You are on page 1of 2

Homework 3

Exercise 3.1. Prove or disprove the following property: if a metric space (X , d ) has at least two elements,
then it admits an open subset which is neither X nor the empty set ∅.
Solution. Suppose x ∈ X . Because X has at least two elements, the complement X r {x } is nonempty. Suppose
y ∈ X r {x } and let r = d (x, y). Then certainly x < Bd (y, r ) because d (x, y) ≮ r . We conclude that X r {x } is
open in X , and X r {x } , X as well as X r {x } , ∅.
Exercise 3.2. Let X = (X , d X ), Y = (Y , dY ), Z = (Z , d Z ) and W = (W , dW ) be metric spaces. Prove that if
f:X Z and д : Y W are each continuous, then

f ×д: X × Y Z ×W

sending (x, y) ( f (x ), д(y)) is continuous, with respect to the product metrics on X × Y and Z × W .
Solution. Suppose that (x 0 , y0 ) ∈ X × Y and ε > 0. By the continuity of f and д there are δ 1 , δ 2 > 0 such that
f (BdX (x 0 , δ 1 )) ⊂ BdZ ( f (x 0 ), ε/2) and д(BdY (y0 , δ 2 )) ⊂ BdW (д(y0 ), ε/2). Let δ = min{δ 1 , δ 2 }.
Suppose that (x, y) ∈ BdX ×Y ((x 0 , y0 ), δ ), that is, d X ×Y ((x 0 , y0 ), (x, y)) < δ . This means that we have
d X (x 0 , x ) + dY (y0 , y) < δ . But then d X (x 0 , x ) ≤ d X (x 0 , x ) + dY (y0 , y) < δ ≤ δ 1 so that x ∈ BdX (x 0 , δ 1 ),
and similarly y ∈ BdY (y0 , δ 2 ). By our choice of δ 1 and δ 2 this implies that d Z ( f (x 0 ), f (x )) < ε/2 and
dW (д(y0 ), д(y)) < ε/2. Combining these we find

d Z ×W ( f × д(x 0 , y0 ), f × д(x, y)) = d Z ( f (x 0 ), f (x )) + dW (д(y0 ), д(y)) < ε/2 + ε/2 = ε.

So f × д(x, y) ∈ BdZ ×W ( f × д(x 0 , y0 ), ε ). We conclude that f × д is continuous with respect to the product
metrics.
Exercise 3.3. The goal of this exercise is to prove continuity of some of the standard algebraic operations on
R, thought of as maps R2 R; we will focus on addition

+ : R2 R
(x, y) x +y

and multiplication

· : R2 R
(x, y) xy.

Use the standard Euclidean metric d (x, y) = |x − y| on R and the “taxicab” metric on R2 given by

d Ta ((x 1 , x 2 ), (y1 , y2 )) = |x 1 − y1 | + |x 2 − y2 |;

this metric is equivalent to the usual Euclidean metric, so anything you prove about these operations being
continuous for d Ta holds for d Eu as well.
(i) Show that addition is continuous.
(ii) Show that multiplication is continuous.
(iii) Conclude a result stated in class: that if f , д : X R are two continuous functions from a metric
space to R, then f + д and f д are continuous.
Solution.
(i) Suppose that (x 0 , y0 ) ∈ R2 and ε > 0. Then, if d ((x 0 , y0 ), (x, y)) = |x 0 − x | + |y0 − y| < ε, we will have

d (x 0 + y0 , x + y) = |x 0 + y0 − x − y| ≤ |x 0 − x | + |y0 − y| < ε.

So we conclude that + is continuous.

1
(ii) Suppose that (x 0 , y0 ) ∈ R2 and ε > 0. Let δ = min{ε/(|x 0 | + |y0 | + 1), 1} and suppose that (x, y) ∈ R2
satisfies d ((x 0 , y0 ), (x, y)) = |x 0 − x | + |y0 − y| < δ . Then
d (x 0y0 , xy) = |x 0y0 − xy| = |x 0 (y0 − y) + (x 0 − x )y0 − (x 0 − x )(y0 − y)| ≤
≤ |x 0 ||y0 − y| + |x 0 − x ||y0 | + |x 0 − x ||y0 − y| <
< |x 0 |δ + |y0 |δ + δ 2 =
= (|x 0 | + |y0 | + δ )δ ≤
ε
≤ (|x 0 | + |y0 | + 1) = ε.
|x 0 | + |y0 | + 1
Hence, we can conclude that · : R2 R is continuous.
(iii) First consider the map ∆ : X X × X with ∆(x ) = (x, x ). We claim that ∆ is continuous. To see this,
let x ∈ X and ε > 0. If y ∈ X with d (x, y) < ε/2, then we find
d ((x, x ), (y, y)) = d (x, y) + d (x, y) < 2ε/2 = ε.
Now, let f , д : X R be continuous. Then, by what we have seen in part (i) and (ii) and exercise 2,
the functions + ◦ ( f × д) ◦ ∆ and {·} ◦ ( f × д) ◦ ∆ are compositions of continous functions. Note that
+ ◦ ( f × д) ◦ ∆(x ) = +( f × д(x, x )) = +( f (x ), д(x )) = f (x ) + д(x ) = ( f + д)(x )
and similarly · ◦ ( f × д) ◦ ∆ = f д. So we conclude that f + д and f д are compositions of continuous
functions and therefore continuous themselves.
Exercise 3.4. Let X = (X , d ) and Y = (Y , d 0 ) be metric spaces and f : X Y a map. By a theorem we
stated in class on Friday and proved on Monday, f is continuous if and only if the preimage under f of any
open (respectively closed) set is open (respectively closed). This suggests an easy way to give many new
examples of open and closed sets in a metric space M: write down a function f : M R, show that it is
continuous, and then take the preimage under f of an open or closed set in R respectively. Using this method:
(i) Fix real positive numbers a 1 , . . . , ak +1 > 0. Show that the generalized ellipsoid
 k+1
X 
Ea1, ...,ak +1 = (x 1 , . . . , x k +1 ) : ai x i2 = 1
i=1

is a closed subset of Rk+1 .


(ii) Let V = {(x 1 , x 2 , x 3 ) : x 12 + x 2 < 1 and (x 3 − 2) 4 > 4}. Show that V is an open subset of R3 with its
standard Euclidean metric.
Solution.
(i) Consider the function f : Rk+1 R with f (x 1 , . . . , x k+1 ) = i ai x i2 . By repeated appliation of
P
exercise 3 part (iii) this function is continuous. Note that {1} ⊂ R is closed, since any singleton in a
metric space is closed. Then Ea1, ...,ak +1 = f −1 ({1}) is a preimage of a closed set under a continuous
functions and therefore closed itself.
(ii) First consider the function f : R3 R with f (x 1 , x 2 , x 3 ) = x 12 + x 2 . Again, by exercise 3 this a
continuous function. Since (−∞, 1) ⊂ R is an open set, its preimage
f −1 ((−∞, 1)) = {(x 1 , x 2 , x 3 ) : x 12 + x 2 < 1}
is an open subset of R3 . Similarly, the function д : R3 R with д(x 1 , x 2 , x 3 ) = (x 3 − 2) 2 is continuous
and (4, ∞) ⊂ R is open. Therefore
д−1 ((4, ∞)) = {(x 1 , x 2 , x 3 ) : (x 3 − 2) 2 > 4}
is open. But then V = f −1 ((−∞, 1)) ∩ д−1 ((4, ∞)) is a finite intersection of open sets and hence open
itself.

You might also like