Elementary theory of metric spaces

Theorem 1 (Schwarz inequality) Let x = (x1 , ..., xN ) and y = (y1 , ..., yN ) be elements of RN . Then: X X 1 X 1 2 2 N N N 2 2 x y x y . i i i=1 i i i=1 i=1 PN
i=1 (yi

Proof. Dene f : R R, f () =

f () =

Note that since the equation depends on and not x or y , we can rearrange the equation as: f () = a2 b + c and nd the that minimizes f . There are two possible cases: PN (a) a = 0. If a = 0 then i=1 x2 i = 0, which is equal to having xi = 0 i. Then, X X 1 X 1 2 2 N N N 2 2 xi yi i=1 xi yi i=1 i=1 X X 1 X 1 2 2 N N N 2 0 yi i=1 0 yi i=1 i=1 00

XN

xi )2 . Note that f () 0 . Also, XN

i=1

i=1

2 yi 2

xi yi + 2

XN

i=1

x2 i.

And the expression is satised. (b) a 6= 0. The value of that minimizes f satises f 0 ( ) = 0. Then, f 0 () = 2a b 2a b = 0 b = 2a It follows that f ( ) is: f ( ) = a = = b 2a 2 b 2a

+c

ab2 b2 +c 4a2 2a b2 b2 +c 4a 2a = b2 +c 4a

b b +4ac 0; therefore, b2 4ac. Recalling the values of a, b, But since f () 0 then 4 a + c 0 and 4a and c: X X 2 X N N N 2 2 2 xi yi 4 xi yi i=1 i=1 i=1

Taking square roots, and noting that x2 = |x|,

There are two useful derivations of this result.

X X 1 X 1 2 2 N N N 2 2 x y x y . i i i i i=1 i=1 i=1

Corollary 2 (Triangle inequality) Let x = (x1 , ..., xN ) and y = (y1 , ..., yN ) be elements of RN . Then: r r r XN XN XN 2 2 2. (xi + yi ) xi + yi
i=1 i=1 i=1

Proof. Note that:

P P 1 P 1 N N N 2 2 2 2 From Theorem 1 we know that i=1 xi yi = . Substituting: i=1 xi i=1 yi XN

i=1

XN

i=1

(xi + yi )2 =

XN

i=1

x2 i +2

XN

i=1

xi yi +

XN

i=1

2 yi .

(xi + yi )2 = x2 i +2

XN

i=1

= since PN

i=1

P xi yi is just a particular (and valid) case of N i=1 xi yi . Taking square roots: r XN

i=1

r XN

XN
N i=1

i=1

x2 i +2

x2 i

1 X 2 N

XN

i=1

xi yi + 1 2

XN

i=1

2 yi

i=1

x2 i

i=1

x2 i +

r XN

y2 i=1 i

!2

XN

i=1

2 yi

(xi + yi )

r XN

x2 i=1 i

r XN

y2 . i=1 i

Proof. Set xi = ai bi , yi = bi ci . Then, by Corollary 2, r r r XN XN XN 2 2 ((ai bi ) + (bi ci )) (ai bi ) + (bi ci )2 .

i=1 i=1 i=1

Corollary 3 Let a = (a1 , ..., aN ), b = (b1 , ..., bN ), c = (c1 , ..., cN ) be elements of RN . Then, r r r XN XN XN 2 2 (ai ci ) (ai bi ) + (bi ci )2 .
i=1 i=1 i=1

Denition 4 (Cartesian product) Let S and T be sets. The cartesian product of S and T , denoted as S T , is the set of all ordered pairs (p, q ) in which p belongs to S and q belongs to T : S T = {(p, q )|p S, q T } . Remark 5 Almost generally, S T 6= T S . Example 6 Let S = {1, 2} and T = {1, 3}. Then S T = {(1, 1), (1, 3), (2, 1), (2, 3)}. However, the product T S is dierent, since T S = {(1, 1), (1, 2), (3, 1), (3, 2)}. Denition 7 The cartesian product of any nite number of sets S1 , ..., SM is the set of ordered N-tuples (p1 , ..., pN ) in which pi Si , i = {1, ..., N }. This is denoted by: ON
i=1

Si .