Norms and Metrics, Normed Vector Spaces and Metric Spaces
Norms and Metrics, Normed Vector Spaces and Metric Spaces
Norms and Metrics, Normed Vector Spaces and Metric Spaces
) := |x x
|.
Remark: The distance function d : R R to R
+
has the following properties:
(D1) d(x, x
) 0 ;
(D2) d(x, x
) = 0 x = x
;
(D3) d(x, x
) = d(x
, x) ;
(D4) d(x, x
) d(x, x
) + d(x
, x
x
2
1
+ x
2
2
+ + x
2
n
.
The Euclidean distance between two points x, x
R
n
is
d(x, x
) := ||x x
||.
Remark: The Euclidean norm function || || : R
n
R
+
has the properties (N1) - (N4); the
Euclidean distance function d : R
n
R
n
R
+
has the properties (D1) - (D4).
Denition: Let V be a vector space. A function || || : V R
+
is a norm on V if it satises
(N1) - (N4). A vector space together with a norm is called a normed vector space.
Denition: Let X be a set. A metric on X is a function d : X X R
+
that satises (D1) -
(D4). The pair (X, d) is called a metric space.
Remark: If || || is a norm on a vector space V , then the function d : V V R
+
dened by
d(x, x
) := ||x x
|| is a metric on V
In other words, a normed vector space is automatically a metric space, by dening the metric in
terms of the norm in the natural way. But a metric space may have no algebraic (vector) structure
i.e., it may not be a vector space so the concept of a metric space is a generalization of the
concept of a normed vector space.
In each of the following examples you should verify that d is a metric by verifying that it satises
each of the four conditions (D1) to (D4). For the norms on R
n
, you should draw the set of all
points in R
2
that satisfy ||x|| = 1.
Example 1: In R
n
dene ||x||
:= max{|x
1
|, . . . , |x
n
|} and dene d : R
n
R
n
R
+
as
d(x, x
) = ||x x
||
= max{|x
1
x
1
|, . . . , |x
n
x
n
|} . This is called the max-norm or sup-norm.
Example 2: In R
n
dene ||x||
1
:=
n
i=1
|x
i
| and d(x, x
) = ||x x
||
1
=
n
i=1
|x
i
x
i
| . This is
called the city block, Manhattan, or taxicab norm.
Example 3: Let X = C[0, 1] = {f : [0, 1] R | f is continuous } . Dene || || : C[0, 1] R by
||f|| := max{|f(x)| | x [0, 1]} . Note that this is well-dened as a consequence of the Weierstrass
Theorem, which says that a continuous real function on a closed interval attains a maximum. We
will introduce the Weierstrass Theorem formally a little later in the course, and generalize it to
compact sets in metric spaces. Also note that C[0, 1] is a vector space under natural denitions of
vector addition and scalar multiplication:
(f + g)(x) := f(x) + g(x) and (f)(x) := f(x).
Exercise: Show that the set C[0, 1] with the operations of vector addition and scalar multipli-
cation dened in Example 3 is a vector space.
2
Example 4: Let X = {a, b} or any other nite set. Dene d : X X R
+
as follows:
d(x, x
) :=
0 if x = x
1 otherwise .
Then d is a metric on X. This is an example of a metric space that is not a normed vector space:
there is no way to dene vector addition or scalar multiplication for a nite set.
Example 5: Let p be a real number satisfying p 1, and dene ||x||
p
on R
n
+
by
x
p
:= (|x
1
|
p
+ |x
2
|
p
+ + |x
n
|
p
)
1/p
.
This is called the
p
norm. Note that the Euclidean norm is the
2
-norm, the city block norm is
the
1
-norm, and the sup-norm is the
-norm.
Remark: If (X, d) is a metric space and S is a subset of X, then (S, d) is a metric space.
Example 6: Let V be a normed vector space for example, R
2
with the Euclidean norm. Let
C be the unit circle {x V | ||x|| = 1}. This is another example of a metric space that is not
a normed vector space: V is a metric space, using the metric dened from || ||, and therefore,
according to the above remark, so is C; but C is not a vector space, so it is not a normed vector
space. C could be replaced here by any subset of V that is not a vector subspace of V i.e., any
subset thats not closed under vector addition and scalar multiplication.
3