Solutions to Final Exam Review Problems

Math 5C, Winter 2007

n=0
(x/4)
n
=

n=0
(1)
n
4
n+1
x
n
. Since the
innite series is geometric, with ratio x/4, it converges for | x/4| < 1, and
thus for |x| < 4. Thus, the radius of convergence is 4.
(Note: it is also possible to use Taylors formula, c
n
= f
(n)
(0)/n!, to determine
the coecients of the Maclaurin series and then use the ratio test to compute the
radius of convergence.)
(b) Find the Taylor series for f(x) centered at x = 1, and compute its radius of
convergence.
Solution.
f(x) =
1
5 + (x 1)
=
1
5(1 (x 1)/5)
=
1
5

n=0
(x 1)
5
=

n=0
(1)
n
5
n+1
(x 1)
n
.
Since the innite series is geometric with ratio (x 1)/5, it converges for |
(x 1)/5| < 1, and thus for |x 1| < 5. So the radius of convergence is 5.
(Note: Again, we could have used Taylors formula c
n
= f
(n)
(1)/n!, and the ratio
test to nd the radius of convergence.)
(c) Find the Taylor series for g(x) = ln(4 + x) centered at x = 1, and compute its
radius of convergence.
Solution. Since g(x) =
_
f(x) dx, we integrate the Taylor series from part (b):
g(x) =
_

n=0
(1)
n
5
n+1
(x 1)
n
=

n=0
(1)
n
5
n+1
(n + 1)
(x 1)
n+1
+ C.
To nd C, we plug in x = 1: C = g(1) = ln 5. Thus, replacing n + 1 by k,
g(x) = ln 5 +

k=1
(1)
k1
5
k
k
(x 1)
k
. The radius of convergence is the same as the
series we integrated, and thus equals 5.
1
2. For what values of x does the series

n=1
x
n
n
x
converge? (Bonus: Does it converge
uniformly on this entire set?)
Solution. We use the ratio test:
lim
n

x
n+1
/(n + 1)
x
x
n
/n
x

= lim
n

x
(1 + 1/n)
x

= |x/1
x
|
= |x| < 1
So it converges for x in the interval (1, 1), but we still need to check whether it con-
verges at the endpoints of this interval: x = 1. If x = 1, the series becomes

n=1
1
n
which diverges by the p-series test. If x = 1, the series becomes

n=1
(1)
n
n which
diverges by the nth term test, since lim
n
(1)
n
n = 0. Thus the series converges only
for x in the interval (1, 1).
3. Show that the series of functions

n=1
ne
nx
converges uniformly on [1/2, ).
Solution. We use the M-test, where u
n
(x) = ne
nx
. Since x 1/2, and each u
n
(x) is
a positive-valued decreasing function, |u
n
(x)| = u
n
(x) u
n
(1/2) = ne
n/2
. Thus, we
let M
n
= ne
n/2
. To show that the series converges uniformly, we must now check that
the series

n=1
M
n
=

n=1
n/e
n/2
converges. We can check this by the ratio test:
lim
n

(n + 1)/e
(n+1)/2
n/e
n/2

= lim
n
1 + 1/n
e
1/2
= 1/

e < 1,
meaning that the series converges by the ratio test.
4. Let f(x) =
_
1, if /2 x /2
1, if x < /2 or /2 < x
(a) Find the Fourier series for f(x), and sketch its graph.
Solution. Notice that f(x) is an even function (its graph is symmetric about
the y-axis). Thus all the b
n
s will equal 0, and we only need to compute the a
n
coecients. We have
a
n
=
1

f(x) cos(nx) dx
2
=
1

_
_
/2

cos(nx) dx +
_
/2
/2
cos(nx) dx +
_

/2
cos(nx) dx
_
=
1

sin(nx)
n
_
/2

+
sin(nx)
n
_
/2
/2
+
sin(nx)
n
_

/2
_
=
1

_
4 sin(n/2)
n
_
=
_
0, n even
4(1)
(n1)/2
/n, n odd
The above is correct for n > 0 only. When n = 0, the three integrals in the second
line above evaluate to /2, , /2, respectively. Thus a
0
= 0, and the Fourier
series is
f(x)

n=1
a
n
cos(nx) =

n=1, n odd
4(1)
(n1)/2
n
cos(nx)
=

k=1
4(1)
k1
(2k 1)
cos(2k 1)x
=
4

_
cos x
1
3
cos(3x) +
1
5
cos(5x)
1
7
cos(7x) +
_
.
(Note: Any of the last 3 expressions would be a correct answer.) Its graph consists
of horizontal segments at y = 1 over the intervals ((2k 1/2), (2k + 1/2)) and
at y = 1 over the intervals ((2k + 1/2), (2k + 3/2)) for each integer k, and it
has points on the x-axis at the endpoints of each of these intervals.
(b) Use part (a) (or other methods) to nd the Fourier series for
g(x) =
_
_
_
x, if x < /2
x, if /2 x /2
x if /2 < x
Solution. Notice that g

(x) = f(x). Thus we can obtain the Fourier series for

k=1
4(1)
k1
(2k 1)
cos(2k 1)x dx
= C +

k=1
4(1)
k1
(2k 1)
2

sin(2k 1)x
= C +
4

_
sin x
1
3
2
sin(3x) +
1
5
2
sin(5x)
1
7
2
sin(7x) +
_
.
3
Since g(x) is an odd function, all the a
n
coecients, including a
0
, equal 0. Thus
C = a
0
/2 = 0, and removing the Cs from the above yields the correct Fourier
series.
(c) Show that 1
1
3
+
1
5

1
7
+ =

4
. (You should practice using one of the Fourier
series above, or else one from lecture, rather than a power series.)
Solution. Plug x = 0 into the Fourier series from (a) to get f(0) =
4

(1 1/3 +
1/5 1/7 + ). Since f(0) = 1, the sum 1 1/3 + 1/5 1/7 + converges to
/4.
5. Find the Fourier series of f(x) = e
x
. (Suggestion: use the complex form of the Fourier
series.)
Solution. The complex form of the Fourier series is f(x)

n=
c
n
e
inx
, where the
coecients c
n
are given by the formula c
n
=
1
2
_

f(x)e
inx
dx. Here,
c
n
=
1
2
_

e
(1in)x
dx
=
e
(1in)x
2(1 in)
_

=
e
(1in)
e
(1in)
2(1 in)
=
e

(cos(n) + i sin(n)) e

(cos(n) + i sin(n))
2(1 in)
=
(1)
n
e

(1)
n
e

2(1 in)
=
(1)
n
(e

)(1 + in)
2(n
2
+ 1)
=
(1)
n
(e

)
2(n
2
+ 1)
+ i
(1)
n
(e

)n
2(n
2
+ 1)
Thus
e
x

n=
_
(1)
n
(e

)
2(n
2
+ 1)
+ i
(1)
n
(e

)n
2(n
2
+ 1)
_
e
inx
.
Since c
n
= (a
n
ib
n
)/2 for n > 0, and c
0
= a
0
/2, we see that a
n
equals twice the real
part of c
n
for each n 0, and b
n
equals (-1) times twice the imaginary part of c
n
for
n > 0. Thus, in terms of sines and cosines,
e
x

_
1
2
+

n=1
_
(1)
n
n
2
+ 1
cos(nx) +
(1)
n+1
n
n
2
+ 1
sin(nx)
_
_
.
4
6. Find the solution of the wave equation

2
u
t
2
4

u
x
2
= 0 with initial displacement
given by f(x) = 0, and initial velocity given by g(x) = sin
2
x. (Hint: use a half-angle
formula!) Sketch the solution when t = /2.
Solution. Due to an error on my part, the Hint is useless. We have to solve the wave
equation with a = 2, f(x) = 0, and g(x) = sin
2
x. The solution has the form
u(x, t) =

n=1
sin(nx)[
n
sin(2nt) +
n
cos(2nt)]
where 2n
n
are the Fourier sine coecients of g(x) = sin
2
x and
n
are the Fourier
sine coecients of f(x) = 0. Thus all the
n
are 0, and the
n
s are computed as in
lecture (see the lecture notes pp. 66-7):

n
=
1
2n
_
2

_

0
sin
2
x sin(nx) dx
_
=
_
0, n even
4
n
2
(n
2
4)
, n odd
Hence, the solution as a Fourier series is
u(x, t) =

n=1, odd
4
n
2
(n
2
4)
sin(nx) sin(2nt)
=

k=1
4
(2k 1)
2
((2k 1)
2
4)
sin((2k 1)x) sin(2(2k 1)t).
When t = /2, sin(2nt) = sin(n) = 0. Thus u(x, /2) = 0, and the string is in
equilibrium position along the x-axis.
5