Continuity
Matt Rosenzweig
Contents
1 Continuity
1.1 Rudin Chapter 4 Exercises .
1.1.1 Exercise 1 . . . . . .
1.1.2 Exercise 2 . . . . . .
1.1.3 Exercise 3 . . . . . .
1.1.4 Exercise 4 . . . . . .
1.1.5 Exercise 5 . . . . . .
1.1.6 Exercise 11 . . . . .
1.1.7 Exercise 12 . . . . .
1.1.8 Exercise 14 . . . . .
1.1.9 Exercise 15 . . . . .
1.1.10 Exercise 20 . . . . .
1.1.11 Exercise 21 . . . . .
1.1.12 Exercise 22 . . . . .
1.1.13 Exercise 23 . . . . .
1.1.14 Exercise 24 . . . . .
1
1.1
1.1.1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
1
1
1
2
2
2
2
3
3
3
3
4
4
5
5
Continuity
Rudin Chapter 4 Exercises
Exercise 1
If f : R R satisfies
lim [f (x + h) f (x h)] = 0,
h0
x R
then f need not be continuous.
Proof. Consider f : R R defined by
f (x) :=
1
0
|x| > 0
x=0
It is clear that f has a discontinuity at x = 0, but is continuous at every x 6= 0. The limit statement follows from
continuity for all x 6= 0. If x = 0, then
f (x + h) f (x h) = f (h) f (h) = 1 1 = 0,
1.1.2
h 6= 0
Exercise 2
If f : (X, d) (Y, ) is a continuous mapping of metric spaces, then
f (E) f (E),
E X
Moreover, the inclusion can be proper.
Proof. Let y = f (x) f (E), where x E. Let > 0 be given, and consider the open neighborhood B(y; ). I
claim that B(y; ) f (E) 6= . Since f is continuous at y, there exists > 0 such that
d(x, x0 ) < = (f (x), f (x0 )) <
By definition of closure, there exists x0 E B(x; ), so that
(f (x), f (x0 )) = (y, f (x0 )) <
But f (x0 ) f (E), which implies that B(y; )f (E) 6= . Since > 0 was arbitrary, we conclude that y f (E).
1
1.1.3
Exercise 3
Let f : (X, d) R be a continuous real-valued function on a metric space (X, d). Define the zero set of f to be
Z(f ) := {p X : f (p) = 0}
Then Z(f ) is closed.
Proof. Let (pn )
n=1 be a sequence in Z(f ) such that pn p X. Since f is continuous, we have that
f (p) = lim f (pn ) = 0 = p Z(f )
n
1.1.4
Exercise 4
Let f, g : (X, d) (Y, ) be continuous mappings, and let E be a dense subset of X. Then f (E) is dense in
f (X). Furthermore, if g(p) = f (p) for all p E, then g(p) = f (p) for all p X.
Proof. Let y = f (x) f (X), and let > 0 be given. Since f is continuous, there exists > 0 such that
d(x, x0 ) < = (y, f (x0 ) <
Since E is dense in X, there exists x0 E B(x; ), so that (y, f (x0 )) < and therefore B(y; ) f (E) 6= .
Suppose g(p) = f (p) for all p E. Let > 0 be given. Since f (E) = g(E) are dense in f (X) and g(X), there
exists q E such that
(f (p), f (q)) <
,
2
(g(p), g(q)) <
2
Since f (q) = g(q), we conclude from the triangle inequality that
(f (p), g(p)) (f (p), f (q)) + (f (q), g(q)) + (g(q), g(p)) = (f (p), f (q)) + (g(q), g(p)) <
+ =
2 2
Since > 0 was arbitrary, we conclude that (f (p), g(p)) = 0 f (p) = g(p).
1.1.5
Exercise 5
If
1.1.6
Exercise 11
Suppose f : (X, d) (Y, ) is uniformly continuous. Then (f (xn ))
n=1 is a Cauchy sequence in Y for every
Cauchy sequence (xn )
n=1 in X. Let > 0 be given. Choose > 0 such that
x, x0 X, d(x, x0 ) < = (f (x0 ), f (x)) <
and choose a positive integer N such that
n, m N = d(xn , xm ) <
Then by uniform continuity,
n, m N = (f (xn ), f (xm ) < ,
If f is a uniformly continuous real function defined on a dense subset E X. Then f has a unique continuous
extension from E to X.
Proof. Uniqueness follows from Exercise 4. For existence, first fix x X. By the density of E, there exists a
sequence (xn )
n=1 in E such that xn x. Then (xn ) is Cauchy and hence (f (xn ))n=1 is a Cauchy sequence in
R by our previous result. By completeness, f (xn ) converges to some limit y. Define
f (x) := y = lim f (xn )
n
We need to show that this definition is well-defined, i.e. does not depend on our choice of sequence (xn ). If
0
(x0n )
n=1 is another sequence in E such that xn x. Then by uniform continuity, for any > 0,
|f (x0n ) f (xn )| < ,
for all n sufficiently large, which implies that f (x0n ) f (x). To see that the defined function f : X R is
continuous, first choose > 0 such that p, q E with d(p, q) < implies that |f (p) f (q)| < 3 . Choose p, q E
such that max {|f (x) f (p)| , |f (y) f (q)|} < 3 . Then
x, y X, d(x, y) < = |f (x) f (y)| |f (x) f (p)| + |f (p) f (q)| + |f (q) f (y)|
< + +
3 3 3
=
1.1.7
Exercise 12
Let f : (X, dX ) (Y, dY ) and g : (Y, dY ) (Z, dZ ) be uniformly continuous functions between metric spaces.
Then h := g f : (X, dX ) (Z, dZ ) is uniformly continuous.
Proof. Let > 0 be given. Choose 1 > 0 such that
y, y 0 Y, dY (y, y 0 ) < 1 = dZ (f (y), f (y 0 )) <
and choose 2 > 0 such that
x, x0 X, dX (x, x0 ) < 2 = dY (f (x), f (x0 )) < 1 = dZ (g(f (x)), g(f (x0 ))) <
We conclude that h is uniformly continuous.
1.1.8
Exercise 14
Let I := [0, 1] be the closed unit interval, and suppose that f : I I is continuous. Then f has at least one
fixed point in I.
Proof. Consider the continuous function g : I R defined by g(x) := f (x) x. If f has no fixed points, then
g(x) 6= 0 for all x I. By the intermediate value theorem, we must have g(x) > 0 for all x I or g(x) < 0 for
all x I. By hypothesis that f (I) I, g(0) > 0 and g(1) = f (1) 1 < 0, which results in a contradiction.
1.1.9
Exercise 15
Every continuous open mapping f : R R is monotonic.
Proof. Let x, y R with x < y.
1.1.10
Exercise 20
Let E be a nonempty subset of a metric space (X, d). Define a function E : X [0, ) by
E (x) := inf d(x, z)
zE
E (x) = 0 if and only if x E.
Proof. Observe that
E (x) = 0 > 0, z E s.t.d(x, z) < > 0, B(x; ) E 6= x E
E is a uniformly continuous function on X.
Proof. To show that E is uniformly continuous, it suffices to show that
|E (x) E (y)| d(x, y)
Let > 0 be small. By the definition of E and the triangle inequality, we obtain the inequalities
E (x) d(x, z) d(x, y) + d(z, y), E (y) d(y, z) d(x, y) + d(x, z)
z E, x, y X
Taking the infimum over all z E. We see that E (x) d(x, y) + E (y) and E (y) d(x, y) + E (x), which
shows that
d(x, y) E (x) E (y) d(x, y) |E (x) E (y)| d(x, y)
1.1.11
Exercise 21
Let K, F be disjoint subsets of a metric space (X, d), with K compact and F closed. There exists > 0 such
that
p K, q F = d(p, q) >
If neither K nor F are compact, then the conclusion may fail.
Proof. Since K, F are disjoint and closed, it follows from Exercise 20 that F (p) > 0 for all p K and F is
continuous on K. Since K is compact, by Weierstrass extreme value theorem, F attains its minimum at some
p0 K. Hence,
0 < F (p0 ) F (p) = inf d(p, q) d(p, q),
qF
p K, q F
Choosing = F (p0 ) completes the proof.
To see that the conlusion may fail if neither Knor F are compact, letour metric space be the rationals
equipped with the Euclidean metric. Then K := ( 2, ) and F := (, 2) are closed in Q, but neitheris
compact. However, inf pK,qF d(p, q) = 0 since points in K and F become arbitrarily close as they approach 2
from the right and left, respectively.
1.1.12
Exercise 22
Let A and B be disjoint nonempty closed sets in a metric space (X, d), and define a function f : X R by
f (p) :=
A (p)
,
A (p) + B (p)
pX
Then f is a continuous function with f (X) [0, 1], f 1 ({0}) = A and f 1 ({1}) = B.
Proof. Recall from Exercise 20 that A (p) = 0 and B (p) = 0 if and only if p A = A and p B = B,
respectively. Since A and B are disjoint, we see that
A (p) = 0 = B (p) > 0,
B (p) = 0 = A (p) > 0
Thus,
f (p) = 0 A (p) = 0 p A,
f (p) = 1 B (p) = 0 p B
For any p X \ A,
0
A (p)
A (p)
=1
A (p) + B (p)
A (p)
which shows that f (X) [0, 1]. f is continuous since the it is the composition of the continuous functions A
and B .
Note that this exercise establishes a converse to Exercise 3: namely, every closed subset of a metric space is
the zero set of some continuous function.
Since the image of f is contained in [0, 1], we have that
1
1
V := f 1 ((, )) = f 1 ([0, ),
2
2
1
1
W := f 1 (( , +)) = f 1 (( , 1])
2
2
where V and W are open, since f is continuous. Hence, A and B are separated by disjoint open sets.
4
1.1.13
Exercise 23
For Exercises 23 and 24, see my post Continuity, Convexity, and Jensens Inequality.
1.1.14
Exercise 24
