Module-5 Inner Product Spaces:

Dot products and inner products – the lengths and angles of

vectors – matrix representations of inner products- Gram-
Schmidt orthogonalisation

5.2
 n
5.1 Length and Dot Product in R
 Length :
The length of a vector v  (v1 , v2 ,, vn ) in Rn is given by

|| v ||  v12  v2 2   vn 2 ( || v || is a real number)

 Notes: The length of a vector is also called its norm

 Properties of length (or norm)

(1) v  0
(2) v  1  v is called a unit vector
(3) v  0 if and only if v  0
(4) cv  c v (proved in Theoerm 5.1)
5.3
 Ex 1:
(a) In R5, the length of v  (0 ,  2 , 1 , 4 ,  2) is given by

|| v ||  0 2  (2) 2  12  4 2  (2) 2  25  5

(b) In R3, the length of v  ( 2

17
, 2
17
, 3
17
) is given by

2 2 2
 2   2   3  17
|| v ||         1
 17   17   17  17

(If the length of v is 1, then v is a unit vector)

5.4
 n
 A standard unit vector in R : only one component of the vector is
1 and the others are 0 (thus the length of this vector must be 1)

R 2 : e1 , e 2   1, 0  ,  0,1

R 3 : e1 , e 2 , e3   1, 0, 0  ,  0,1, 0  ,  0, 0,1

R n : e1 , e 2 , , e n   1, 0, , 0  ,  0,1, , 0 , ,  0, 0, ,1

 Notes: Two nonzero vectors are parallel if u  cv

(1) c  0  u and v have the same direction
(2) c  0  u and v have the opposite directions

5.5
 Theorem 5.1: Length of a scalar multiple
Let v be a vector in Rn and c be a scalar. Then
|| cv ||  | c | || v ||
Pf:
v  (v1 , v2 ,  , vn )
 cv  (cv1 , cv2 ,  , cvn )
|| cv ||  || ( cv1 , cv2 ,  , cvn ) ||
 (cv1 ) 2  (cv2 ) 2    (cvn ) 2
2 2 2
 c 2 (v1  v2    vn )
2 2 2
 | c | v1  v2    vn
 | c | || v ||
5.6
 Theorem 5.2: How to find the unit vector in the direction of v
n v
If v is a nonzero vector in R , then the vector u 
|| v ||
has length 1 and has the same direction as v. This vector u
is called the unit vector in the direction of v
Pf:
1
v is nonzero  v  0  0
v
1
If u  v (u has the same direction as v)
v
|| cv ||  | c | || v ||
v 1
|| u ||   || v ||  1 (u has length 1)
|| v || || v ||
5.7
 Notes:
v
(1) The vector is called the unit vector in the direction of v
|| v ||
(2) The process of finding the unit vector in the direction of v
is called normalizing the vector v

5.8
 Ex 2: Finding a unit vector
Find the unit vector in the direction of v = (3, –1, 2), and verify
that this vector has length 1
Sol:

v  (3 ,  1 , 2)  v  32   1  22  14
2

v (3 ,  1 , 2) 1
   (3 ,  1 , 2)
|| v || 3  (1)  2
2 2 2
14
 3 1 2 
 , , 
 14 14 14 
2 2 2
 3   1   2  14
       14  1
 14   14   14 
v
 is a unit vector
v
5.9
 Distance between two vectors:
n
The distance between two vectors u and v in R is
d (u , v )  || u  v ||

 Properties of distance
(1) d (u , v )  0
(2) d (u , v)  0 if and only if u = v
(3) d (u , v )  d ( v , u) (commutative property (交換律) of the distance function)
5.10
 Ex 3: Finding the distance between two vectors
The distance between u = (0, 2, 2) and v = (2, 0, 1) is

d (u , v )  ||u  v||  ||(0  2 , 2  0 , 2  1)||

 ( 2) 2  22  12  3

5.11
 Dot product in Rn:
The dot product of u  (u1 , u 2 ,  , u n ) and v  (v1 , v2 ,  , vn )
returns a scalar quantity
u  v  u1v1  u2v2   un vn (u  v is a real number)
(The dot product is defined as the sum of component-by-component
multiplications)

 Ex 4: Finding the dot product of two vectors

The dot product of u = (1, 2, 0, –3) and v = (3, –2, 4, 2) is
u  v  (1)(3)  (2)(2)  (0)(4)  (3)(2)  7

 Matrix Operations in Excel

SUMPRODUCT: calculate the dot product of two vectors
5.12
 Theorem 5.3: Properties of the dot product
If u, v, and w are vectors in Rn and c is a scalar,
then the following properties are true
(1) u  v  v  u (commutative property of the dot product)
(distributive property of the dot
(2) u  ( v  w )  u  v  u  w product over vector addition)
(3) c(u  v)  (cu)  v  u  (cv) (associative property of the scalar
multiplication and the dot product)
(4) v  v  || v || 2  v  v  0
(5) v  v  0 if and only if v  0 (straightforwardly true according to (4))

※ The proofs of the above properties simply follow the definition of the
dot product in Rn

5.13
 Euclidean n-space:
n
– In section 4.1, R was defined to be the set of all order n-
tuples of real numbers
– When Rn is combined with the standard operations of
vector addition, scalar multiplication, vector length,
and dot product, the resulting vector space is called
Euclidean n-space

5.14
 Ex 5: Find dot products

u  (2 ,  2) , v  (5 , 8), w  (4 , 3)
(a) u  v (b) (u  v)w (c) u  (2 v) (d) || w || 2 (e) u  ( v  2w)
Sol:
(a) u  v  (2)(5)  (2)(8)  6
(b) (u  v)w  6w  6(4 , 3)  (24 , 18)
(c) u  (2v)  2(u  v)  2(6)  12
(d) || w || 2  w  w  (4)(4)  (3)(3)  25
(e) v  2w  (5  (8) , 8  6)  (13 , 2)
u  ( v  2w)  (2)(13)  (2)(2)  26  4  22

5.15
 Ex 6: Using the properties of the dot product
Given u  u  39, u  v  3, v  v  79,
find (u  2 v)  (3u  v)

Sol:
(u  2v)  (3u  v)  u  (3u  v)  2v  (3u  v)
 u  (3u)  u  v  (2 v)  (3u)  (2v)  v
 3(u  u)  u  v  6( v  u)  2( v  v)
 3(u  u)  7(u  v)  2( v  v)
 3(39)  7(3)  2(79)  254

5.16
 Theorem 5.4: The Cauchy-Schwarz inequality
If u and v are vectors in Rn, then
| u  v |  || u || || v || ( | u  v | denotes the absolute value of u  v )
(The geometric interpretation for this inequality is shown on the next slide)

 Ex 7: An example of the Cauchy-Schwarz inequality

Verify the Cauchy-Schwarz inequality for u = (1, –1, 3)
and v = (2, 0, –1)
Sol:
u  v  1, u  u  11, v  v  5
 u  v  1  1
u v  u  u  v  v  11  5  55
 uv  u v
5.17
 Dot product and the angle between two vectors
To find the angle  (0     ) between two nonzero vectors
u = (u1, u2) and v = (v1, v2) in R2, the Law of Cosines can be
applied to the following triangle to obtain
2 2 2
v  u  v  u  2 v u cos
(The length of the subtense of θ can be expressed in terms of
the lengths of the adjacent sides and cos θ)

2
 v  u  (u1  v1 ) 2  (u2  v2 ) 2
2
v  v12  v22
2
u  u12  u22
u1v1  u2v2 uv
 cos   ※ You can employ the fact that |cos θ|  1 to
v u v u prove the Cauchy-Schwarz inequality in R2
5.18
 The angle between two nonzero vectors in Rn:
uv
cos  , 0  
|| u || || v ||
Opposite Same
u v  0 u v  0 u v  0 direction
direction

   
  
     0    0
2 2 2
cos   1 cos   0 cos   0 cos   0 cos   1

 Note:
The angle between the zero vector and another vector is
not defined (since the denominator cannot be zero)
5.19
 Ex 8: Finding the angle between two vectors
u  (4 , 0 , 2 ,  2) v  ( 2 , 0 ,  1 , 1)
Sol:
u  uu   4 2
 0  2   2  24
2 2 2

v  v  v  2  0   1  12  6
2 2 2

u  v  (4)(2)  (0)(0)  (2)(1)  (2)(1)  12

uv  12 12
 cos     1
|| u || || v || 24 6 144

u and v have opposite directions

(In fact, u = –2v and according to the
arguments on Slide 5.4, u and v are parallel
and with different directions) 5.20
 Orthogonal vectors:
n
Two vectors u and v in R are orthogonal (perpendicular) if
uv0
 Note:
The vector 0 is said to be orthogonal to every vector

5.21
 Ex 10: Finding orthogonal vectors
n
Determine all vectors in R that are orthogonal to u = (4, 2)
Sol:
u  (4 , 2) Let v  (v1 , v2 )
 u  v  (4 , 2)  (v1 , v2 )
 4v1  2v2
0
t
 v1  , v2  t
2
 t 
 v   ,t  , t  R
 2 

5.22
 Theorem 5.5: The triangle inequality
If u and v are vectors in Rn, then || u  v ||  || u ||  || v ||
Pf:
|| u  v || 2  (u  v)  (u  v)
 u  (u  v)  v  (u  v)  u  u  2(u  v)  v  v
 || u || 2 2(u  v) || v || 2  || u || 2 2 | u  v |  || v || 2 (c  |c|)
 || u || 2 2 || u || || v ||  || v || 2 (Cauchy-Schwarz inequality)
 (|| u ||  || v ||) 2
(The geometric representation of the triangle inequality:
|| u  v ||  || u ||  || v || for any triangle, the sum of the lengths of any two sides is
larger than the length of the third side (see the next slide))
 Note:
Equality occurs in the triangle inequality if and only if
the vectors u and v have the same direction (in this
situation, cos θ = 1 and thus u  v  u v  0) 5.23
 Theorem 5.6: The Pythagorean theorem
If u and v are vectors in Rn, then u and v are orthogonal
if and only if
(This is because u·v = 0 in the
|| u  v || || u ||  || v ||
2 2 2
proof for Theorem 5.5)

※ The geometric meaning: for any right triangle the sum of the squares of the lengths of
two legs equals the square of the length of the hypotenuse.

|| u  v ||  || u ||  || v || || u  v ||2  || u ||2  || v ||2 5.24

 Similarity between dot product and matrix multiplication:

 u1   v1 
u  v  (A vector u = (u1, u2,…, un) in Rn can be
u   2 v   2 represented as an n×1 column matrix)
 
   
u n  v n 

v1 
v 
u  v  uT v  [u1 u2 un ]  2   [u1v1  u2v2   un vn ]
 
 
vn 
(The result of the dot product of u and v is the same as the result
of the matrix multiplication of uT and v)

5.25
Keywords in Section 5.1:
 length:
 norm:
 unit vector:
 standard unit vector:
 Distance:
 dot product:
 Euclidean n-space:
 Cauchy-Schwarz inequality:
 angle:
 triangle inequality:
 Pythagorean theorem:

5.26
 5.2 Inner Product Spaces
 Inner product: represented by angle brackets〈 u , v〉
Let u, v, and w be vectors in a vector space V, and let c be
any scalar. An inner product on V is a function that associates
a real number〈 u , v〉 with each pair of vectors u and v and
satisfies the following axioms

(1)〈 u , v〉 〈 v , u〉 (commutative property of the inner product)

(distributive property of the inner product
(2)〈 u , v  w〉 〈 u , v〉 〈 u , w〉 over vector addition)
(3) c〈 u , v〉 〈 cu , v〉 (associative property
the inner product)
of the scalar multiplication and

(4) 〈 v , v〉  0
(5) 〈 v , v〉  0 if and only if v  0 (straightforwardly true according to (4))
5.27
 Note:
u  v  dot product (Euclidean inner product for R n )
 u , v  general inner product for a vector space V

 Note:
A vector space V with an inner product is called an inner
product space

Vector space: (V , , )
Inner product space: (V , ,  , , >)

5.28
 Ex 1: The Euclidean inner product for Rn
n
Show that the dot product in R satisfies the four axioms
of an inner product
Sol:
u  (u1 , u2 ,, un ) , v  (v1 , v2 ,, vn )
〈 u , v〉  u  v  u1v1  u2v2    un vn

By Theorem 5.3, this dot product satisfies the required four axioms.
n
Thus, the dot product can be a sort of inner product in R

5.29
 Ex 2: A different inner product for Rn
Show that the following function defines an inner product
on R2. Given u  (u1 , u2 ) and v  (v1 , v2 ),
〈 u , v〉  u1v1  2u 2 v 2
Sol:
(1) 〈 u , v〉  u1v1  2u2v2  v1u1  2v2u2 〈 v , u〉
(2) w  ( w1 , w2 )
〈 u , v  w〉  u1 (v1  w1 )  2u2 (v2  w2 )
 u1v1  u1w1  2u2v2  2u2 w2
 (u1v1  2u2v2 )  (u1w1  2u2 w2 )
〈 u , v〉 〈 u , w〉

5.30
 (3) c〈 u , v〉  c(u1v1  2u2v2 )  (cu1 )v1  2(cu2 )v2 〈 cu , v〉
(4) 〈 v , v〉  v12  2v22  0
(5) 〈 v , v〉  0  v12  2v22  0  v1  v2  0 ( v  0)

 Note: Example 2 can be generalized such that

〈 u , v〉  c1u1v1  c2u2v2   cnunvn , for all ci  0
can be an inner product on Rn

5.31
 Ex 3: A function that is not an inner product
Show that the following function is not an inner product on R3
〈 u , v〉  u1v1  2u2v2  u3v3
Sol:
Let v  (1 , 2 , 1)

Then 〈 v , v〉  (1)(1)  2(2)(2)  (1)(1)  6  0

Axiom 4 is not satisfied
Thus this function is not an inner product on R3

5.32
 Theorem 5.7: Properties of inner products
Let u, v, and w be vectors in an inner product space V, and
let c be any real number
(1)〈 0, v〉 〈 v, 0〉  0
(2)〈 u  v, w〉 〈 u, w〉 〈 v, w〉
(3)〈 u, cv〉  c〈 u, v〉
※ To prove these properties, you can use only the four axioms for
defining an inner product (see Slide 5.26)
Pf:
(3)
(1)〈 0, v〉 =〈 0u, v〉  0〈 u, v〉  0
(1) (2) (1)
(2)〈 u  v, w〉 〈 w, u  v〉 〈 w, u〉 +〈 w, v〉 〈 u, w〉 +〈 v, w〉
(1) (3)
(3) 〈 u, cv〉 〈 cv, u〉  〈c u, v〉

5.33
※ The definition of norm (or length), distance, angle, orthogonal, and
normalizing for general inner product spaces closely parallel to
those based on the dot product in Euclidean n-space

 Norm (length) of u:
|| u ||  〈 u , u〉

 Distance between u and v:

d (u , v) || u  v || u  v, u  v

 Angle between two nonzero vectors u and v:

〈 u , v〉
cos  , 0  
|| u || || v ||
 Orthogonal: (u  v)
u and v are orthogonal if 〈 u , v〉  0 5.34
 Normalizing vectors
(1) If || v ||  1 , then v is called a unit vector
(Note that v is defined as v, v )

(2) v  0 
v (the unit vector in the
Normalizing

v direction of v)
(if v is not a
zero vector)

5.35
 Ex 6: An inner product in the polynomial space
For p  a0  a1 x    an x n and q  b0  b1 x    bn x n ,
and p , q  a0b0  a1b1   anbn is an inner product

Let p( x)  1  2 x2 , q( x)  4  2 x  x 2 be polynomials in P2
(a) p , q  ? (b) || q || ? (c) d ( p , q)  ?
Sol:
(a) p , q  (1)(4)  (0)(2)  (2)(1)  2
(b) || q ||  q , q  42  (2)2  12  21
(c) p  q  3  2 x  3x 2
 d ( p , q)  || p  q ||   p  q, p  q
 (3)2  22  (3)2  22
5.36
 Properties of norm: (the same as the properties for the dot
n
product in R on Slide 5.2)
(1) || u ||  0
(2) || u ||  0 if and only if u  0
(3) || cu ||  | c | || u ||

 Properties of distance: (the same as the properties for the dot

product in Rn on Slide 5.9)
(1) d (u , v )  0
(2) d (u , v )  0 if and only if u  v
(3) d (u , v )  d ( v , u)

5.37
 Theorem 5.8：
Let u and v be vectors in an inner product space V
(1) Cauchy-Schwarz inequality:
〈| u , v〉 |  || u || || v || Theorem 5.4
(2) Triangle inequality:
|| u  v ||  || u ||  || v || Theorem 5.5
(3) Pythagorean theorem:
u and v are orthogonal if and only if

|| u  v || 2  || u || 2  || v || 2 Theorem 5.6

5.38
 Orthogonal projection: For the dot product function in Rn, we
define the orthogonal projection of u onto v to be projvu = av (a
scalar multiple of v), and the coefficient a can be derived as
follows
Consider a  0, av  a v  a v  u cos 
u
|| u || || v || cos  || u || || v || u  v uv
  
 v v || u || || v || v
v
projv u  av, a  0 uv uv uv
a   projv u  v
v
2
vv vv
 For inner product spaces:
Let u and v be two vectors in an inner product space V.
If v  0 , then the orthogonal projection of u onto v is
given by
u, v
projv u  v
 v, v 5.39
 Ex 10: Finding an orthogonal projection in R3
Use the Euclidean inner product in R3 to find the
orthogonal projection of u = (6, 2, 4) onto v = (1, 2, 0)

Sol:
 u , v  (6)(1)  (2)(2)  (4)(0)  10
 v , v  12  22  0 2  5
u , v uv 10
 projv u  v v  (1 , 2 , 0)  (2 , 4 , 0)
 v , v vv 5

5.40
5.41
 Theorem 5.9: Orthogonal projection and distance
Let u and v be two vectors in an inner product space V,
and if v ≠ 0, then
 u , v
d (u, projv u)  d (u, cv) , c 
 v , v

u u d (u, cv )
d (u, projv u)

v v
projv u cv

※ Theorem 5.9 can be inferred straightforwardly by the Pythagorean Theorem,

i.e., in a right triangle, the hypotenuse (斜邊) is longer than both legs (兩股)
5.42
Keywords in Section 5.2:
 inner product:
 inner product space:
 norm:
 distance:
 angle:
 orthogonal:
 unit vector:
 normalizing:
 Cauchy-Schwarz inequality:
 triangle inequality:
 Pythagorean theorem:
 orthogonal projection:

5.43
 5.3 Orthonormal Bases: Gram-Schmidt Process
 Orthogonal set :
A set S of vectors in an inner product space V is called an
orthogonal set if every pair of vectors in the set is orthogonal

S  v1 , v 2 ,  , v n   V
 v i , v j   0, for i  j

 Orthonormal set :
An orthogonal set in which each vector is a unit vector is
called orthonormal set
S   v1 , v 2 , , v n   V
 For i  j ,  vi , v j    vi , vi   vi 2
1

For i  j ,  vi , v j   0
5.44
 Note:
– If S is also a basis, then it is called an orthogonal basis or

an orthonormal basis
n
– The standard basis for R is orthonormal. For example,

S  (1,0,0),(0,1,0),(0,0,1)
is an orthonormal basis for R3

 This section identifies some advantages of orthonormal bases,

and develops a procedure for constructing such bases, known
as Gram-Schmidt orthonormalization process

5.45
 Ex 1: A nonstandard orthonormal basis for R3
Show that the following set is an orthonormal basis
v1 v2 v3
 1 1   2 2 2 2 2 2 1 
S   , , 0 ,     , 
 6 , 6 , 3 , 3
,
3 3 
 2 2   
Sol:
First, show that the three vectors are mutually orthogonal
v1  v 2   16  16  0  0
2 2
v1  v 3   0 0
3 2 3 2
2 2 2 2
v 2  v3     0
9 9 9

5.46
 Second, show that each vector is of length 1
|| v 1 ||  v 1  v 1  1
2
 12  0  1
|| v 2 ||  v 2  v 2  2
36
 2
36
 8
9
1
|| v 3 ||  v 3  v 3  4
9
 94  1
9
1
Thus S is an orthonormal set

Because these three vectors are linearly independent (you can

check by solving c1v1 + c2v2 + c3v3 = 0) in R3 (of dimension 3), by
Theorem 4.12 (given a vector space with dimension n, then n
linearly independent vectors can form a basis for this vector space),
these three linearly independent vectors form a basis for R3.
 S is a (nonstandard) orthonormal basis for R3
5.47
 Ex : An orthonormal basis for P2(x)
In P2 ( x), with the inner product  p, q  a0b0  a1b1  a2b2,
the standard basis B  {1, x, x 2 } is orthonormal
Sol:
v1  1  0 x  0 x 2 , v2  0  x  0x2 , v3  0  0x  x 2 ,
Then
 v1 , v 2   (1)(0)  (0)(1)  (0)(0)  0
 v1 , v3   (1)(0)  (0)(0)  (0)(1)  0
 v 2 , v3   (0)(0)  (1)(0)  (0)(1)  0

v1   v1 , v1   11   0  0    0  0   1
v2   v2 , v2    0  0   11   0  0   1
v3   v3 , v3    0  0    0  0   11  1 5.48
 Theorem 5.10: Orthogonal sets are linearly independent
If S  {v1 , v 2 ,  , v n } is an orthogonal set of nonzero vectors
in an inner product space V, then S is linearly independent
Pf:
S is an orthogonal set of nonzero vectors,
i.e.,  vi , v j   0 for i  j, and  vi , vi   0
(If there is only the trivial solution for ci’s,
For c1 v1  c2 v 2   cn v n  0 i.e., all ci’s are 0, S is linearly independent)
 c1 v1  c2 v 2   cn v n , v i    0, v i   0 i
 c1  v1 , v i   c2  v 2 , v i    ci  v i , v i    cn  v n , v i 
 ci  v i , vi   0 (because S is an orthogonal set of nonzero vectors)
 vi , vi   0  ci  0 i  S is linearly independent
5.49
 Corollary to Theorem 5.10:
If V is an inner product space with dimension n, then any
orthogonal set of n nonzero vectors is a basis for V
1. By Theorem 5.10, if S = {v1, v2, …, vn} is an orthogonal set of n
vectors, then S is linearly independent
2. According to Theorem 4.12, if S = {v1, v2, …, vn} is a linearly
independent set of n vectors in V (with dimension n), then S is a
basis for V
※ Based on the above two arguments, it is straightforward to
derive the above corollary to Theorem 5.10

5.50
 Ex 4: Using orthogonality to test for a basis
4
Show that the following set is a basis for R
v1 v2 v3 v4
S {( 2 , 3 , 2 ,  2) , (1 , 0 , 0 , 1) , (1 , 0 , 2 , 1) , (1 , 2 ,  1 , 1)}
Sol:
v1 , v 2 , v3 , v 4 : nonzero vectors
v1  v 2  2  0  0  2  0 v 2  v 3  1  0  0  1  0
v1  v 3  2  0  4  2  0 v 2  v 4  1  0  0  1  0
v1  v 4  2  6  2  2  0 v3  v 4  1  0  2  1  0
 S is orthogonal
 S is a basis for R 4 (by Corollary to Theorem 5.10)
※ The corollary to Thm. 5.10 shows an advantage of introducing the concept of
orthogonal vectors, i.e., it is not necessary to solve linear systems to test
whether S is a basis (e.g., Ex 1 on 5.44) if S is a set of orthogonal vectors 5.51
 Theorem 5.11: Coordinates relative to an orthonormal basis
If B  {v1 , v 2 ,  , v n } is an orthonormal basis for an inner
product space V, then the unique coordinate representation of a
vector w with respect to B is
w   w, v1  v1   w, v 2  v 2    w, v n  v n
※ The above theorem tells us that it is easy to derive the coordinate
representation of a vector relative to an orthonormal basis, which is
another advantage of using orthonormal bases
Pf:
B  {v1 , v 2 , is
an, vorthonormal
n} basis for V
w  k1v1  k 2 v 2    k n v n V (unique representation from Thm. 4.9)
1 i j
Since  v i , v j    , then
0 i j
5.52
 w, vi  (k1v1  k2 v 2   kn v n ), v i
 k1 v1 , v i   ki vi , v i   kn v n , v i
 ki for i = 1 to n
 w   w, v1  v1   w, v 2  v 2     w, v n  v n

 Note:
If B  {v1 , v 2 ,  , v n } is an orthonormal basis for V and w V ,
Then the corresponding coordinate matrix of w relative to B is
  w , v1  
 w , v  
w B  2 

  
 
 w , v n  
5.53
 Ex
For w = (5, –5, 2), find its coordinates relative to the standard
basis for R3
 w, v1   w  v1  (5 ,  5 , 2)  (1 , 0 , 0)  5
 w, v 2   w  v 2  (5,  5 , 2)  (0 , 1 , 0)  5
 w, v 3   w  v 3  (5 ,  5 , 2)  (0 , 0 , 1)  2
5 
 [w ]B   5
 2 
※ In fact, it is not necessary to use Thm. 5.11 to find the coordinates relative
to the standard basis, because we know that the coordinates of a vector
relative to the standard basis are the same as the components of that vector
※ The advantage of the orthonormal basis emerges when we try to find the
coordinate matrix of a vector relative to an nonstandard orthonormal basis
(see the next slide) 5.54
 Ex 5: Representing vectors relative to an orthonormal basis
Find the coordinates of w = (5, –5, 2) relative to the following
orthonormal basis for R 3
v1 v2 v3
B  {( 53 , 54 , 0) , ( 54 , 53 , 0) , (0 , 0 , 1)}
Sol:
 w, v1   w  v1  (5 ,  5 , 2)  ( 53 , 54 , 0)  1
 w, v 2   w  v 2  (5,  5 , 2)  ( 54 , 53 , 0)  7
 w, v 3   w  v 3  (5 ,  5 , 2)  (0 , 0 , 1)  2
  1
 [w ]B   7 
 2 
5.55
 The geometric intuition of the Gram-Schmidt process to find an
orthonormal basis in R2

v2
w2
v2 w1  v1
v1
projw1 v 2

w 2  v 2  projw1 v 2 is
v1 , v 2 is a basis for R 2
orthogonal to w1  v1

w1 w2
{ , } is an orthonormal basis for R 2
w1 w2
5.56
 Gram-Schmidt orthonormalization process:
B  {v1 , v 2 , , v n } is a basis for an inner product space V
Let w1  v1 S1  span({w1})
v 2 , w1
w 2  v 2  projS1 v 2  v 2  w1 S2  span({w1 , w 2 })
w1 , w1
v3 , w1 v3 , w 2
w 3  v3  projS2 v3  v3  w1  w2
w1 , w1 w2 , w2
The orthogonal projection onto a
n 1
vn , wi subspace is actually the sum of
w n  v n  projSn1 v n  v n   w i orthogonal projections onto the vectors
i 1 w i , w i in an orthogonal basis for that subspace
(I will prove it on Slides 5.67 and 5.68)
 B'  {w1 , w 2 , , w n } is an orthogonal basis
w1 w 2 wn
 B ''  { , , , } is an orthonormal basis
w1 w 2 wn
5.57
 Ex 7: Applying the Gram-Schmidt orthonormalization process
Apply the Gram-Schmidt process to the following basis for R3
v1 v2 v3
B  {(1 , 1 , 0) , (1 , 2 , 0) , (0 , 1 , 2)}
Sol:

w1  v1  (1 , 1 , 0)
v 2  w1 3 1 1
w2  v2  w1  (1 , 2 , 0)  (1 , 1 , 0)  ( , , 0)
w1  w1 2 2 2
v 3  w1 v3  w 2
w3  v3  w1  w2
w1  w1 w2  w2
1 1/ 2 1 1
 (0 , 1 , 2)  (1 , 1 , 0)  ( , , 0)  (0 , 0 , 2)
2 1/ 2 2 2 5.58
Orthogonal basis
1 1
 B'  {w1 , w 2 , w 3}  {(1, 1, 0), ( , , 0), (0, 0, 2)}
2 2

Orthonormal basis
w1 w 2 w 3 1 1 1 1
 B' '  { , , }  {( , , 0), ( , , 0), (0, 0, 1)}
w1 w 2 w 3 2 2 2 2

5.59
 Ex 10: Alternative form of Gram-Schmidt orthonormalization process
Find an orthonormal basis for the solution space of the
homogeneous system of linear equations
x1  x 2  7 x4  0
2 x1  x 2  2 x3  6 x 4  0

5.60
 Ex 10: Alternative form of Gram-Schmidt orthonormalization process
Find an orthonormal basis for the solution space of the
homogeneous system of linear equations
x1  x 2  7 x4  0
2 x1  x 2  2 x3  6 x 4  0

Sol:

 x1   2s  t   2  1 
 x   2s  8t   2   8
  2     s   t  
 x3   s   1   0 
       
x
 4  t  0 1
5.61
Thus one basis for the solution space is
B  {v1 , v 2 }  {( 2 , 2 , 1 , 0) , (1 ,  8 , 0 , 1)}
w1 1  2 2 1 
w1  v1 and u1    2, 2, 1, 0    , , , 0 
w1 3  3 3 3 
 v 2 , u1 
w 2  v 2   v 2 , u1  u1 (due to w 2  v 2  u1 and u1 , u1   1)
u1 , u1 
  2 2 1    2 2 1 
 1,  8, 0, 1  1,  8, 0, 1   , , , 0    , , , 0 
  3 3 3   3 3 3 
  3,  4, 2, 1
※ In this alternative form,
we always normalize wi
to be ui before
w2 1
u2    3,  4, 2, 1 processing wi+1
w2 30 ※ The advantage of this
method is that it is
  2 2 1    3  4 2 1  easier to calculate the
 B' '   , , ,0  ,  , , ,  orthogonal projection
 3 3 3   30 30 30 30  of wi+1 on u1, u2,…, ui
5.62
 Alternative form of the Gram-Schmidt orthonormalization process:
B  {v1 , v 2 , , v n } is a basis for an inner product space V

w1 v1
u1  
w1 v1
w2
u2  , where w 2  v 2  v 2 , u1 u1
w2
w3
u3  , where w 3  v 3  v 3 , u1 u1  v 3 , u 2 u 2
w3

n 1
wn
un  , where w n  v n   v n , ui ui
wn i 1

 {u1 , u 2 , , u n } is an orthonormal basis for V

5.63