You are on page 1of 4

Midterm Solutions

1. Prove that if (xn ) of real numbers is convergent then (|xn |) is also


convergent. Give an example to show that the convergence of (|xn |)
need not imply the convergence of (xn ).
Solution. Since (xn ) is convergent, given  > 0, there exists N N
such that |xn x| < , for all n N .
From the triangle inequality, we know that ||xn | |x|| |xn x|.
Consequently, we have that ||xn | |x|| < , for all n N , which proves
the convergence of (|xn |).
The sequence ((1)n ) is non-convergent, but converges absolutely to 1.
2. Let X = (xn ) be a sequence of positive real numbers such that
 
xn+1
lim = L > 1.
xn
Show that X is not a bounded sequence and hence is not convergent.
 
Solution. Since lim xxn+1 n
= L, given  > 0, there exists N N such

that xxn+1 L < , for all n N . This would imply that for n N ,

n
xn+1
xn
> L  = r, where r > 1 for sufficiently small .
Therefore, for n N , we obtain
xn+1 > rxn > r2 xn1 > . . . > xN rnN +1 .
By setting C = xN /rN , we get xn+1 > Crn+1 for all n N . Since
r > 1, this sequence will be unbounded, and hence diverges.
3. If 0 < r < 1 and |xn+1 xn | < rn for all n N, show that (xn ) is a
Cauchy sequence.
Solution. For n > m, we have that
n n
X X
i1 rm rn
|xn xm | |xi xi1 | < r = .
i=m+1 i=m+1
1r

Since 0 < r < 1, lim(rn ) = 0, and hence (rn ) is Cauchy. Therefore,


given  > 0, there exists N N such that for all m, n N , we have
|rm rn | < (1 r).

1
Hence for all m, n N , we have that |xn xm | < , showing that (xn )
is Cauchy.

4. Let (fn ) C[0, 1] be such that there exists M > 0 such that kfn k
M , for all n N. Define
Z x
Fn (x) = fn (t) dt.
0

Show that Fn has a uniformly convergent subsequence.


Solution. From the Fundamental Theorem of Calculus, we have that
Fn0 (x) = fn (x), and hence kFn0 k = kfn k M .
By applying the Mean Value Theorem to the interval [x, y] [0, 1], we
obtain a c (x, y) such that Fn (y) Fn (x) = Fn0 (c)(y x). This would
imply that

|Fn (y) Fn (x)| kFn0 k |y x| = M |y x|. (1)

Therefore, given  > 0, if we choose = /M , we have that |Fn (y)


Fn (x)| < , whenever |xy| < . This shows that the set {Fn : n N}
is equicontinuous.
Moreover, putting x = 0 in Equation (1), we obtain |Fn (y)| M |y|
M , and taking supremum over all y [0, 1], we get that kFn k M ,
that is, Fn is bounded for all n N.
From the Corollary to the Ascoli-Arzela Theorem, we conclude that
(Fn ) has a uniformly convergent subsequence.
R1 R1
5. Let lim(fn ) = f in C[0, 1]. Then show that lim 0 fn (x) dx = 0 f (x) dx.
Solution. Since lim(fn ) = f , given  > 0 and x [0, 1], there exists
N = N (x, ) N such that |fn (x) f (x)| < , for all n N .
Therefore, for all n N , we have that
Z 1 Z 1

(fn (x) f (x)) dx |fn (x) f (x)| dx <  .

0 0

R1 R1
This shows that lim 0
fn (x) dx = 0
f (x) dx.

2
6. Let f : (0, 1) R given by f (x) = 1/x cannot be approximated
uniformly by polynomials on (0, 1). In other words, show that given
 > 0 there exists no polynomial p(x) on (0, 1) such that kp f k < .
Solution. Suppose that given  > 0, there exists a polynomial p such
that kf pk < . Then this would imply that kf k < kpk + ,
showing that f is bounded in (0, 1). But this is a contradiction to the
fact that 1/x is unbouded (0, 1).
7. Consider the metric space

X
`2 = {(xn ) : xi R and |xn |2 < }
n=1

with the usual metric. Consider the sequence (en ) `2 given by


en = (0, 0, . . . , 0, 1, 0, 0, . . .)
(ie. en has 1 in the nth position and zero elsewhere). Show that (en ) is
a bounded sequence, but does not have a convergent subsequence.
Solution. Let (an ) be a sequence of sequences in `2 . We will denote
the ith term of the nth sequence an by an (i). For the sequence to be
bounded, we need to have a M > 0 such that kxn k2 < M , that is,
P 2
i=1 |an (i)| < , for all n N. Since ken k2 = 1 for all n, we have
that the sequence (en ) is bounded.
Suppose that (ank ) is a convergent
subesequence of (an ). Then we
can see that kank anm k2 = 2, for all k, m N, which shows that
the subsequence will not be Cauchy. Hence no subsequence can be
convergent.
8. (Bonus) Consider the space C[0, 1], and define
Z 1
kf k1 = |f (t)| dt
0

Show that k k1 is a norm on C[0, 1].


Solution.
(i) Since |(f (t)| R0, we have that kf k1 0. Moreover, since f is
1
continuous and 0 |f (t)| dt = 0, which would imply that f (t) = 0,
for all t [0, 1]. (This is an exercise in calculus.)

3
(ii) For c R and f C[0, 1],
Z 1 Z 1
kcf k1 = |cf (t)| dt = |c| |f (t)| dt = |c|kf k1 .
0 0

(iii) For f, g C[0, 1], we have that


Z 1 Z 1 Z 1
kf +gk1 = |f (t)+g(t)| dt |f (t)| dt+ |g(t)| dt = kf k1 +kgk1 .
0 0 0

You might also like