t) =
m=1
n=1
_
A
mn
cos a
_
m
2
+n
2
t +B
mn
sina
_
m
2
+n
2
t
_
sinmxsinny.
The initial condition u
t
(x, y, 0) = 0 implies B
mn
= 0 and
u(x, y, t) =
m=1
n=1
A
mn
cos a
_
m
2
+n
2
t sinmxsinny.
At t = 0 we have
xy(x )(y ) =
m=1
n=1
A
mn
sinmxsinny.
Using (11) and (12) in the text, it follows that
A
mn
=
4
2
_

0
_

0
xy(x )(y ) sinmxsinny dxdy
=
4
2
_

0
x(x ) sinmxdx
_

0
y(y ) sinny dy
=
16
m
3
n
3
2
[(1)
m
1][(1)
n
1].
658
Exercises 13.8
4. The conditions X(0) = 0 and Y (0) = 0 give c
1
= 0 and c
3
= 0. The conditions X(b) = 0 and Y (c) = 0 yield
two sets of eigenvalues
= m/b, m = 1, 2, 3, . . . and = n/c, n = 1, 2, 3, . . . .
A product solution of the partial dierential that satises the boundary conditions is
u
mn
(x, y, t) = (A
mn
cos a
mn
t +B
mn
sina
mn
t) sin
_
m
b
x
_
sin
_
n
c
y
_
,
where
mn
=
_
(m/b)
2
+ (n/c)
2
. To satisfy the initial conditions we use the superposition principle:
u(x, y, t) =
m=1
n=1
(A
mn
cos a
mn
t +B
mn
sina
mn
t) sin
_
m
b
x
_
sin
_
n
c
y
_
.
At t = 0 we have
f(x, y) =
m=1
n=1
A
mn
sin
_
m
b
x
_
sin
_
n
c
y
_
and
g(x, y) =
m=1
n=1
B
mn
a
mn
sin
_
m
b
x
_
sin
_
n
c
y
_
.
Using (11) and (12) in the text, it follows that
A
mn
=
4
bc
_
c
0
_
b
0
f(x, y) sin
_
m
b
x
_
sin
_
n
c
y
_
dxdy
B
mn
=
4
abc
mn
_
c
0
_
b
0
g(x, y) sin
_
m
b
x
_
sin
_
n
c
y
_
dxdy.
To separate Laplaces equation in three dimensions we try u(x, y, z) = X(x)Y (y)Z(z):
X
Y Z +XY
Z +XY Z
= 0
X
X
=
Y
Y

Z
Z
=
2
.
Then
X
+
2
X = 0 (4)
Y
Y
=
Z
Z
+
2
=
2
Y
+
2
Y = 0 (5)
Z
(
2
+
2
)Z = 0. (6)
The general solutions of equations (4), (5), and (6) are, respectively
X(x) = c
1
cos x +c
2
sinx
Y (y) = c
3
cos y +c
4
siny
Z(z) = c
5
cosh
_
2
+
2
z +c
6
sinh
_
2
+
2
z.
5. The boundary and initial conditions are
u(0, y, z) = 0, u(a, y, z) = 0
u(x, 0, z) = 0, u(x, b, z) = 0
u(x, y, 0) = 0, u(x, y, c) = f(x, y).
659
Exercises 13.8
The conditions X(0) = Y (0) = Z(0) = 0 give c
1
= c
3
= c
5
= 0. The conditions X(a) = 0 and Y (b) = 0 yield
two sets of eigenvalues:
=
m
a
, m = 1, 2, 3, . . . and =
n
b
, n = 1, 2, 3, . . . .
By the superposition principle
u(x, y, t) =
m=1
n=1
A
mn
sinh
mn
z sin
m
a
xsin
n
b
y
where
2
mn
=
m
2
2
a
2
+
n
2
2
b
2
and
A
mn
=
4
ab sinh
mn
c
_
b
0
_
a
0
f(x, y) sin
m
a
xsin
n
b
y dxdy.
u(0, y, z) = 0,
u(x, 0, z) = 0,
u(x, y, 0) = f(x, y),
u(a, y, z) = 0,
u(x, b, z) = 0,
u(x, y, c) = 0.
The conditions X(0) = Y (0) = 0 give c
1
= c
3
= 0. The conditions X(a) = Y (b) = 0 yield two sets of eigenvalues:
=
m
a
, m = 1, 2, 3, . . . and =
n
b
, n = 1, 2, 3, . . . .
Let
2
mn
=
m
2
2
a
2
+
n
2
2
b
2
.
Then the boundary condition Z(c) = 0 gives
c
5
cosh c
mn
+c
6
sinhc
mn
= 0
from which we obtain
Z(z) = c
5
_
cosh w
mn
z
cosh c
mn
sinhc
mn
sinhz
_
=
c
5
sinhc
mn
(sinhc
mn
cosh
mn
z cosh c
mn
sinh
mn
z)
= c
mn
sinh
mn
(c z).
By the superposition principle
u(x, y, t) =
m=1
n=1
A
mn
sinh
mn
(c z) sin
m
a
xsin
n
b
y
where
A
mn
=
4
ab sinhc
mn
_
b
0
_
a
0
f(x, y) sin
m
a
xsin
n
b
y dxdy.
u(0, y, z) = 0,
u(x, 0, z) = 0,
u(x, y, 0) = u
0
,
u(1, y, z) = 0,
u(x, 1, z) = 0,
u(x, y, 1) = u
0
.
660
Chapter 13 Review Exercises
Applying the superposition principle to the solutions in Problems 5 and 6, with a = b = c = 1 and f(x, y) = u
0
in Problem 5 and f(x, y) = u
0
in Problem 6, we get
u(x, y, t) =
m=1
n=1
A
mn
[sinh
mn
z sinh
m
(1 z)] sinmxsinny
where
2
mn
= (m
2
+n
2
)
2
and
A
mn
=
4
sinh
mn
_
1
0
_
1
0
u
0
sinmxsinny dxdy =
4u
0
sinh
mn
__
1
0
sinmxdx
___
1
0
sinny dy
_
=
4u
0
sinh
mn
_
1
m
_
1 (1)
n
__
1
n
_
1 (1)
n
_
=
4u
0
mn
2
sinh
mn
_
1 (1)
m
_
1 (1)
n
.
1. Let u = XY so that
u
yx
= X
,
X
= XY,
and
X
X
=
Y
Y
=
2
.
Then
X
2
X = 0 and Y
2
Y
= 0.
For
2
> 0 we obtain
X = c
1
e
2
x
and Y = c
2
e
y/
2
.
For
2
< 0 we obtain
X = c
1
e
2
x
and Y = c
2
e
y/
2
.
For
2
= 0 we obtain X = c
1
and Y = 0. The general form is
u = XY = A
1
e
A2x+y/A2
.
2. If u = XY then
u
x
= X
Y,
u
y
= XY
,
u
xx
= X
Y,
u
yy
= XY
,
and
X
Y +XY
+ 2X
Y + 2XY
= 0
661
so that
(X
+ 2X
)Y +X(Y
+ 2Y
) = 0,
X
+ 2X
X
=
Y
+ 2Y
Y
=
2
,
X
+ 2X
2
X = 0,
and
Y
+ 2Y
2
Y = 0.
Using
2
as a separation constant we obtain
Y = c
1
e
(1+
1+
2
)y
+c
2
e
(1
1+
2
)y
.
If 1
2
< 0 then
X = e
x
_
c
3
cos
_
2
1 x +c
4
sin
_
2
1 x
_
.
If 1
2
> 0 then
X = c
3
e
(1+
1
2
)x
+c
4
e
(1
1
2
)x
.
If 1
2
= 0 then
X = c
3
e
x
+c
4
xe
x
so that
u = XY =
_
A
1
e
(1+
1+A5 )y
+A
2
e
(1
1+A5 )y
_
e
x
_
A
3
cos
_
A
5
1 x +A
4
sin
_
A
5
1 x
_
=
_
A
1
e
(1+
1+A5 )y
+A
2
e
(1
1+A5 )y
__
A
3
e
(1+
1A5 )x
+A
4
e
(1
1A5 )x
_
=
_
A
1
e
(1+
2 )y
+A
2
e
(1
2 )y
_
_
A
3
e
x
+A
4
xe
x
_
.
Using
2
we obtain the same three solutions except that x and y are interchanged. Using
2
= 0 we obtain
u = XY =
_
A
1
+A
2
e
2x
_ _
A
3
+A
4
e
2y
_
.
3. Substituting u(x, t) = v(x, t) +(x) into the partial dierential equation we obtain
k

2
v
x
2
+k
(x) =
v
t
.
k
= 0 or = c
1
x +c
2
.
Considering
u(0, t) = v(0, t) +(0) = u
0
we set (0) = u
0
so that (x) = c
1
x +u
0
. Now
u
x
x=
=
v
x
x=
(x) = v(, t) +() u

1
is equivalent to
v
x
x=
+v(, t) = u
1
(x) () = u
1
c
1
(c
1
+u
0
),
u
1
c
1
c
1
u
0
= 0 or c
1
=
u
1
u
0
1 +
.
662
The steady-state solution is
(x) =
_
u
1
u
0
1 +
_
x +u
0
.
4. The solution of the problem represents the heat of a thin rod of length . The left boundary x = 0 is kept at
constant temperature u
0
for t > 0. Heat is lost from the right end of the rod by being in contact with a medium
that is held at constant temperature u
1
.
a
2

2
u
x
2
=

2
u
t
2
, 0 < x < 1, t > 0,
u(0, t) = 0, u = (1, t) = 0, t > 0,
u(x, 0) = 0,
u
t
t=0
= g(x), 0 < x < 1.
From Section 13.4 in the text we see that A
n
= 0,
B
n
=
2
na
_
1
0
g(x) sinnxdx =
2
na
_
3/4
1/4
hsinnxdx
=
2h
na
_
1
n
cos nx
_
3/4
1/4
=
2h
n
2
2
a
_
cos
n
4
cos
3n
4
_
and
u(x, t) =
n=1
B
n
sinnat sinnx.
2
u
x
2
+x
2
=

2
u
t
2
, 0 < x < 1, t > 0,
u(0, t) = 1, u(1, t) = 0, t > 0,
u(x, 0) = f(x), u
t
(x, 0) = 0, 0 < x < 1.
Substituting u(x, t) = v(x, t) +(x) into the partial dierential equation gives
2
v
x
2
+
(x) +x
2
=

2
v
t
2
.
This equation will be homogeneous provided
(x) +x
2
= 0 or
(x) =
1
12
x
4
+c
1
x +c
2
.
From (0) = 1 and (1) = 0 we obtain c
1
= 11/12 and c
2
= 1. The new problem is
2
v
x
2
=

2
v
t
2
, 0 < x < 1, t > 0,
v(0, t) = 0, v(1, t) = 0, t > 0,
v(x, 0) = f(x) (x), v
t
(x, 0) = 0, 0 < x < 1.
From Section 13.4 in the text we see that B
n
= 0,
A
n
= 2
_
1
0
[f(x) (x)] sinnxdx = 2
_
1
0
_
f(x) +
1
12
x
4
+
11
12
x 1
_
sinnxdx,
663
and
v(x, t) =
n=1
A
n
cos nt sinnx.
Thus
u(x, t) = v(x, t) +(x) =
1
12
x
4
11
12
x + 1 +
n=1
A
n
cos nt sinnx.
7. Using u = XY and
2
X
2
X = 0,
X(0) = 0,
and
Y
+
2
Y = 0,
Y (0) = 0,
Y () = 0.
Then
Y = c
1
sinny and X = c
2
sinhnx
for n = 1, 2, 3, . . . so that
u =
n=1
A
n
sinhnx sinny.
Imposing
u(, y) = 50 =
n=1
A
n
sinhn sinny
gives
A
n
sinhn =
100
n
1 (1)
n
sinhn
so that
u(x, y) =
100
n=1
1 (1)
n
nsinhn
sinhnx sinny.
8. Using u = XY and
2
X
2
X = 0,
and
Y
+
2
Y = 0,
Y
(0) = 0,
Y
() = 0.
Then
Y = c
1
cos ny
for n = 0, 1, 2, . . . and
X = c
2
or X = c
2
e
nx
for n = 1, 2, 3, . . . (since u must be bounded as x ) so that
u = A
0
+
n=1
A
n
e
nx
cos ny.
664
Imposing
u(0, y) = 50 = A
0
+
n=1
A
n
cos ny
gives
A
0
=
1
_

0
50 dy = 50
and
A
n
=
2
_

0
50 cos ny dy = 0
for n = 1, 2, 3, . . . so that
u(x, y) = 50.
9. Using u = XY and
2
X
2
X = 0,
and
Y
+
2
Y = 0,
Y (0) = 0,
Y () = 0.
Then
Y = c
1
sinny and X = c
2
e
nx
for n = 1, 2, 3, . . .(since u must be bounded as x ) so that
u =
n=1
A
n
e
nx
sinny.
Imposing
u(0, y) = 50 =
n=1
A
n
sinny
gives
A
n
=
2
_

0
50 sinny dy =
100
n
[1 (1)
n
]
so that
u(x, y) =
n=1
100
n
[1 (1)
n
]e
nx
sinny.
k

2
u
x
2
=
u
t
, L < x < L, t > 0,
u(L, t) = 0, u(L, t) = 0, t > 0,
u(x, 0) = u
0
, L < x < L.
Referring to Section 13.3 in the text we have
X(x) = c
1
cos x +c
2
sinx
and
T(t) = c
3
e
k
2
t
.
665
Using the boundary conditions u(L, 0) = X(L)T(0) = 0 and u(L, 0) = X(L)T(0) = 0 we obtain
c
1
cos(L) +c
2
sin(L) = 0
c
1
cos L +c
2
sinL = 0
or
c
1
cos L c
2
sinL = 0
c
1
cos L +c
2
sinL = 0.
Adding, we nd cos L = 0 which gives the eigenvalues
=
2n 1
2L
, n = 1, 2, 3, . . . .
Thus
u(x, t) =
n=1
A
n
e
(
2n1
2L
)
2
kt
cos
_
2n 1
2L
_
x.
From
u(x, 0) = u
0
=
n=1
A
n
cos
_
2n 1
2L
_
x
we nd
A
n
=
2
_
L
0
u
0
cos
_
2n1
2L
_
xdx
2
_
L
0
cos
2
_
2n1
2L
_
xdx
=
u
0
(1)
n+1
2L/(2n 1)
L/2
=
4u
0
(1)
n+1
(2n 1)
.
11. The coecients of the series
u(x, 0) =
n=1
B
n
sinnx
are
B
n
=
2
_

0
sinxsinnxdx =
2
_

0
1
2
[cos(1 n)x cos(1 +n)x] dx
=
1
_
sin(1 n)x
1 n
sin(1 +n)x
1 +n
0
_
= 0 for n = 1.
For n = 1,
B
1
=
2
_

0
sin
2
xdx =
1
_

0
(1 cos 2x) dx = 1.
Thus
u(x, t) =
n=1
B
n
e
n
2
t
sinnx
reduces to u(x, t) = e
t
sinx for n = 1.
k

2
v
x
2
+k
+ sin2x =
v
t
.
k
+ sin2x = 0.
The general solution of this equation is
(x) =
1
4k
2
sin2x +c
1
x +c
2
.
666
From (0) = (1) = 0 we nd that c
1
= c
2
= 0 and
(x) =
1
4k
2
sin2x.
k

2
v
x
2
=
v
t
, 0 < x < 1, t > 0
v(0, t) = 0, v(1, t) = 0, t > 0
v(x, 0) = sinx (x), 0 < x < 1.
If we let v = XT then
X
X
=
T
kT
=
2
gives the separated dierential equations
X
+
2
X = 0 and T
+k
2
T = 0.
The respective solutions are
X(x) = c
3
cos x +c
4
sinx
T(t) = c
5
e
k
2
t
.
From X(0) = 0 we get c
3
= 0 and from X(1) = 0 we nd = n for n = 1, 2, 3, . . . . Consequently, it follows
that
v(x, t) =
n=1
A
n
e
kn
2
2
t
sinnx
where
v(x, 0) = sinx
1
4k
2
sin2x = 0
implies
A
n
= 2
_
1
0
_
sinx
1
4k
2
sin2x
_
sinnxdx.
By orthogonality A
n
= 0 for n = 3, 4, 5, . . . , and only A
1
and A
2
can be nonzero. We have
A
1
= 2
__
1
0
sin
2
xdx
1
4k
2
_
1
0
sin2xsinxdx
_
= 2
_
1
0
1
2
(1 cos 2x) dx = 1
and
A
2
= 2
__
1
0
sinxsin2xdx
1
4k
2
_
1
0
sin
2
2xdx
_
=
1
2k
2
_
1
0
1
2
(1 cos 4x) dx =
1
4k
2
.
Therefore
v(x, t) = A
1
e
k
2
t
sinx +A
2
e
k4
2
t
sin2x
= e
k
2
t
sinx
1
4k
2
e
4k
2
t
sin2x
and
u(x, t) = v(x, t) +(x) = e
k
2
t
sinx +
1
4k
2
(1 e
4k
2
t
) sin2x.
13. Using u = XT and
2
as a separation constant we nd
X
+ 2X
+
2
X = 0 and T
+ 2T
+ (1 +
2
)T = 0.
667
Thus for > 1
X = c
1
e
x
cos
_
2
1 x +c
2
e
x
sin
_
2
1 x
T = c
3
e
t
cos t +c
4
e
t
sint.
For 0 1 we only obtain X = 0. Now the boundary conditions X(0) = 0 and X() = 0 give, in turn,
c
1
= 0 and

2
1 = n or
2
= n
2
+1, n = 1, 2, 3, . . . . The corresponding solutions are X = c
2
e
x
sinnx.
The initial condition T
(0) = 0 implies c
3
= c
4
and so
T = c
4
e
t
_
_
n
2
+ 1 cos
_
n
2
+ 1 t + sin
_
n
2
+ 1 t
_
.
Using u = XT and the superposition principle, a formal series solution is
u(x, t) = e
(x+t)
n=1
A
n
_
_
n
2
+ 1 cos
_
n
2
+ 1 t + sin
_
n
2
+ 1 t
_
sinnx.
14. Letting c = XT and separating variables we obtain
kX
hX
X
=
T
T
or
X
X
=
T
kT
=
2
where = h/k. This leads to the separated dierential equations
X
+
2
X = 0 and T
+k
2
T = 0.
The solution of the second equation is
T(t) = c
3
e
k
2
t
.
For the rst equation we have m =
1
2
(
2
4
2
), and we consider three cases using the boundary conditions
X(0) = X(1) = 0:
2
> 4
2
The solution is X = c
1
e
m1x
+c
2
e
m2x
, where the boundary conditions imply c
1
= c
2
= 0, so X = 0.
(Note in this case that if = 0, the solution is X = c
1
+ c
2
e
x
and the boundary conditions again imply
c
1
= c
2
= 0, so X = 0.)
2
= 4
2
The solution is X = c
1
e
m1x
+c
2
xe
m1x
, where the boundary conditions imply
c
1
= c
2
= 0, so X = 0.
2
< 4
2
The solution is
X(x) = c
1
e
x/2
cos
4
2
2
2
x +c
2
e
x/2
sin
4
2
2
2
x.
From X(0) = 0 we see that c
1
= 0. From X(1) = 0 we nd
1
2
_
4
2
2
= n or
2
=
1
4
(4n
2
2
+
2
).
Thus
X(x) = c
2
e
x/2
sinnx,
and
c(x, t) =
n=1
A
n
e
x/2
e
k(4n
2
2
+
2
)t/4
sin2x.
The initial condition c(x, 0) = c
0
implies
c
0
=
n=1
A
n
e
/2
sinnx. (1)
668
From the self-adjoint form
d
dx
[e
x
X
] +
2
e
x
X = 0
the eigenfunctions are orthogonal on [0, 1] with weight function e
x
. That is
_
1
0
e
x
(e
x/2
sinnx)(e
x/2
sinmx) dx = 0, n = m.
Multiplying (1) by e
x
e
x/2
sinmx and integrating we obtain
_
1
0
c
0
e
x
e
x/2
sinmxdx =
n=1
A
n
e
x
e
x/2
(sinmx)e
x/2
sinnxdx
c
0
_
1
0
e
x/2
sinnxdx = A
n
_
1
0
sin
2
nxdx =
1
2
A
n
and
A
n
= 2c
0
_
1
0
e
x/2
sinnxdx =
4c
0
[2e
/2
n 2n(1)
n
]
e
/2
(
2
+ 4n
2
2
)
=
8nc
0
[e
/2
(1)
n
]
e
/2
(
2
+ 4n
2
2
)
.
669

AEMC13

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AEMC13

Uploaded by

Copyright:

Available Formats

13 Boundary-Value Problems in

(L) = 0. From X(0) = X

(0) = 0 we nd, in turn, c

(0) = 0 (since the left end of the rod is insulated), we nd c

(1) = hX(1) implies

(1) = h(1) +hu

(a) = hX(a) and

(0) = 0 and X(x)Y

(1) = hX(x)Y (1)

(1) = hY (1) implies

(1) +hY (1) = 0,

(L) = 0 we obtain cos L = 0 and

(1) = 0 we obtain cos = 0 and

cos + cosh sin + sinh

(0) = 0, X(1) = 0, and X

cos cosh sin sinh

(x) = v(, t) +() u

You might also like