MA2223 – Tutorial solutions

Part 1. Metric spaces

p
T1–1. Show that the function d(x, y) = |x − y| defines a metric on R.

The given function is symmetric and non-negative with d(x, y) = 0 if and only if x = y.
It remains to check that the triangle inequality holds. This is the case because
p p
d(x, y) = |x − y| ≤ |x − z| + |z − y|
p p
≤ |x − z| + |z − y| = d(x, z) + d(z, y).

T1–2. Show that d(x, y) = |x − y|3 does not define a metric on R.

The given function is symmetric and non-negative with d(x, y) = 0 if and only if x = y.
However, it does not satisfy the triangle inequality because

d(1, 3) = 23 > 13 + 13 = d(1, 2) + d(2, 3).

T1–3. Compute the distances d1 (f, g) and d∞ (f, g) when f, g ∈ C[0, 1] are the functions
defined by f (x) = x and g(x) = x3 .

One has x3 ≤ x for all 0 ≤ x ≤ 1 and this implies that

Z 1 Z 1
3 1 1 1
d1 (f, g) = |x − x | dx = (x − x3 ) dx = − = .
0 0 2 4 4

To compute the distance d∞ (f, g), we need to compute the maximum value of

h(x) = |f (x) − g(x)| = x − x3 , 0 ≤ x ≤ 1.

Since h(0) = h(1) = 0 and h′ (x) = 1 − 3x2 , it easily follows that

√
2 2 3
d∞ (f, g) = max h(x) = h(3−1/2 ) = 3−1/2 (1 − 3−1 ) = √ = .
0≤x≤1 3 3 9
 T1–4. Compute the distance d∞ (x2 , x + 2) in the space C[0, 2].

We need to compute the maximum value of the function

h(x) = |x2 − x − 2|, 0 ≤ x ≤ 2.

Since x2 − x − 2 = (x − 2)(x + 1) ≤ 0 for each 0 ≤ x ≤ 2, we have

h(x) = |x2 − x − 2| = x + 2 − x2 =⇒ h′ (x) = 1 − 2x.

This means that h(x) is increasing for x < 1/2 and decreasing for x > 1/2, so

d∞ (x2 , x + 2) = max h(x) = h(1/2) = 9/4.

0≤x≤2

T1–5. Show that d1 (x1/n , 1) → 0 as n → ∞ in the space C[0, 1]. Does the same
statement hold in the case of the d∞ metric?

Since x1/n ≤ 1 for each 0 ≤ x ≤ 1, the first distance is given by

Z 1
1/n 1
d1 (x , 1) = (1 − x1/n ) dx = 1 −
0 1/n + 1

and it converges to 1 − 1 = 0 as n → ∞. On the other hand, the second distance is

d∞ (x1/n , 1) = max (1 − x1/n ) = 1.

0≤x≤1

T1–6. Show that the d∞ metric in R2 is the limit of the dp metric by showing that
1/p
limp→∞ |x|p + |y|p

= max(|x|, |y|) for all x, y ∈ R.

Suppose first that |y| < |x|. Then x is nonzero and | xy | < 1, so

1/p h y p i1/p
lim |x|p + |y|p

= lim |x| · 1 + = |x|.
p→∞ p→∞ x
This settles the case |y| < |x| and the case |x| < |y| is similar. When |x| = |y|, we have
1/p 1/p
lim |x|p + |y|p = lim 2|x|p = lim 21/p |x| = |x|.
 
p→∞ p→∞ p→∞
 T2–1. Let (X, d) be a metric space. Given a point x ∈ X and a real number r > 0,
show that A = {y ∈ X : d(x, y) > r} is open in X.

Let y ∈ A be given and note that ε = d(x, y) − r is positive. We claim that B(y, ε) is
contained entirely within A. In fact, one has

z ∈ B(y, ε) =⇒ d(z, y) < ε

=⇒ d(z, y) + r < ε + r = d(x, y) ≤ d(x, z) + d(z, y)
=⇒ r < d(x, z)
=⇒ z ∈ A.

This shows that B(y, ε) is contained entirely within A and so A is open.

T2–2. Show that A = {(x, y) ∈ R2 : x2 + y 2 < 2y} is open in R2 .

Completing the square, one may express the given set in the form

A = (x, y) ∈ R2 : x2 + y 2 − 2y < 0


= (x, y) ∈ R2 : x2 + (y − 1)2 < 1 = B((0, 1), 1).



In particular, A is open in R2 because every open ball is open in R2 .

T2–3. Show that A = {x ∈ R : x3 + 2x2 − 3x ≤ 0} is closed in R.

It suffices to show that the complement of A is open in R. Since

x3 + 2x2 − 3x = x(x2 + 2x − 3) = x(x + 3)(x − 1),

one has x3 + 2x2 − 3x > 0 if and only if x ∈ (−3, 0) ∪ (1, ∞). Thus, the complement of A is
the union of two open sets, so the complement of A is open and A is closed.

T2–4. Consider the set X = A ∪ B, where A = (0, 1) and B = [2, 3). Show that A, B
are both open in X and thus closed in X as well.

The set A is open because it is the open ball B(1/2, 1/2). Similarly, the set B is open
because it is the open ball B(5/2, 1). Since the sets A and B are the complements of one
another, the fact that they are both open implies that they are both closed as well.

T2–5. Find a sequence of closed intervals C1 ⊂ C2 ⊂ C3 ⊂ · · · such that their union is

an open interval.

Consider the closed intervals Cn = [1/n, 3 − 1/n], for instance. Since the left endpoint is
decreasing to 0 and the right one is increasing to 3, the union of the intervals is (0, 3).
 T2–6. Let (X, d) be a metric space whose metric d is discrete. What can you say about
a sequence {xn } which is convergent in X?

Suppose the sequence converges to x. Then there exists an integer N > 0 such that

d(xn , x) < 1 for all n ≥ N .

In particular, xn = x for all n ≥ N , so the sequence is eventually constant.

T3–1. Suppose (X, d) is a metric space and f : X → R is continuous. Show that

A = {x ∈ X : |f (x)| < r} is open in X for each r > 0.

The given set can be expressed in the form

A = {x ∈ X : −r < f (x) < r} = f −1 (−r, r).

Since (−r, r) is open in R, its inverse image must then be open in X by continuity.

T3–2. Show that every function f : X → Y is continuous when X, Y are metric spaces
and the metric on X is discrete.

To show that f is continuous at x, let ε > 0 be given and take δ = 1. Then

dX (x, y) < δ =⇒ x=y =⇒ dY (f (x), f (y)) = dY (f (x), f (x)) = 0 < ε.

T3–3. Suppose f : X → Y is a constant function between metric spaces, say f (x) = y0

for all x ∈ X. Show that f is continuous.

To show that f is continuous at x, let ε > 0 be given and δ > 0 be arbitrary. Then

dX (x, y) < δ =⇒ dY (f (x), f (y)) = dY (y0 , y0 ) = 0 < ε.

T3–4. Show that f (x) = cos(sin(2x)) is Lipschitz continuous on [0, 1].

This follows immediately by Theorem 1.8 because f is differentiable and

|f ′ (x)| = | sin(sin(2x))| · | cos(2x)| · 2 ≤ 2.

 x
T3–5. Show that fn (x) = x2 +n2
converges uniformly on [0, ∞).

The given sequence converges pointwise to the zero function f (x) = 0. To show that the
convergence is uniform, one needs to show that

sup |fn (x) − f (x)| = sup fn (x)

x≥0 x≥0

x
goes to zero as n → ∞. Consider the function fn (x) = x2 +n2
on [0, ∞). Since

x2 + n2 − 2x2 n2 − x 2
fn′ (x) = = ,
(x2 + n2 )2 (x2 + n2 )2

this function is increasing when x < n and decreasing when x > n, so its maximum value is
n 1
sup |fn (x) − f (x)| = sup fn (x) = fn (n) = = .
x≥0 x≥0 2n 2 2n

This expression does go to zero as n → ∞, so the convergence is uniform, indeed.

x
T3–6. Show that fn (x) = x+n
does not converge uniformly on [0, ∞).

The given sequence converges pointwise to the zero function f (x) = 0. To show that the
convergence is not uniform, we consider the expression
x
sup |fn (x) − f (x)| = sup fn (x) = sup .
x≥0 x≥0 x≥0 x+n

As one can easily check, fn′ (x) = n/(x + n)2 is positive, so fn (x) is increasing and
x x
sup |fn (x) − f (x)| = sup = lim = 1.
x≥0 x≥0 x+n x→∞ x+n

This expression does not go to zero as n → ∞, so the convergence is not uniform.

T4–1. Show that the set A = {(x, y) ∈ R2 : y > 0} is not complete.

It suffices to find a sequence of points in A such that their limit is not in A. There are
obviously lots of examples, but the simplest one is probably

(xn , yn ) = (0, 1/n) ∈ A, lim (xn , yn ) = (0, 0) ∈

/ A.
n→∞
 T4–2. Show that (X, d) is complete, if the metric d is discrete.

Let {xn } be a Cauchy sequence in X. Then there exists an integer N > 0 such that

d(xm , xn ) < 1 for all m, n ≥ N .

Since the metric d is discrete, this actually gives xm = xn for all m, n ≥ N . Thus,

xm = xN for all m ≥ N

and the given Cauchy sequence converges to the point xN ∈ X.

T4–3. Let (X, d) be a metric space and suppose A, B ⊂ X are complete. Show that
the union A ∪ B is complete as well.

Suppose {xn } is a Cauchy sequence of points in A ∪ B. There must be infinitely many

terms that lie in A or else infinitely many terms that lie in B. Consider the former case, as
the other case is similar. Since a subsequence lies in A and A is complete, this subsequence
must converge to a point x ∈ A. In view of Theorem 1.13, the original sequence must also
converge to x, so it converges to a point x ∈ A ∪ B.

T4–4. Show that f (x) = ln(x + 2) has a unique fixed point in [0, 2].

The function f is increasing, so it maps the interval [0, 2] into the interval

[f (0), f (2)] = [ln 2, ln 4] ⊂ [ln 1, ln e2 ] = [0, 2].

Since [0, 2] is closed in R, it is complete by Theorem 1.16. To show that f is a contraction,

we use the mean value theorem to write

|f (x) − f (y)| = |f ′ (c)| · |x − y|

for some point c between x and y. Since f ′ (x) = 1/(x + 2), this actually gives
1 1
|f (x) − f (y)| = · |x − y| ≤ · |x − y|.
c+2 2
In particular, f is a contraction and the result follows by Banach’s fixed point theorem.
 T4–5. Show that f (x) = x3 is a contraction on X = (0, 1/2), but it has no fixed point
in X. Does this contradict Banach’s theorem?

The function f is increasing, so it maps the interval (0, 1/2) into the interval

(0, 1/8) ⊂ (0, 1/2).

It does not have any fixed points because x3 = x implies that x = 0, ±1 and none of these
points lies in X. To show that f is a contraction, we use the mean value theorem to write

|f (x) − f (y)| = |f ′ (c)| · |x − y|

for some point c between x and y. Since f ′ (x) = 3x2 , we conclude that
3
|f (x) − f (y)| = 3c2 · |x − y| ≤ · |x − y|.
4
This example does not contradict Banach’s theorem because X = (0, 1/2) is not complete.

T4–6. Suppose that f : X → Y is a continuous function and {xn } is a Cauchy sequence

in X. Does {f (xn )} have to be Cauchy as well?

No. For instance, f (x) = 1/x is continuous on (0, ∞) and xn = 1/n is Cauchy, but the
sequence f (xn ) = n is not Cauchy because f (xn ) − f (xn−1 ) = 1 for all n.