Exercise Sheet 6
Exercise Sheet 6
1.1.˚ Define the interior of A, denoted by Å, as the collection of interior all points of A. Note
that, it is clear from the definition that A is open if and only if Å “ A.
(a) Show that Å is an open subset of X contained in A.
(b) Show that if an open subset U of X is contained in A then U Ň Å.
(c) Show that, Å is the union of all open subsets of X that are contained in A.
The derived set of A, denoted by A1 , is defined to be the collection of all limit points of A. We
def
define the closure of A to be the set A “ A Y A1 . Thus, it is clear that x P A if and only if, for
every r ą 0, Bpx; rq X A ‰ H. It is immediate from the definition that A is closed if and only if
A “ A.
1.4.˚ Let x P X and r ą 0. Show that Bpx; rq Ň Brx; rs. Give an example where the inclusion
is proper.
1.6.˚ Let U and F be open and closed subsets of R respectively. Show that, there exists A Ň R
such that Å “ U and A “ F .
` ˘
Hint. Does U Y pBF q Y pQ X F qzU work?
1.7. Suppose further that X is a normed linear space and d has been induced by the norm
on X. We say that A Ň X is said to be convex if @x, y P A and α P r0, 1s, one has
p1 ´ αqx ` αy P A. Show that, Å and A are both convex whenever A is convex.
def
1.8. Recall that, for A, B Ň X, dpA, Bq “ inftdpa, bq : a P A, b P Bu. When B “ txu, we simply
write dpx, Aq. Show the following:
(a) @r ą 0, tx P X : dpx, Aq ă ru and tx P X : dpx, Aq ą ru are open subsets of X.
(b) @r ą 0, tx P X : dpx, Aq ď ru and tx P X : dpx, Aq ě ru are closed subsets of X.
(c) x P X is an exterior point of A, i.e., an interior point of XzA, if and only if dpx, Aq ą 0.
(d) x P A ðñ dpx, Aq “ 0 ðñ dpx, Aq “ 0.
(e) Can 1.8.d be generalised to dpA, Bq “ 0 ðñ A X B ‰ H?
1.10.˚ Show that if x ‰ y P X, there exists open subsets U and V of X such that U X V “ H.
1.11.˚ Prove that disjoint closed sets in X can be enlarged to disjoint open sets.
1.12.˚ Suppose that X is infinite. Show that X contains an open subset U such that both U and
XzU are infinite.
1.14. Show that, if U is open for every open subset U of X, then X has no limit point.
1.15.˚ Let U be an open subset of R. For every x P U , consider the union Ix of all open intervals
I Ň U such that x P I.
(a) Show that Ix is an open interval contained in U .
(b) Show that, for x, y P U , either Ix “ Iy or Ix X Iy “ H.
(c) Conclude from 1.15.a and 1.15.b that U is the disjoint union of distinct Ix ’s.
(d) Show that, the union mentioned in 1.15.c is countable.
Exercise Sheet 6 (MTH 301) 3
(e) Further to the above, show that, if tJn u8 n“1 is a some mutually disjoint countable
collection of open intervals such that
ď8
U“ Jn , (1.1)
n“1
1.22.˚ Show that every closed infinite subset of Rn , where n P N, is the closure of a countable
subset of Rn .
4 Exercise Sheet 6 (MTH 301)
Hint. Let A Ň Rn be closed and infinite. Denote the collection of all open balls Bpx, rq
such that x P Qn and r P Qą0 by B. Note that B is countable. For each ball B P B such
that B X A ‰ H, we choose a point from B X A ‰ H. Denote the collection of those chosen
points by A1 . Clearly A1 is a countable subset of A, and hence A1 Ň A. Now suppose a P A
and r ą 0. Then there exists B P B with radius ă r such that a P B (why?). Thus we
obtain a point in A1 whose distance from a is atmost 2r.
Hint. Observe that,` if txn u8 ˘ is 1Cauchy, then one can always choose a subsequence
n“1
8
txnk uk“1 such that d xnk`1 , xnk ă 2k , for all k P N.
2.2. Let pX, } ¨ }q be a normed linear space over K “ R or C. For a sequence txn u8 n“1 in X,
8
ÿ
def
consider sn “ x1 ` ¨ ¨ ¨ ` xn , for all n P N. We say the series xn is convergent if tsn u8
n“1
n“1
8
ÿ 8
ÿ
is convergent. In that case, we denote lim sn by xn . xn is said to be absolutely
nÑ8
n“1 n“1
8
ÿ
convergent if }xn } ă 8. Show that the following are equivalent:
n“1
(S.1) X is complete with respect to the metric induced by } ¨ }.
(S.2) Every absolutely convergent series is convergent.
Hint. To show (S.1) ùñ (S.2), just mimic the proof for series of real or complex numbers.
2.1. can be used to show the converse.
2.3.˚ The aim of this exercise is to show that none of the hypotheses of the contraction principle
can be dropped. Give an example of a map T : X ÝÑ X without a fixed point in each of
the following cases:
(a) X is not a complete metric space but T is a contraction.
(b) X is a complete metric space but d pT x, T yq ă dpx, yq for all x, y P X.
2.4. Prove the following extension of the contraction principle: let X be a complete metric space
def
and there exists n P N such that T n “ T ˝ ¨ ¨ ¨ ˝ T (n times) is a contraction. Then show
that T has a unique fixed point.
Remark. It may happen sometime that while the given map T does not contract, some
iterate actually does. Even in such cases, 2.4. yields a unique fixed point.
def
Hint. Suppose x P X is the unique fixed point of T n . Let y “ T x. What is T n y?
2.5.˚ Show that if the assumption that diam Fn ÝÝÝÑ 0 is dropped from the hypothesis of
nÑ8
Cantor’s intersection theorem, then the intersection can be empty.
Exercise Sheet 6 (MTH 301) 5
2.6.˚ Suppose that a metric space X satisfied the following property: whenever tFn u8 n“1 is a
sequence of closed subsets in X such that Fn`1 Ň Fn , for all n P N, and diam Fn ÝÝÝÑ 0,
Ş8 nÑ8
n“1 Fn ‰ H. Show that X must be complete.
2.7.˚ Let pX, } ¨ }q be a normed linear space over K “ R or C. Show that the following are
equivalent:
(N.1) X is complete with respect to the metric induced by } ¨ }.
If tBn u8
(N.2) Ş n“1 is a collection of closed balls in X such that Bn`1 Ň Bn , for all n P N, then
8
n“1 Bn ‰ H.
Remark. Note that, if Bn ’s are closed balls, the condition diam Bn ÝÝÝÑ 0 is not required,
nÑ8
provided X is a compete normed linear space. But 8
Ş
B
n“1 n can be empty if X is just a
complete metric space (see 2.8.).
Hint. Adopt the strategy of 2.6. to prove (N.2) ùñ (N.1). To see the converse, let
tBrxn , rn su8
n“1 be a sequence of closed balls in X such that Brxn`1 , rn`1 s Ň Brxn , rn s,
for all n P N. Then clearly trn u8n“1 is decreasing. Observe that, for all n P N, one has
}xn`1 ´ xn } ď rn ´ rn`1 (why?). This shows that txn u8n“1 is Cauchy. Now show that, if
xn ÝÝÝÑ x, then x P Brxn`1 , rn`1 s, for all n P N.
nÑ8
2.10.˚ Let X be a complete metric space and U Ň X be open. Then show that the conclusion of
Baire category theorem holds for U .
Hint. All the balls in U we need to consider can be made balls in X.
6 Exercise Sheet 6 (MTH 301)
2.11. Let X be a complete metric space and tFn u8 n“1 be a sequence of closed subsets of X. Show
that if each Fn has empty interior, then so does their union.
2.13. Let X and Y be metric spaces and f : X ÝÑ Y . Assume that X is complete. Suppose Ť8 that
there exists a sequence tAn u8
n“1 of nowhere dense subsets of X such that Dpf q “ n“1 An .
Then show that there is a dense subset of X on which f is continuous.
2.14.˚ Let X be a complete metric space and A Ň X be dense. Assume further that A Ň X 1 .
Does there exist a function f : X ÝÑ R such that lim |f pxq| “ `8, for every a P A?
xÑa
Ť8
Hint. Observe that, X “ n“1 tx P X : |f pxq| ă nu. Then some open ball B must be
contained in the closure of some tx P X : |f pxq| ă N u (why?). Pick a P B X A. Then
there must exist an r ą 0 such that Bpa; rq Ň B and @x P Bpa; rqztau, |f pxq| ě N . Since
pBpa; rqztauqXA ‰ H (why?), it follows that, there exists a1 P B XA such that |f pa1 q| ě N .
Now show that a1 does not lie in the closure of tx P X : |f pxq| ă N u.
2.15. Let X and Y be two metric spaces. Denote by BpX, Y q the set of all bounded maps from
X to Y . For f, g P BpX, Y q, define
def
dpf, gq “ sup d pf pxq, gpxqq . (2.2)
xPX
2.16. Let X “ ra, bs, where a ă b P R, and Y be as in 2.15.. We denote the set of all continuous
maps from ra, bs to Y by C pra, bs, Y q. For 1 ď p ă 8, define
ˆż b ˙ p1
def
dpf, gq “ dpf ptq, gptqqp dt . (2.3)
a
(a) Why does integral appearing in the right hand side of (2.3) make sense?
Hint. You may use 3.1..
(b) Show that (2.3) defines a metric on Cpra, bs, Y q. Is this metric space complete?
2.17.˚ Fix α P p0, 1s. Let C 0,α ra, bs be the collection of functions f : ra, bs ÝÑ R such that
|f pxq ´ f pyq|
sup ă 8.
0ďx‰yď1 |x ´ y|α
def |f pxq ´ f pyq|
For f P C 0,α ra, bs, let }f }0,α “ |f paq| ` sup . Show that } ¨ }0,α defines a
0ďx‰yď1 |x ´ y|α
norm on C 0,α ra, bs. Is C 0,α ra, bs complete with respect to the metric induced by this norm?
0,α
Hint. Observe that tfn u8
n“1 is Cauchy in C ra, bs, then tfn u8
n“1 is Cauchy in Cra, bs,
}¨}8
where the latter is equipped with the sup metric. Suppose fn ÝÝÝÑ f . Show that f satisfies
nÑ8
(2.17.) so that f P C 0,α ra, bs. Finally show that tfn u8
n“1 converges to f in C
0,α
ra, bs.
Exercise Sheet 6 (MTH 301) 7
3. Continuous functions
3.1. Let pX, dq be a metric space. Assume that X ˆ X is equipped with the following metric:
def
d8 ppx1 , y1 q, px2 , y2 qq “ maxtdpx1 , x2 q, dpy1 , y2 qu.
Show that d : X ˆ X ÝÑ R is continuous.
3.2. Let pV, x , yq be an inner product space over K “ R or C. Show that the following is a
continuous function:
V ˆ V ÝÑ K, pv, wq ÞÑ xv, wy.
3.3. Let X and Y be metric spaces and f : X ÝÑ Y . Show that the following are equivalent:
(a) f is continuous.
` ˘
(b) @A Ň X. f A Ň f pAq.
` ˘
(c) @B Ň Y , f ´1 pBq Ň f ´1 B .
3.5.˚ (a) Let X, Y and Z be metric spaces and f : X ÝÑ Y and g : Y ÝÑ Z. Show that if f is
continuous at x0 P X and g is continuous at f px0 q, then g ˝ f is continuous at x0 .
(b) What will be an appropriate analogue of 3.5.a for limits?
3.8.˚ Let C 1 ra, bs denote the set of all continuously differentiable functions on ra, bs. Assume
that C 1 ra, bs is equipped with the sup norm. Is the map C 1 ra, bs ÝÑ Cra, bs, f ÞÑ f 1
continuous?
def
Hint. Consider fn pxq “ sin?nx
n
, for all x P r0, 1s. Recall from 3.20. (f ) of Exercise Sheet 4
x
that lim sup cos nx “ 1, when 2π R Q.
nÑ8
3.9.˚ Let X and Y be metric spaces and f : X ÝÑ Y . Assume that X has no isolated point and
lim f ptq exists for any x P X. Consider the following function:
tÑx
def
gpxq “ lim f ptq, @x P X.
tÑx
Show that g is continuous.
3.10. Let X be a metric space and A Ň X. Find a necessary and sufficient condition on A so
that χA , i.e., the characteristic function of A, is continuous.
3.11. Let X and Y be metric spaces and f : X ÝÑ Y . Show that f is continuous if and only if
the graph of f , i.e., the set tpx, f pxqq : x P Xu is closed in X ˆ Y .
3.12. Let X be a metric space. Show that X does not have a limit point if and only if for every
metric space Y and every map f : X ÝÑ Y , f is continuous.
3.13.˚ Let pa, bq be an open bounded interval in R. Give an example of a bijective continuous
function f : pa, bq ÝÑ R such that f ´1 is also continuous.
` ˘
Hint. If pa, bq “ ´ π2 , π2 , can you find such an f ?
3.14. Let X and Y be metric spaces and f, g : X ÝÑ Y be continuous. Show the following:
(a) tx P X : f pxq ‰ gpxqu is open.
(b) tx P X : f pxq “ gpxqu is closed.
3.16. Let X be a metric space and γ, σ : r0, 1s ÝÑ X be continuous. Assume that γp1q “ σp0q.
Define γ ˚ σ as follows:
#
def γp2tq, if 0 ď t ď 12 ;
γ ˚ σptq “
σp2t ´ 1q, if 21 ď t ď 1.
Show that γ ˚ σ is continuous.
3.17. Show that if every real valued continuous function on a metric space is bounded then every
real valued continuous function must have points of maximum and minimum.
3.18.˚ Let A and B be two disjoint closed subsets of a metric space X. Show that there exists a
continuous f : X ÝÑ r0, 1s satisfying the following properties simultaneously:
Exercise Sheet 6 (MTH 301) 9
3.20. Suppose that F is a family of real valued functions on a complete metric space X. Assume
that, for every x P X, there exists Mx ą 0 such that |f pxq| ă Mx , for all f P F . Show
that, there exists an open subset U of X and M ą 0 such that |f pxq| ă M , for all f P F
and x P U .
def
Hint. For n P N, consider Fn “ tx P X : |f pxq| ď n, @f P F u. Use Baire category
theorem.
4. Uniform continuity
4.1.˚ Let X and Y be metric spaces and f : X ÝÑ Y . Show that the following are equivalent:
(a) f is uniformly continuous.
(b) If txn u8 8
n“1 and tyn un“1 are sequences in X such that dpxn , yn q ÝÝÝÑ 0, then
nÑ8
4.1. is very useful while showing some function is not uniformly continuous. Typical examples are
ex , x sin x and sin px2 q etc.
4.4.˚ Let X and Y be normed linear spaces over K “ R or C, and T : X ÝÑ Y be linear. Show
that the following are equivalent:
(i) T is continuous at 0.
(ii) T is continuous.
(iii) DM ą 0 such that, for all x P X, }T x} ď M }x}.
(iv) T is Lipschitz.
(v) T is uniformly continuous.
Hint. (ii) ùñ (iii) is the only nontrivial part. As T is continuous at› 0,ˆthere˙›exists δ ą 0
›δ x ››
such that }x} ă δ ùñ }T x} ă 1. Observe that, for any given x P X, ›› ă δ.
2 }x} ›
4.5. For any set X, BpXq denotes the set of all bounded real valued functions on X. Assume that
BpXq is equipped with the sup norm. Show that each of the following maps is Lipschitz?
(a) BpXq ÝÑ R, f ÞÑ f px0 q, where x0 P X is fixed.
(b) BpXq ÝÑ ℓ8 , f ÞÑ tf pxn qu8 8
n“1 , where txn un“1 Ň X is fixed, and ℓ8 is equipped with
the sup norm.
(c) What if one considers Cra, bs, equipped with the } ¨ }1 norm, in 4.5.a and 4.5.b?
żb
(d) Cra, bs ÝÑ R, f ÞÑ f ptq dt, where Cra, bs is equipped with the sup norm.
a
(e) What if one considers } ¨ }1 in 4.5.d?
(f) ℓ1 ÝÑ ℓ1 , txn u8 8 8
n“1 ÞÑ txn yn un“1 , where tyn un“1 is fixed.
4.6.˚ (a) Suppose that f : R ÝÑ R is continuous. Assume that lim f pxq exist in R. Then
xÑ˘8
show that f is uniformly continuous.
(b) Show that if f : ra, 8q ÝÑ R is continuous and lim f pxq exist in R, then f is uniformly
xÑ8
continuous.
4.9. Let f : r0, 8q ÝÑ R be uniformly continuous. Assume that, for all x ě 0 f px ` nq ÝÝÝÑ 0.
nÑ8
Show that lim f pxq “ 0.
xÑ8
Hint. Let ε ą 0. Since f is uniformly continuous on r0, 1s, there exists N P N such that
for all x, y P r0, 1s with |x ´ y| ă N1 , one has |f pxq ´ f pyq| ă ε. Consider the partition of
r0, 1s into N equal parts. Let x1 ă x2 ă ¨ ¨ ¨ ă xN be the points of subdivision. Then there
Exercise Sheet 6 (MTH 301) 11
exists n0 P N such that, for all n ě n0 , |f pxi ` nq| ă ε, for all i ´ 1, . . . , N . For any x ě n0 ,
there exists xi such that |x ´ rxs ´ xi | ă N1 . (why?)
4.10. Let f : r1, 8q ÝÑ R be uniformly continuous. Show that there exists M ą 0 such that
|f pxq|
x
ď M , for all x ě 1.
Hint. There exists N δ ą 0 such that |f pxq ´ f pyq| ă 1 whenever x, y ě 1 and |x ´ y| ď δ.
By repeated application of triangle inequality, we obtain
|f pxq| ď |f p1q| ` |f pxq ´ f p1q| ď f p1q ` pn ` 1q,
where x ´ 1 “ nδ ` r, n P Zě0 and 0 ď r ă δ.