HOMEWORK 5 SOLUTIONS

Section 3.3
1
Well solve the inhomogeneous diffusion equation on the half-line with Dirichlet boundary conditions using the reflection method. The setup of the problem is as follows:
u kuxx = f (x, t) 0 < x < , 0 < t <
u(0, t) = 0
u(x, 0) = (x).
Because we have a Dirichlet boundary condition, well use odd extensions of f and . Now we must
solve the problem
vt kvxx = fodd (x, t) < x < , 0 < t <
v(0, t) = 0
v(x, 0) = odd (x).
Now we use the formula for the solution on the whole line:
v(x, t) =

S(x y, t)(y)dy +

S(x y, t)odd (y)dy +

S(x y, t s)f (y, s)dyds

S(x y, t)((y))dy +

t
0

S(x y, t s)fodd (y, s)dyds

S(x y, t s)(f (y, s))dyds

[S(x y, t)(y) S(x + y, t)(y)] dy +

[S(x y, t) S(x + y, t)] (y)dy +

0
0

[S(x y, t s)f (y, s) S(x + y, t s)f (y, s)] dyds

[S(x y, t s) S(x + y, t s)] f (y, s)dyds

3
Let W (x, t) = w(x, t) xh(t). Then we have
Wt kWxx = wt kwxx xh (t) = xh (t)
Wx (0, t) = wx (0, t) h(t) = h(t) h(t) = 0
W (x, 0) = w(x, 0) xh(0) = (x) xh(0).
Now we solve the inhomogeneous Neumann problem on the whole line using even extentions.
Let (x) = (x) xh(0) and f (x, t) = xh (t) for 0 < x < . Then we have
1

HOMEWORK 5 SOLUTIONS

W (x, t) =

0
t

t
0

S(x y, t s)feven (y, s)dyds

S(x y, t)(y)dy

S(x y, t s)f (y, s)dyds +

S(x y, t s)f (y, s)dyds

S(x + y, t)(y)dy

S(x y, t s)f (y, s)dyds

[S(x y, t) + S(x + y, t)](y)dy +

S(x + y, t s)f (y, s)dyds

[S(x y, t s) + S(x + y, t s)]f (y, s)dyds

[S(x y, t) + S(x + y, t)][(y) yh(0)]dy

t

S(x y, t)(y)dy

= +

S(x y, t)(y)dy +

S(x y, t)even (y)dy +

[S(x y, t s) + S(x + y, t s)][yh (s)]dyds.

Finally, we use the fact that W (x, t) = w(x, t) xh(t) so we have

w(x, t) =

[S(xy, t)+S(x+y, t)][(y)yh(0)]dy

t
0

[S(xy, ts)+S(x+y, ts)][yh (s)]dyds+xh(t)

3.4
1
We use Theorem 1 with f (x, t) = xt, (x) = 0 and (x) = 0. This gives
u(x, t) =
=
=
=
=

1
f
2c

t
x+c(ts)
1
ysdyds

2c 0 xc(ts)
t
1
1 2 x+c(ts)
s [ y xc(ts) ] ds
2c 0
2
t
1
1
s((x2 + 2xc(t s) + c2 (t s)2 ) (x2 2xc(t s) + c2 (t s)2 ))ds

2c 0 2
t
1
s(4xc(t s))ds
4c 0

xts xs2 ds

t
1
1
= xts2 xs3
0
2
3
1 3 1 t
= xt xt
2
3
1 3
= xt .
6

HOMEWORK 5 SOLUTIONS

3
We use Theorem 1 with f (x, t) = cos(x), (x) = sin(x) and (x) = 1 + x. This gives
x+ct
1
1
1
u(x, t) = [(x + ct) + (x ct)] +
(s)ds + f
2
2c xct
2c

x+ct
t
x+c(ts)
1
1
1
= [sin(x + ct) + sin(x ct)] +
(1 + s)ds +
cos(y)dyds
2
2c xct
2c 0 xc(ts)
Well evaluate each term separately. The first term simplifies to sin(x) cos(ct) by the angle sum
formula (although you are not required to perform this simplification).
The second term is:
x+ct
x+ct
1
1
1
(s)ds = (s + s2 )

xct
2c xct
2c
2
1
= (2ct + 2xct)
2c
= t + xt.
The third term is:
t
x+c(ts)
t
x+c(ts)
1
1
cos(y)dyds =

sin(y)xc(ts) ds
2c 0 xc(ts)
2c 0
t
1
=
2 cos(x) sin(c(t s))ds
2c 0
t
1
= 2 cos(x) cos(c(t s))
0
c
1
= 2 cos(x)(1 cos(ct)).
c
Putting it all together we have
1
u(x, t) = sin(x) cos(ct) + t + xt + 2 cos(x)(1 cos(ct)).
c
5
This problem isnt pretty, but here it is worked out in excruciating detail. Well make extensive
use of Theorem 3 from Appendix A.3, which says the following:
b(t)
b(t) d
d
f (x, t)dx =
f (x, t)dx f (a(t), t)a (t) + f (b(t), t)b (t).

dt a(t)
a(t) dt
First well verify the initial conditions:
0
x+c(0s)
1
f (y, s)dyds

2c 0 xc(0s)
=0

u(x, 0) =
since were integrating from 0 to 0.

t
x+c(ts)
d 1
f (y, s)dyds

dt 2c 0 xc(ts)
t d
x+c(ts)
x+c(tt)
1
=
(
[
f (y, s)dy] ds +
f (y, t)dy) .
2c 0 dt xc(ts)
xc(tt)

ut (x, t) =

When we set t = 0 the first term disappears because were integrating from 0 to 0 and the second
term disappears because were integrating from x to x. Thus, ut (x, 0) = 0 as required.

HOMEWORK 5 SOLUTIONS

Now we proceed to verify that utt = c2 uxx + f . Continuing the computation from above we have
ut (x, t) =

t d
x+c(ts)
x+c(tt)
1
(
[
f (y, s)dy] ds +
f (y, t)dy)
2c 0 dt xc(ts)
xc(tt)

t d
x+c(ts)
1
[
f (y, s)dy] ds

2c 0 dt xc(ts)

t
x+c(ts) d
1
f (y, s)dy f (x c(t s), s)(c) + f (x + c(t s), s)(c)) ds
(
2c 0
xc(ts) dt

Since neither y nor s depend on t, we have

d
dt f (y, s)

= 0 so this becomes

t
1
cf (x c(t s), s) + cf (x + c(t s), s)ds.
2c 0
Now we differentiate with respect to t again:
t
d 1
utt (x, t) =
cf (x c(t s), s) + cf (x + c(t s), s)ds
dt 2c 0
t d
1
(
[cf (x c(t s), s) + cf (x + c(t s), s)] ds + cf (x c(t t), t) + cf (x + c(t t), t))
=
2c 0 dt
t d
1
(
[cf (x c(t s), s) + cf (x + c(t s), s)] ds + 2cf (x, t))
=
2c 0 dt
t d
1
=
(
[cf (x c(t s), s) + cf (x + c(t s), s)] ds) + f (x, t)
2c 0 dt
In order to differential the terms inside the integral, well need to apply the chain rule. To make it
easier to read, lets let a = x c(t s) and b = x + c(t s). Then we have
t d
1
=
(
[cf (a, s) + cf (b, s)] ds) + f (x, t)
2c 0 dt
t
1
f b f s
f a f s
=
( c [
+
] + c[
+
]) ds + f (x, t)
2c 0
a t s t
b t s t
t f a
1
f s
f b f s
= ( [
+
]+[
+
]) ds + f (x, t)
2 0 a t s t
b t s t

Since s doesnt depend on t,

finally get a formula for utt :

s
t

= 0. We also have

utt (x, t) =

a
t

= c and

b
t

= c. Putting this together we

t
c
[fa (a, s) + fb (b, s)]ds + f (x, t).
2 0

Now well compute uxx :

t
x+c(ts)
d d 1
f (y, s)dyds

dx dx 2c 0 xc(ts)
t d
x+c(ts)
d 1
=
[
f (y, s)dy] ds

dx 2c 0 dx xc(ts)

uxx (x, t) =

t
x+c(ts) d
d 1
(
f (y, s)dy f (x c(t s), s) + f (x + c(t s), s)) ds

dx 2c 0
xc(ts) dx

As before, since f (y, s) doesnt depend on x, the

t

d
dx f (y, s)

term drops out and we have

d 1
(f (x c(t s), s) + f (x + c(t s), s)) ds
dx 2c 0
t d
1
=
[f (x c(t s), s) + f (x + c(t s), s)] ds

2c 0 dx
=

HOMEWORK 5 SOLUTIONS

As before, lets let a = x c(t s) and b = x + c(t s). Then we have

t d
1
=
[f (a, s) + f (b, s)] ds

2c 0 dx
t
1
f a f s
f b f s
=
+
]+[
+
] ds
[
2c 0
a x s x
b x s x
t
1
=
fa (a, s) + fb (b, s)ds
2c 0
t
1
=
fa (a, s) + fb (b, s)ds
2c 0
That was a lot of work, but now we have nice expressions for utt and uxx . They are as follows:
t
c
utt = fa (a, s) + fb (b, s)ds + f (x, t)
2 0
t
1
uxx =
fa (a, s) + fb (b, s)ds
2c 0
It is now immediately clear that utt = c2 uxx + f (x, t) as desired.