You are on page 1of 6

MA1506 TUTORIAL 11 SOLUTIONS

QUESTION 1.
(a) Setting u(x, y) = X(x)Y (y), the p.d.e. yux xuy = 0 becomes yX 0 Y xXY 0 = 0.
Dividing by XY gives
X0 Y0
y x = 0
X Y
1 X 0 1 Y0
i.e. = = k1 (constant)
x X y Y
1 X0
This gives two separable o.d.e., the first of which is = k1 . Integrating this yields
x X
1
ln |X| = k1 x2 + c1
2
2 1
i.e. X = ec1 ecx , where c = k1
2
1 Y0 2
Similarly, integrating the second separable o.d.e. = k1 gives Y = ec2 ecy .
y Y
Thus,
2 2
u(x, y) = XY = ec1 ecx ec2 ecy
2 +y 2 )
= kec(x .

(b) Setting u(x, y) = X(x)Y (y), the p.d.e. ux = yuy becomes X 0 Y = yXY 0 . Dividing by XY
gives
X0 Y0
= y = c (constant)
X Y
X0
This gives two separable o.d.e., the first of which is = c. Integrating this yields
X
ln |X| = cx + k1
i.e. X = ek1 ecx .

Y0 c
Similarly, integrating the second separable o.d.e. = gives
Y y

ln |Y | = c ln |y| + k2
i.e. Y = ek2 y c .

Thus,

u(x, y) = XY = ek1 ecx ek2 y c


= ky c ecx .

(c) Setting u(x, y) = X(x)Y (y), the p.d.e. uxy = u becomes X 0 Y 0 = XY . Dividing by XY gives

X0 Y 0
= 1.
X Y
MA1505 Tutorial 11 Solutions

X0 Y0 X0 Y0 1
This implies that both and are nonzero constants and we set = c and = . This
X Y X Y c
X0
gives two separable o.d.e., the first of which is = c. Integrating this yields
X

ln |X| = cx + k1
i.e. X = ek1 ecx .

Y0 1
Similarly, integrating the second separable o.d.e. = gives
Y c
y
ln |Y | = + k2
c
i.e. Y = ek2 ey/c .

Thus,

u(x, y) = XY = ek1 ecx ek2 ey/c


= kecx+y/c .

(d) Setting u(x, y) = X(x)Y (y), the p.d.e. xuxy + 2yu = 0 becomes xX 0 Y 0 + 2yXY = 0.
Dividing by 2yXY gives
X0 1 Y0
  
x = 1.
X 2y Y
X0 1 Y 0 X0 1 Y 0 1
This implies that both x and are nonzero constants. We set x = c and = ,
X 2y Y X 2y Y c
X0 c Y0 2y
which respectively give two separable o.d.e. = and = .
0
X x Y c
X c
Integrating the first o.d.e. = yields
X x

ln |X| = c ln x + k1
i.e. X = ek1 xc .

Y0 2y
Similarly, integrating the second separable o.d.e. = gives
Y c
y2
ln |Y | = + k2
c
2
i.e. Y = ek2 ey /c .

Thus,
2 /c
u(x, y) = XY = ek1 xc ek2 ey
2 /c
= kxc ey .

2
Question 2
According to the lecture notes, we just need to extend f(x) = x to an odd function on the
domain (, ) [which is immediate, because y = x is already odd!], and then turn that
into a periodic function of period 2. The result is a discontinuous function with jumps;
at those jumps, the average value is zero [which is also the value of y(0, )]. Next, we just
need the Fourier sine coefficients of y = x, that is,
" #
2Z 2 xcos(nx) sin(nx) 2 2
bn = xsin(nx)dx = + 2
= cos(n) = (1)n+1 .
0 n n 0
n n

So again according to the lecture notes, the solution of the wave equation in this case is

2
(1)n+1 sin(nx)cos(nct).
X
y(t, x) =
n=1 n

Notice that y(0, ) from this formula is zero, as it should be.


For the heat equation, we just need to substitute into the general solution given in the
notes [with L = ], and so we get

2
(1)n+1 sin(nx)exp(n2 c2 t).
X
y(x, t) =
n=1 n

Question 3
We want to solve
c2 yxx = ytt ,
subject to the four conditions

y(t, 0) = y(t, ) = 0, y(0, x) = f (x), yt (0, x) = 0,

where f(x) is a given function which is zero at 0 and , and which we extend when
necessary to be an odd periodic function of period 2. DAlembert claimed that the
following object satisfies all of the above:
1h i
y(t, x) = f (x + ct) + f (x ct) .
2
First, note that if y(t,x) is given by this formula, then

c2 h 00 i
ytt = f (x + ct) + f 00 (x ct) ,
2
while
1 h 00 i
yxx = f (x + ct) + f 00 (x ct) .
2
So it does satisfy the equation. We also see that
1h i
y(t, 0) = f (ct) + f (ct) ,
2
which is zero because f(x) is odd. Then
1h i 1h i
y(t, ) = f ( + ct) + f ( ct) = f ( + ct) + f ( ct) ,
2 2
because f(x) is periodic with period 2, and this too vanishes because f(x) is odd. Next,
clearly
1h i
y(0, x) = f (x) + f (x) = f (x),
2
and finally
1h i
yt (0, x) = cf 0 (x) cf 0 (x) = 0.
2
So DAlembert was right.

Question 4
If you take the wave equation
c2 yxx = ytt ,
and change the variables from (x,y) to (, ), where = ax, = at, for any constant a,
then the equation continues to hold because you get a factor involving a2 coming out on
both sides of the equation. We can use this here, because suppose that x ranges from 0
to L: then define
x t
= , = .
L L
Then the range of is 0 to , so all our familiar formulae for the wave equation hold when
we use and . Thus for example the function f(x) giving the initial shape of the string,
when expressed in terms of , will have a Fourier sine series with coefficients of the form
2Z
bn = f ()sin(n)d,
0
which, returning to the variable x, can be written [since d = dx/L]

2ZL nx
bn = f (x)sin( )dx.
L 0 L
The solution of the wave equation likewise takes the form

X
bn sin(n)cos(nc ),
n=1

which means that it is


X nx nct
bn sin( )cos( ),
n=1 L L
and this is the answer to our question. In summary: all you have to do is to put in factors
of /L, BUT you have to remember to do it for both x AND t.

2
Question 5
To solve
ytt = c2 yxx byt ,
set y(t,x) = T(t)X(x), and get

T 00 X = c2 T X 00 bT 0 X.

Divide throughout by TX:


T 00 2X
00
T0
=c b ,
T X T
from which we get
T 00 T0 X 00
+ b = c2 .
T T X
As usual, both sides have to be equal to some constant K, and so we get a pair of
ordinary differential equations:

T 00 + bT 0 + KT = 0,

X 00 + c2 KX = 0,
and so we have successfully separated the variables. To get y(t,0) = y(t,) = 0, we want
X(0) = X() = 0, and, as usual, that means that we want trigonometric solutions for
X(x); so from the second equation we see that K is positive. No doubt you now recognise
the first equation as the equation of motion of the damped simple harmonic oscillator.
[Remember that, in the ODE for the damped harmonic oscillator, all of the coefficients
were positive, and that was very important.] So the solutions for T(t) will be [if b is small
enough] products of exponentials with trigonometric functions, as in Chapter 2.

Question 6
Because the second derivative of Tx/L with respect to x, and its first derivative with
respect to t, are both zero, it is easy to see that if u(x, t) satisfies the heat equation, then
so must u (x, t). Furthermore u (0, t) = u(0, t) = 0, and u (L, t) = u(L, t) (T L/L) =
T T = 0, so actually u (x, t) has the same value, zero, at the two ends of the bar,
as in the lecture notes. But we have to be careful: the initial condition has changed:
u (x, 0) = u(x, 0) (T x/L) = f (x) (T x/L). So we can use the result given in the
lecture notes, but using f (x) (T x/L) instead of f (x):

2 n 2 c2
!
nx
 

X
u (x, t) = Bn sin exp t ,
n=1 L L2

where the Bn are the Fourier sine coefficients of f (x) (T x/L). You can see that
2 Z L Tx nx
 
Bn = bn sin dx,
L 0 L L
where the bn are the Fourier sine coefficients of f (x). Integrating by parts [as in question
2, above] we get
2T
Bn = bn + (1)n ,
n
3
and so, since u(x, t) = u (x, t) + (T x/L), we have finally

2 n 2 c2
!
Tx X 2T nx
   
u(x, t) = + bn + (1)n sin exp t .
L n=1 n L L2

Notice that this tends to Tx/L as t , which makes sense physically (the temperature
should change linearly from the left to the right once things settle down.)

Question 7
Here we use the following useful observation: the Laplace equation is completely sym-
metric between x and y. That is, if you switch x and y in the Laplace equation, nothing
changes. Furthermore, the boundary conditions we have here [with u(x,y) equal to zero
on all sides of the square except the left vertical side, ie along the y axis] are obtained by
flipping x and y in the problem in the lecture notes [the one at the very end of Chapter
8, where u(x,y) is equal to zero on all sides of the square except the bottom, ie along the
x axis]. Lets call solution of that problem in the lecture notes u1 (x,y), so

X
u1 (x, y) = cn sin(nx)sinh[n( y)],
n=1

where R
2
0 f (x)sin(nx)dx
cn = .
sinh(n)
Exchanging x with y we get the solution of our problem here, which I will call u2 (x,y):

X
u2 (x, y) = dn sin(ny)sinh[n( x)],
n=1

where R
2
0 g(y)sin(ny)dy
dn = .
sinh(n)
[If you dont like this symmetry argument, you can do this problem by just following the
steps in the lecture notes, making the necessary changes at each step.]
For the second part of the problem, I claim that the answer is simply u(x,y) u1 (x,y)
+ u2 (x,y). Since each of these is a solution, the linearity of the Laplace equation ensures
that their sum is likewise a solution. But now we have to check the boundary conditions
to see that they agree with the ones we want. [It might be helpful here to draw two
squares, marking on the edges of each one the values of u1 (x,y) and u2 (x,y) there, so you
can take the sums easily.]

u(x, 0) = u1 (x, 0) + u2 (x, 0) = f (x) + 0 = f (x);

u(x, ) = u1 (x, ) + u2 (x, ) = 0 + 0 = 0;


u(, y) = u1 (, y) + u2 (, y) = 0 + 0 = 0;
u(0, y) = u1 (0, y) + u2 (0, y) = 0 + g(y) = g(y),
all of which are correct. So this is indeed the solution.