Math 462: HW2 Solutions

Due on July 25, 2014

Jacky Chong
1
Jacky Chong
Remark: We are working in the context of Riemann Integrals.
Problem 1
2.1.1 Solve u
tt
= c
2
u
xx
, u(x, = e
x
, u
t
(x, 0) = sin x.
Solution: By dAlemberts formula, we have that
u(x, t) =
1
2
[e
x+ct
+ e
xct
] +
1
2c

x+ct
xct
sin s ds
= e
x

e
ct
+ e
ct
2

+
1
2c
[cos(x ct) cos(x + ct)]
= e
x
cosh ct +
1
2c
[cos xcos ct + sin xsin ct cos xcos ct + sin xsin ct]
= e
x
cosh ct +
1
c
sin xsin ct.
Problem 2
2.1.5 (The hammer blow) Let (x) 0 and (x) = 1 for |x| < a and (x) = 0 for |x| a. Sketch the string
prole (u versus x) at each of the successive instants t = a/2c, a/c, 3a/2c, 2a/c, and 5a/c.
Solution: At t = a/2c, we have that
u(x, a/2c) =
1
2c

x+a/2
xa/2
(s) ds
=

0 for x
3
2
a,
1
2c

3
2
a + x

for
3
2
a < x
1
2
a,
a
2c
for
1
2
a < x <
1
2
a,
1
2c

3
2
a x

for
1
2
a x <
3
2
a,
0 for x
3
2
a.
At t = a/c, we have that
u(x, a/c) =
1
2c

x+a
xa
(s) ds
=

0 for x 2a,
1
2c
(2a + x) for 2a < x < 0,
1
2c
(2a x) for 0 x < 2a,
0 for x 2a.
Page 2 of 7
Jacky Chong Problem 2
At t = 3a/2c, we have that
u(x, 3a/2c) =
1
2c

x+3a/2
x3a/2
(s) ds
=

0 for x
5
2
a,
1
2c

5
2
a + x

for
5
2
a < x <
1
2
a,
a
c
for
1
2
a x
1
2
a,
1
2c

5
2
a x

for
1
2
a < x <
5
2
a,
0 for x
5
2
a.
At t = 2a/c, we have that
u(x, 2a/c) =
1
2c

x+2a
x2a
(s) ds
=

0 for x 3a,
1
2c
(3a + x) for 3a < x < a,
a
c
for a x a,
1
2c
(3a x) for a < x < 3a,
0 for x 3a.
And nally, at t = 5a/c, we have that
u(x, 5a/c) =
1
2c

x+5a
x5a
(s) ds
=

0 for x 6a,
1
2c
(6a + x) for 6a < x < 4a,
a
c
for 4a x 4a,
1
2c
(6a x) for 4a < x < 6a,
0 for x 6a.
Problem 3
2.1.8 A spherical wave is a solution of the three-dimensional wave equation of the form u(r, t), where r is
the distance to the origin (the spherical coordinate). The wave equation takes the form
u
tt
= c
2

u
rr
+
2
r
u
r

(spherical wave equation).

(a) Change variables v = ru to get the equation for v : v
tt
= c
2
u
rr
.
(b) Solve for v using (3) and thereby solve the spherical wave equation.
(c) Use (8) to solve it with initial conditions u(r, 0) = (r), u
t
(r, 0) = (r), taking both (r) and (r) to be
even functions of r.
Solution:
Problem 3 continued on next page. . . Page 3 of 7
Jacky Chong Problem 3 (continued)
(a) Dene v(r, t) := ru(r, t), then we see that v
t
= ru
t
, v
tt
= ru
tt
, v
r
= u+ru
r
, and v
rr
= 2u
r
+ru
rr
. Hence
it follows
v
tt
c
2
v
rr
= ru
tt
c
2
(2u
r
+ ru
rr
)
= r

u
tt
c
2
2
r
u
r
c
2
u
rr

= 0.
(b) By equation (3), we get that
v(r, t) = f(r + ct) + g(r ct)
which means
u(r, t) =
f(r + ct)
r
+
g(r ct)
r
.
(c) Observe

u
tt
= c
2

u
rr
+
2
r
u
r

0 < r < and < t <

u(r, 0) = (r), u
t
(r, 0) = (r)
gives rise to the problem

v
tt
= c
2
v
rr
< r < and < t <
v(r, 0) =

(r) := r(r), v
t
(r, 0) =

(r) := r(r)
since

(r) and

(r) are both odd functions. Using dAlemberts formula, we get that
v(r, t) =
1
2
[

(r + ct) +

(r ct)] +
1
2c

r+ct
rct

(s) ds
=
1
2
[(r + ct)(r + ct) + (r ct)(r ct)] +
1
2c

r+ct
rct
s(s) ds
then it follows
u(r, t) =
1
2r
[(r + ct)(r + ct) + (r ct)(r ct)] +
1
2cr

r+ct
rct
s(s) ds
for r > 0 and < t < .
Problem 4
2.2.6 Prove that, among all possible dimensions, only in three dimensions can one have distortionless spher-
ical wave propagation with attenuation. This means the following. A spherical wave in n-dimensional space
satises the PDE
u
tt
= c
2

u
rr
+
n 1
r
u
r

,
where r is the spherical coordinate. Consider such a wave that has the special form u(r, t) = (r)f(t (r)),
where (r) is called the attenuation and (r) the delay. The question is whether such solutions exist for
arbitrary functions f.
(a) Plug the special form into the PDE to get an ODE for f.
(b) Set the coecients of f

, f

, and f equal to zero.

(c) Solve the ODEs to see that n = 1 or n = 3 (unless u 0).
Problem 4 continued on next page. . . Page 4 of 7
Jacky Chong Problem 4 (continued)
(d) If n = 1, show that (r) is a constant (so that there is no attenuation).
Solution:
(a) If u(r, t) = (r)f(t (r)), then we see that
u
r
(r, t) =

(r)f(t (r)) (r)

(r)f

(t (r))
and
u
rr
(r, t) =

(r)f(t (r)) 2

(r)

(r)f

(t (r))
(r)

(r)f

(t (r)) + (r)[

(r)]
2
f

(t (r)).
Then it follows
u
tt
c
2

u
rr
+
n 1
r
u
r

= f

c
2

f 2

+ [

]
2
f

+
n 1
r
[

= (1 c
2
[

]
2
)f

+ c
2

+
n 1
r

c
2

n 1
r

f = 0.
(b) Assume (r) = 0. Setting the coecients of f

, f

, and f equal to zero, we get the following system of

dierential equations

1 c
2
[

]
2
= 0
2

+
n1
r

= 0
n1
r

= 0.
(c) By the rst equation, we see that

(r) =
1
c
which means either
(r) =
r
c
+
0
or (r) =
r
c
+
0
.
By the rst ode, we see that the second ode reduces to

+
n 1
2r
= 0
which means
(r) =

C if n = 1
Cr
(n1)/2
if n > 1
Next, let us solve the third ode. Its clear that when n = 1, then C is also a solution to the third ode.
If n > 1, then we see that
n 1
r

= C

(n 1)
2
2
+
(n 1)(n + 1)
4

r
(n+3)/2
= C
(n 1)(n 3)
4
r
(n+3)/2
= 0
which means n = 3. Thus, the system has a solution provided n = 1 or n = 3.
(d) Its clear from (c) that when n = 1, we have that (r) = const.
Page 5 of 7
Jacky Chong Problem 4
Problem 5
2.3.3 Consider the diusion equation u
t
= u
xx
in the interval (0, 1) with u(0, t) = u(1, t) = 0 and u(x, 0) =
1 x
2
. Note that this initial function does not satisfy the boundary condition at the left end, but that the
solution will satisfy it for all t > 0.
(a) Show that u(x, t) > 0 at all interior points 0 < x < 1, 0 < t < .
(b) For each t > 0, let (t) = the maximum of u(x, t) over 0 x 1. Show that (t) is a decreasing (i.e.,
non increasing) function of t.
(c) Draw a rough sketch of what you think the solution looks like (u versus x) at a few times.
Solution:
(a) If u(x, t) solves the above diusion equation, then v(x, t) := u(x, t) also solves the following Cauchy
problem

v
t
v
xx
= 0 on 0 < x < 1, 0 < t <
v(0, t) = v(1, t) = 0, v(x, 0) = x
2
1
since
v
t
v
xx
= (u
t
u
xx
) = 0.
Fix T > 0. Consider the rectangle R
T
= {(x, t) | 0 x 1 and 0 t T} and the parabolic boundary,

T
= {(x, t) | x = 0 or x = 1 for 0 t T} {(x, t) | 0 x 1 and t = 0}.
By the strong maximum principle, we know that the maximum of v(x, t) in R
T
is only attained on
T
,
i.e. v(x, t) < 0 for all (x, t) R
T

T
since the maximum value of v on
T
is zero. Hence it follows
u(x, t) > 0 for all (x, t) R
T

T
. Moreover, since this holds for all xed T, then we see that u(x, t) > 0
for all (x, t) {(x, t) | 0 x 1 and 0 t < }.
(b) Assume the maximum of u(x, t) at time t occurs at X(t). By part (a), we know that X(t) (0, 1) since
u(x, t) > 0 for (x, t) R
T

T
. Fix T > 0 and dene (x) = u(x, T), then it follows

x
u(x, t)

(x,t)=(X(T),T)
=
d
dx
(x)

x=X(T)
= 0
and

2
x
2
u(x, t)

(x,t)=(X(T),T)
=
d
2
dx
2
(x)

x=X(T)
0
since X(T) is a maximum on {(x, T) | 0 x 1}. Now, dene (t) = u(X(t), t) and assuming X(t) is
dierentiable, we get that
d
dt
(t) = u
t
(X(t), t) + u
x
(X(t), t)X

(t)
= u
xx
(X(t), t) + u
x
(X(t), t)X

(t)
= u
xx
(X(t), t) 0.
Thus, we see that (t) must be a decreasing function.
(c) Left to the reader.
Page 6 of 7
Jacky Chong Problem 5
Problem 6
2.3.4 Consider the diusion equation u
t
= u
xx
in {0 < x < 1, 0 < t < } with u(0, t) = u(1, t) = 0 and
u(x, 0) = 4x(1 x).
(a) Show that 0 < u(x, t) < 1 for all t > 0 and 0 < x < 1.
(b) Show that u(x, t) = u(1 x, t) for all t 0 and 0 x 1.
(c) Use the energy method to show that

1
0
u
2
dx is strictly decreasing function of t.
Solution:
(a) Fix T > 0. By the strong maximum principle, we know that the maximum of u(x, t) in R
T
is only
attained on
T
, i.e. u(x, t) < 1 for all (x, t) R
T

T
since the maximum value of u on
T
is 1 which
occurs at the point (1/2, 0).
Likewise, applying the strong maximum principle to v(x, t) := u(x, t), we see that v(x, t) < 0 on
R
T

T
, i.e. u(x, t) > 0 on R
T

T
. Thus, we have that 0 < u(x, t) < 1 on R
T

T
. Since this holds
for all T, then we see that 0 < u(x, t) < 1 for all (x, t) {(x, t) | 0 x 1 and 0 t < }.
(b) Dene v(x, t) = u(1 x, t). Observe v
t
(x, t) = u
t
(1 x, t), v
x
(x, t) = u
x
(1 x, t), and u
xx
(1 x, t)
which means
v
t
v
xx
= u
t
(1 x, t) u
xx
(1 x, t) = 0
since 0 < 1 x < 1. Now, its clear that v(x, t) solves the same Dirichlet problem as u(x, t)

v
t
v
xx
= 0 on 0 < x < 1, 0 < t <
v(0, t) = v(1, t) = 0, v(x, 0) = 4x(1 x).
By the uniqueness theorem, we have that u(x, t) = v(x, t) = u(1 x, t) for t 0 and 0 x 1.
(c) Observe
d
dt

1
0
u(x, t)
2
dx = 2

1
0
u
x
(x, t)
2
dx.
If u
x
(x, t) 0, then we see that u
xx
0 which means u
t
0. Thus, it follows u(x, t) = const, which
clearly does not satisfy the initial condition. Hence u
x
(x, t) = 0 which means

1
0
u
x
(x, t)
2
dx > 0
i.e., we have that
d
dt

1
0
u(x, t)
2
dx = 2

1
0
u
x
(x, t)
2
dx < 0.
Page 7 of 7