MTH 2102 LINEAR ALGEBRA II

TEACHING NOTES

COURSE OUTLINE

1. Linear Transformations
2. Linear functionals and dual space
3. Eigenvalues and Eigenvectors
4. The characteristic polynomial
5. Annihilators
6. Bilinear quadratic and Harmitian forms

Chapter 1: LINEAR TRANSFORMATIONS

Definition: Let V and W be vector spaces. The mapping T from a vector space V to vector space W is a
simple rule that associates with each vector V (x ) in vector V exactly one vector y in W and this is
denoted by T ( x )= y or simply Tx= y .

Note:

(i) The vector space V is called the domain of the mapping and the set of vectors that map
onto in vector space W is called the range of the transformation.

V W

X T

(ii) T :V → W denotes a mapping from V into W .

(iii) The basic operations of addition and scalar multiplication in vector space help in combining
vectors only linearly that is to form linear combinations of vectors. Since the
transformations of T ( c1 v 1 +c 2 v 2 +…+ cn v n) =c 1 T (v 1)+c 2 T (v 2)+…+ c n T ( v n) for all
vectors v 1 , v 2 , … , v n and all scalars c 1 , c2 , … , c n and we also define that:
 T ( x + y )=T ( x )+T ( y ) ∀ x , y in V
 T ( cx )=cT ( x ) ∀ x in V and scalars c i

Linear Operator or Transformation

Note: The mapping which does not satisfy the above definition is said to be a nonlinear transformation.
Example 1. Defining T :C 1 (1)→ C0 (1) by T ( f )=f 1 show that T is a linear transformation.

Solution

Let f and g be in C 1(1) and C be a real number or a scalar.

Then :

(i) T ( f + g )=T ( f )+T ( g )

1 1
f +g

T ( f ) +T ( g)

Hence T ( f + g )=T ( f )+T ( g)

(ii) T ( c f )=cT (f ). By definition of a linear operator and thus the mapping T satisfies both
properties of vector addition and scalar multiplication.

Example 2. Defining T : R 2 → R 2 by T (x 1 , x 2)→(x 1−x 2 , x 1+ x 2 ) . Show that T is a linear

transformation.

Solution

Let x=(x 1 , x 2) and y=( y 1 , y 2 ) be arbitrary vectors in R2

Such that T ( x + y )=T [ ( x 1 , x 2 ) + ( y 1 , y 2) ]

¿ T (x 1+ y 1 , x 2+ y 2 )

By definition we get T ( x 1 , x 2 ) =( x 1−x 2 , x 1+ x 2 )

¿ [ ( x 1 + y 1 )−( x 2 + y 2 ) , ( x 1+ y 1 ) + ( x2 + y 2 ) ]

¿ [ ( x 1−x 2 ) + ( y 1− y 2) , ( x 1 + x 2) + ( y 1+ y 2 ) ]

¿ T ( x 1 , x 2 ) +T ( y 1 , y2 )

¿ T ( x ) +T ( y )

Thus T (x+ y)=T ( x )+ T ( y)

Similarly for any real number C we have T ( C x ) =T [C ( x 1 , x 2)]

¿ T [C x ¿ ¿ 1 , C x 2 ]¿

¿ C [x ¿ ¿ 1−x 2 , x1 + x 2 ]¿
 ¿ CT (x ¿ ¿ 1 , x2 )¿

¿ CT (x )

Thus TC ( x )=CT ( x)

Therefore T is a linear transformation.

Exercise

1. Let A be an m ×n matrix and defining a transformationT : R n → Rmby T ( x )= A (x) . Show that

T is a linear transformation.
2. Explicitly write out a transformation T : R 3 → R 2 defined by T ( x )= A (x). Where A is the matrix

[ 12 3 −1
1 5 ]
Solutions

1. From the definition of T and properties of addition and scalar multiplication, it follows that
T ( x + y )= A ( x + y)
¿ A ( x )+ A ( y )

¿ T ( x ) +T ( y )

T ( x + y )=T ( x )+T ( y )

Similarly, T ( C x ) =A ( C x )

¿ CA x

Thus T ( C x ) =CT (x ). Thus T is a linear transformation.

2. Let x=(x 1 , x 2 , x 3) in R3 for any vector space x .

We can have T ( x )= A ( x )

[]
x
¿
[
1 3 −1 1
2 1 5
x2
x3
]
¿
[ x 1+3 x 2−x 3
2 x 1+ x 2 +5 x 3 ]
This implies T ( x 1 , x 2 , x3 ) =(x 1 +3 x 2−x 3 , 2 x1 + x 2 +5 x3 )
Thus T : R 3 → R 2 is satisfied.
Question

DefiningT : R 2 → R 2by T ( x 1 , x 2 ) =( x 1+1 , x 2−2 ). Show that T is a linear transformation.

Again considering a finite dimensional space we can let { v 1 , v2 , … , v n } be a basis for vector space V ,
then for all scalars c 1 , c2 , … , c n such that V =c 1 v 1 +c 2 v 2 +…+ c n v n. Thus applying transformation T
on vector space V we have T (V )=T (c ¿ ¿ 1 v ¿ ¿ 1+ c 2 v 2+ …+c n v n )¿ ¿

¿ c 1 T (v ¿¿ 1)+c 2 T (v ¿¿ 2)+ …+c n T (v n)¿ ¿

Note: Bearing in mind that T preserves linear combinations.

Example

Let T : R 2 → R 2 be a linear transformation and suppose the basis vectors v 1=(1 , 0) and v 2=(0 , 1) are
transformed as T ( v 1 )=(3 ,−2)and T ( v 2 )=(4 ,5).

(a) Determine T ( x ) for an arbitrary vector x=(x 1 , x 2)

(b) From part (a) above find T (2 ,−1) and T (−3 , 5)

Solution

(a) We can express x=(x 1 , x 2) in terms of the basis { v 1 , v2 } as

x=x 1 ( 1 , 0 ) + x 2 (0 , 1)

¿ x 1 v 1+ x 2 v 2

This implies T (x)=T [ x 1 v 1+ x 2 v 2 ]

¿ x 1 T (v1 )+ x 2 T (v 2 )by definition

¿ x 1 (3 ,−2)+ x 2(4 , 5)

T ( x )=¿¿

(b) T ( 2 ,−1 )=( 3 ×2+ 4 ×−1 ,−2 ×2+5 ×−1 )

¿ ( 6−4 ,−4±5 )
¿(2,−9)

T (−3 , 5 ) =( 3×−3+4 × 5 ,−2 ×−3+5 ×5 )

 ¿(11,31)

Definition: If T 1 :V →W and T 2 :V →W are two linear transformations we say that T 1=T 2 if and
only if T 1 ( v )=T 2(v ) for all v in vector space V .

The addition of two linear transformations and scalar multiplication of the same can be defined as
( T 1+ T 2 ) v =T 1 ( v ) +T 2 (v ) and it is true for all v . (cT ¿¿ 1) v=c T 1 ( v ) ∀ v ¿.

When T 1 :V →W and T 2 :V →W are linear transformations and C is a scalar then T 1+T 2 and CT 1 are
also linear transformations.

Proof

Here we need to show that the properties

(a) T ( x + y )=T ( x )+T ( y ) is true for all x , y in V

(b) CT ( x )=CT (x)is also true for all x in V and ∀ Ci are satisfied.

By considering T 1+T 2 for any arbitrary vectors v 1 and v 2 in vector space V .

Then we have (T ¿ ¿ 1+ T 2 ) ( v 1 +v 2 ) =T 1 ( v 1 + v 2) + T 2 ( v 1 + v 2) ¿

¿ T 1( v ¿¿ 1)+ T 1 ( v 2 ) +T 2 (v ¿¿ 1)+T 2 ( v 2) ¿ ¿

¿ T 1( v ¿¿ 1)+ T 2 ( v 1 ) +T 1 (v ¿¿ 2)+T 2 ( v 2) ¿ ¿

¿(T ¿ ¿ 1+T 2)v 1 +¿(T ¿ ¿ 1+T 2)v 2 ¿ ¿

Thus (T ¿ ¿ 1+ T 2 ) ( v 1 +v 2 ) =T 1 ( v 1 + v 2) + T 2 ( v 1 + v 2) ¿

Secondly if K is any scalar then (T ¿ ¿ 1+ T 2 ) ( Kv )=T 1 ( Kv ) +T 2 (Kv )¿

¿ KT 1 (v)+ KT 2(v )

KT 1 ( v ) + KT 2 ( v )=K (T ¿ ¿ 1+ T 2 )v ¿

(T ¿ ¿ 1+ T 2 ) ( Kv )=K (T ¿ ¿ 1+T 2)v ¿ ¿

Thus the second property is satisfied. This follows that T 1+T 2 is a linear transformation.

Activity
 1. Let A= [ 2 1
−1 3 ]
and B=
1 −1
2 1 [ ] 2 2 2 2
. Defining T 1 : R → R and T 2 : R → R by T 1 ( x )= Ax and

T 2 ( x )=Bx .
Find (a) T 1+T 2
(b)C T 1

2. Let T : R 2 → R 3 be a linear transformation defined by T ( 1 , 0 )=(1 ,−1 , 2) and T ( 0 , 1 )=(2 , 1,−1) .

2
Show that for all ( x 1 , x 2 ) ∈ R , T ( x 1 , x 2 )=(x 1+2 x 2 , x 1+ x2 , 2 x 1−x 2 )

Chapter 2:LINEAR FUNCTIONALS AND DUAL SPACES

When V is a vector space over a field of scalars K the mapping ∅ :V → K is called a linear functional.
Sometimes it is also referred to as a linear form if for every vector u , v ∈V and scalars a , b ∈ K the
mapping ∅ ( a u+b v )=a ∅ ( u ) +b ∅ (v ) by definition of linear form.

Note: A linear functional on V is a mapping from vector space V into K .

Example 1. Let the linear mapping ∅ : R2 → R and δ : R 2 → R be linear functionals defined by

∅ ( x , y )=2 x + y and δ ( x , y )=3 x− y . Find:

(a) ∅ +δ
(b) 2 ∅ −5 δ

Solution

(a) ( ∅ +δ ) ( x , y )= ∅ ( x , y ) +δ (x , y )
¿ 2 x+ y+3 x− y
¿5 x
(b) 2 ∅ −5 δ=2 ∅ ( x , y )−5 δ (x , y )
¿ 2 ( 2 x + y )−5(3 x− y )
¿ 4 x+2 y−15 x +5 y
¿−11 x+7 y

Example 2. Let V be a vector space of polynomials in t over R . Let g :V → R be the integral operator
b
defined by g ( P ( t ) ) =∫ P(t) dt where g is linear and hence it is a linear functional on V . Given that
a
 1
V = { a+bt : a , b ∈ R } . Let ∅ 1 : V → R and ∅ 2 : V → R be defined by ∅ 1 f ( t )=∫ f (t)dt and
0
2
∅ 2 f ( t )=∫ f (t)dt . Find the basis { v 1 , v2 } of V which is dual to { ∅1 , ∅ 2 }.
0

Solution

Let
v 1=a+bt
v 2=c +dt }
by definition of dual basis

Such that ∅ 1 ( v 1 )=1 , ∅ 1 ( v 2 )=0

∅ 2 ( v 1 )=0 , ∅2 ( v 2) =1

[ ]
1 1
bt2
Case I⟹ ∅ 1 ( v 1 )=∫ ( a+bt ) dt=1 ; at + =1
0
2 0

b
a+ =1or2 a+b=2 …………………………………..(1)
2

[ ]
2 2
b t2
Again∅ 2 ( v 1 )=∫ ( a+bt ) dt=0 ; at + =0
0
2 0

2 a+2 b=0ora+ b=0............................(2)

Solving (1) and (2) gives a=2 and b=−2

[ ]
1 1
d t2
Case II ∅ 1 ( v 2 )=∫ ( c+ dt ) dt=0 ; ct+ =0
0
2 0

2 c +2 d=0orc +d=0............................(3)

[ ]
2 2
dt2
Again ∅ 2 ( v 2 )=∫ ( c+ dt ) dt=1 ; ct + =1
0
2 0

2 c +2 d=1 … … … … …..........................(4)

−1
Solving (3) and (4) gives c= and d=1
2
 { 1
}
Thus 2−2 t ,− +t becomes the basis of V which is dual to { ∅1 , ∅ 2 }
2

Example 3: Let V be a vector space of m square matrices over K and let T :V → K be a trace mapping
such that T ( A )=a 11 +a22 +…+ ann where A=(aij ). T assigns to a matrix A a sum of diagonal
elements.

Note: A set of linear functionals on a vector space V over a field of scalars K is also a vector space over
K with addition and scalar multiplication defined by ( ∅ +δ ) ( v )= ∅ ( v )+ δ(v ) and ( K ∅ ) v=K ∅ (v) where
∅ and δ are linear functionals of V and k ∈ K . This space is referred to as dual space of vector space V
¿
and the dual space of V is denoted by V

For instance: when the vector space V =K n is the vector space of n−¿turples we may write them as
¿
column vectors and the dual space V is identified with space of row vectors. Particularly any linear
functional ∅ =( a1 , a2 , … , an ) in V can be represented by
¿

()
x1
∅ ( x i )=( a1 , a 2 , ⋯ , a n ) x 2 =a1 x 1+ a2 x 2 +…+ an xn which is linear form.
⋮
x3

Dual basis

¿
Each basis of V determines the basis of the dual space V

Theorem: Let { v 1 , v2 , ⋯ , v n } be a basis of V over K . Let ∅ 1 , ∅2 , ⋯ ∅n ∈V be linear functionals defined

¿

by ∅ i ( v i ) =δ ij = {10ifif i=i ≠ jj }
Then { ∅1 , ∅ 2 , … , ∅ n } is a basis of V and δ ij is a short way of writing
¿

∅ 1 (v¿ ¿1)=1 ∅ 1 ( v¿ ¿2)=0 ∅1 (v ¿¿ 3)=0 ⋯ ∅1 ( v ¿¿ n)=0 ¿ ¿ ¿ ¿

∅ 2 (v¿ ¿1)=0 ∅2 (v ¿¿ 2)=1 ∅ 2(v ¿¿ 3)=1 ⋯ ∅ 2 (v¿ ¿ n)=0 ¿ ¿ ¿ ¿

∅ n (v ¿¿ 1)=1 ∅n (v ¿¿ 2)=0 ∅ n (v ¿¿ 3)=1 ⋯ ∅ n (v ¿¿ n)=1 ¿ ¿ ¿ ¿

Example 1. Considering the following basis of R = {v 1= (2 , 1 ) , v 2=(3 ,1) }.

2

Find the dual basis { ∅1 , ∅ 1 }.

Solution

Looking for linear functionals

∅ 1 ( x , y )=ax +by

∅ 2 ( x , y )=cx + dy

Such that ∅ 1 (v¿ ¿1)=1 ∅ 2 (v¿ ¿1)=0¿ ¿

∅ 2 (v¿ ¿1)=0 ∅2 ( v ¿¿ 2)=1 ¿¿

⟹ ∅ 1 (v ¿¿ 1)= ∅1 ( 2 ,1 )=2a +b=1 ⋯ (1)¿

∅ 1 (v¿ ¿ 2)= ∅1 ( 3 , 1 )=3 a+b=0 ⋯ (2)¿

∅ 2 (v¿ ¿1)= ∅2 ( 2 ,1 )=2 c+ d=0 ⋯ (3)¿

∅ 2 (v¿ ¿1)= ∅2 ( 3 , 1 )=3 c +d=1 ⋯ (4 )¿

Solving ( 1 ) and (2)gives a=−1and b=3

Solving (3) and (4)gives c=1 and d=−2

Hence the dual basis is { ∅1 ( x , y )=−x+3 y , ∅ 2 ( x , y )=x−2 y }

Question

Considering the following basis of R3 . { v 1=( 1 ,−1 , 3 ) , v 2=( 0 , 1 ,−1 ) , v 3=(0 , 3 ,−2) }

Find the dual basis { ∅1 , ∅ 2 , ∅ 3 }

HINT

We need to find the linear functionals

∅ 1 ( x , y , z )=a1 x +a 2 y + a3 z ∅ 2 ( x , y , z )=b 1 x +b2 y +b3 z ∅3 ( x , y , z )=c 1 x +c 2 y+ c 3 z

Such that

∅ 1 (v¿ ¿1)=1 ∅ 1 ( v¿ ¿2)=0 ∅1 (v ¿¿ 3)=0 ¿ ¿ ¿

∅ 2 (v¿ ¿1)=0 ∅2 (v ¿¿ 2)=1 ∅ 2(v ¿¿ 3)=0 ¿ ¿ ¿

∅ 3 (v ¿¿ 1)=0 ∅ 3( v ¿¿ 2)=0 ∅ 3 ( v¿ ¿ 3)=1¿ ¿ ¿

Chapter 3: EIGEN VALUES AND EIGEN VECTORS

n
Considering f (t) as a polynomial over a field of scalars K as f ( t )=an t + ⋯ +a1 t+ a0

n
If A is a square matrix over a field of scalars K we always define f ( A )=a n A + ⋯ +a 1 A+ a0 I where I
is an identity matrix.

Note: We say that A is a root or is a zero of a polynomial f ( t ) if f ( A )=0.

Example 1
If A=[ ]1 2
3 4
prove that matrix A is a root of

(a) f ( t )=2 t 2−3 t+7

(b) g ( t )=t 2 −5t−2

Solution

(a) Given f ( t )=2 t 2−3 t+7 and A= [ 13 24 ]

[ 13 24 ] −3[ 13 24 ]+7 [10 01]
2
Then f ( A )=2

¿2
[ 13 24 ][ 13 24]−[39 126 ]+[70 07 ]
¿2
[ 157 1022]−[−49 65 ]
¿
[1430 2044]−[−49 65]
¿
[1821 1439]
In this case A is not a root of the polynomial f ( t )=2 t 2−3 t+7

(b) Given g ( t )=t 2 −5t−2 and A= [ 13 24 ]

[ 13 24] −5[ 13 24 ]−2[ 10 01]
2
Then g ( A )=

¿
[13 24 ][ 13 24 ]−[155 1020]+[ 20 02]
¿
[157 1522]−[ 105 1520 ]+[ 20 02]
¿
[00 00]
In this case A is a root of the polynomial g ( t )=t 2 −5t−2

Example 2

Show that A= [ 12 43] is a zero of f ( t )=t 2−4 t−¿5

Solution
 Since A= [ ]
1 4
2 3
and f ( t )=t 2−4 t−¿5 ,

Then f ( A )= A2−4 A−5

[ ] [ ] [ ]
2
1 4 1 4 1 0
¿ −4 −5
2 3 2 3 0 1

¿
[12 43 ][ 12 43]−[ 48 1612]−[ 50 05]
¿
[00 00]
Example 3

Find f ( A) where A= [ 14 −25 ] and f ( t )=t 2−3 t+ 7

Solution

Since A= [ 14 −25 ] and f ( t )=t 2−3 t+ 7

Then f ( A )= A2−4 A−5

[ ] [ ] [ ]
2
1 −2 1 −2 1 0
¿ −3 +7
4 5 4 5 0 1

¿
[ 14 −25 ][ 14 −25 ]−[123 −615 ]+[ 70 07]
¿
[−724 −12
21 ] [ 12 15 ] [ 0 7 ]
−
3 −6 7 0
+

¿
[−312 −613 ]
Theorem

When f and g are polynomials over a field of scalars K and A is an n - square matrix over K then we
have

(i) ( f +g ) A=f ( A ) +g ( A)
(ii) fg ( A ) =f ( A) g( A)
(iii) ( kf ) A=kf ( A ) for k ∈ K

Note: Since f ( t ) g ( t )=g ( t ) f ( t ) for polynomials f (t)and g(t ), then f ( A ) g ( A ) =g ( A)f ( A).

This implies that for any two polynomials of matrix A they commute.

Proof of the Theorem

 n n
Let f =a n t + ⋯ + a1 t +a 0 and g=bn t + ⋯ + b1 t+b 0

n n
Then by definition of f ( A)=a n A + ⋯ +a1 A+a 0 I and g( A)=bn A + ⋯ +b 1 A+ b0 I

(i) Suppose m ≤n and let b i=0 if i>m

n
Then ( f +g ) A=( an +bn ) A + ⋯ + ( a1+ b1 ) A + ( a 0 +b0 ) I

n n
So ( f +g ) A=a n A + bn A + ⋯ + a1 A +b1 A +a 0 I +b 0 I

¿ f ( A )+ g ( A)

Thus ( f +g ) A=f ( A ) +g ( A)

(ii) By definition

n+m
fg=C n+m t n +m + ⋯ + C1 t+C 0= ∑ C k t K
K=0

K
Where C K =a 0 b K + a1 bK −1+ ⋯ + aK b 0= ∑ ai bK −i
K=0

n+m
Hence fg ( A ) = ∑ C k A
K

K =0

( )(∑ )
n m
And f ( A ) g ( A ) = ∑ ai A i bj Aj
i=0 j=0

n m
¿ ∑ ∑ ai b j A i+ j
i=0 j=0

n +m
¿ ∑ C k A K =fg( A)
K=0

(iii)By definition of scalar multiplication

n n−1
Kf =K a n t + K an−1 t + ⋯ K a1 t+ K a0

n n−1
Where f ( t )=an t +a n−1 t + ⋯ + a1 t

n n−1
⟹ ( Kf ) A=K an A + K an−1 A + ⋯ K a1 A + K a 0 I
 n n−1
¿ K (a¿¿ n A +a n−1 A + ⋯ a1 A+ a0 I )¿

¿ Kf ( A)

⇔ ( Kf ) A=Kf ( A )

Eigen values and Eigenvectors

LetT :V ⟶ V be a linear operator on a vector space V over a field of scalars K , a scalar λ ∈ K is called
an eigen value of T if there exists a nonzero vector V which is a member of a vector space V s.t
T ( v )=λv .

Note:

(i) Every vector that satisfies this relation is called an eigen vector of T belonging to eigen value λ .
(ii) Each vector Kv is an eigen vector s.t T ( Kv )=KT ( v )=K ( λv )=(Kv)
(iii) Terms like characteristic value/ proper value and characteristic vector/ proper vector are
synonymously used as eigen value and eigen vector respectively.

Example 1

Find the eigen values and associated eigen vectors of a matrix A= [ 13 24 ]

Solution

We need the scalar t and a nonzero vector X = ( xy) s.t we have Ax=tx and since A=[ 13 24 ] we find
that: (13 24 )( xy)=t ( xy)
x +2 y=tx
 3 x+ 2 y =ty

( 1−t ) x+2 y=0

3 x+ ( 2−t ) y =0

For homogeneous equation to have a nonzero solution the determinant shoulb equal to zero (0).

|1−t3 2
2−t |
=0

( 1−t ) ( 2−t )−( 2 ×3 ) =0

2
t −3 t−4=0

( t+ 1 )( t−4 ) =0

So t=4 or t=−1

⟹ t is an eigen value iff t=4 or t=−1

When t=4

( 1−4 ) x +2 y=0

−3 x+ ( 2−4 ) y=0

−3 x+ 2 y =0 ⋯ (1)

3 x−2 y=0 ⋯ (2)

⟹ 3 x−2 y=0

x=2and y=3

X= ( xy)=(23)
The eigen vector V = ( xy )=(23) is the eigen value t=4
When t=−1

( 1+1 ) x+2 y=0

 3 x+ ( 2+ 1 ) y =0

2 x+ 2 y =0

3 x+ 3 y=0

Thus v= (−11) or v=(−11 ) is a nonzero eigen vector belonging to the eigen value t=−1 and the other
eigen vector belonging to t=−1 is a multiple of v .

Diagonalisation and Eigen Vectors

Let T :V → V be a linear operator on a vector space V with a finite dimension n and T can be
represented by a diagonal matrix

[ ]
K1 0 ⋯ 0
0 K2 ⋯ 0
⋯ ⋯ ⋯ ⋯
0 0 ⋯ Kn

If and only if there exists a basis (v 1 , v 2 , ⋯ , v n ) of V for which

T ( v 1 )=K 1 v 1

T ( v 2 )=K 2 v 2

⋯ ⋯⋯⋯

T ( v n )=K n v n

is such that the vectors v 1 , v 2 , ⋯ , v n are eigen vectors of the operator T belong to the eigen values
K 1 , K 2 , ⋯ , K n respectively.

Theorem

An n -square matrix A is similar to a diagonal matrix B iff A has n linearly independent eigen vectors in
this case the diagonal elements of B are the corresponding eigen values.

From the above theorem, if we let P be the matrix whose columns are the n independent eigen vectors
of A then B=P−1 AP

Example
Let A= (12 43).
(a) Find all eigenvalues of A and corresponding eigenvectors.
(b) Find an invertible matrix P such that P−1 AP is diagonal.

Solution

(a) The characteristic polynomial tI −A of A is tI −A= (0t 0t )−(12 43)

¿ ( t−1
−2
−4
t−3 )
The characteristic polynomial ∆ t of A is given by ∆ t=tI − A

¿|t−1
−2
−4
t−3 |
¿ ( t−1 ) ( t−3 )−8

2
¿ t −4 t−5

¿(t−5)(t+1)

The roots of ( t−5 ) ( t+1 )=0 are 5 and 1

To obtain the eigenvectors belonging to the eigenvalue 5 we substitute t=5 in the characteristic

matrix and obtain (5−1

−2
−4
5−3
= )(
4 −4
−2 2 )
The homogeneous system determined by the above matrix is

(−24 −4 x
2 y )( )
=0

4 x−4 y=0

−2 x+ 2 y =0

Or x− y =0

This yields x=1 and y=1 (And the corresponding multiples of 1)

The eigenvector is v= ( 11)
All multiples of v= ( 11) generate the eigen space of 5.
Secondly, the eigenvectors corresponding to the eigenvalue of -1 are obtained as follows

(−2
−2
−4 x
−4 y )( )
=0

−2 x−4 y=0

x +2 y=0

This yields x=2 and y=−1

The eigenvector is v= (−12 )

All multiples of -1 generate the eigen space of -1.

(b) Let P be the matrix whose columns are the above eigenvectors such that P= (11 −12 )
Then B=P−1 AP

( )(
1 2
¿
3
1
3 1 4 1 2
−1 2 3 1 −2 )( )
3 3

( )(
1 2
¿
3
1
3 5 −2
−1 5 1 )
3 3

¿ ( 50 −10 )
Bis a matrix representation of the operator A in the new basis.
 −1
P AP is diagonal.

Question

[ ]
1 −3 3
Find all the eigenvalues and the basis of the following eigen space A= 3 −5 3
6 −6 4

Can it be diagonalised? Why?

The Characteristic Polynomial

Let A be an n square matrix over a field of scalars K ,

[ ]
a11 a 12 ⋯ a1 n
a a 22 ⋯ a2 n
Where A= 21
⋯ ⋯ ⋯ ⋯
a n1 an 2 ⋯ ann

The matrix tI −A

Where I n is an n square identity matrix and t is a determinant called the Characteristic matrix of .

i.etI −A=¿

The determinant of the above ∆ A ( t )=det ⁡(t I n− A) or |tI − A| which is a polynomial in t is called a
Characteristic Polynomial of matrix A and we call ∆ A ( t )=|tI− A|=0 the characteristic equation of A .

Example

( )
1 3 0
Find the characteristic polynomial of A= −2 2 −1
4 0 −2

Solution

| |
t−1 −3 0
∆ t=tI − A= 2 t−2 1
−4 0 t+2

¿ ( t−1 ) |t−20 t +21 |+3|−42 t+21 |+0

 ¿ ( t−1 ) [ ( t−2 )( t +2 )−0 ] + 3[2 ( t +2 ) + 4]

¿ ( t−1 ) ( t 2 −4 ) +6 t+12+12
3 2
¿ t −t −4 t +6 t +24
3 2
¿ t −t +2 t+ 24

A monic polynomial of degree 3.

Note: A monic polynomial is when the coefficient of t 3 is 1.

Cayley-Hamiliton Theorem

It states that “Every matrix is a zero of its characteristic polynomial”

Example

Given a matrix A= (13 22). Verify the Cayley – Hamiliton Theorem

Solution

∆ t=|tI− A|

¿t (| 10 00)−(13 22)|
¿ |t−1
−3
−2
t−2 |
2
¿ t −3 t−4

But we are expected from Cayley-Hamiliton Theorem, A is a zero of its characteristic polynomial such
that ∆ t=t 2−3 t−4

( ) ( ) ( )
2
¿ 1 2 −3 1 2 −4 1 0
3 2 3 2 0 1

¿ ( 79 106 )−(39 66)−( 40 04)

 ¿ ( 00 00)
Therefore matrix A is a zero of its characteristic polynomial.

Note: There exists an intimate relationship between the characteristic polynomials and eigenvalues.

Theorem

Let A be an n square matrix over a field of scalars K . A scalar λ ∈ K is an eigenvalue of A if and only if
λ is a root of the characteristic polynomial ∆ (t)of A .

Corollary

If the characteristic polynomial ∆ ( t ) of an n square matrix A is a product of distinct linear factors

∆ t=( t−a1 ) ( t−a2 ) ( t−a2 ) ⋯ (t−an )

Where a 1 , a2 , ⋯ , a n are distinct roots of ∆ (t), then A is similar to a diagonal matrix whose elements are
ai

Secondly: Let A be an n square matrix over the complex field C . Then A has at least one eigenvalue.

For example

( )
3 0 0
Let A= 0 2 −5 . We have its characteristic polynomial as ∆ (t) as
0 1 −2

| |
t−3 0 0
∆ t= 0 t−2 5
0 −1 t+2

¿ ( t−3 ) |t−2 5
−1 t+2|−0|
0 5
0 t+2|+ 0|
0
0
t−2
−1 |
¿ ( t−3 ) [ ( t −4 ) +5 ]=(t−3)(t +1)
2 2

3 2
¿ t −3 t +t−3

From the above we consider two cases where

(i) A is a matrix over the real field; In this case A has only one eigenvalue 3. And since 3 has only
one independent eigenvector A is not diagonalizable.
 (ii) A is a matrix over a complex field C then A has three distinct eigenvalues 3 , i and −i . Thus
there exists an invertible matrix P over a complex field C for which

( )
3 0 0
−1
P AP= 0 i 0
0 0 −i
Thus A is diagonalizable.

Question

Let A= (31 −1
1 )
and B=
1 −1
2 1 ( )
. Find all eigenvalues and corresponding eigenvectors of A and B

viewed at as matrices over

(i) The real field R

(ii) The complex field C

ANNIHILATORS
¿
When W is a subset of a vector space , a linear functional ∅ ∈V is called an Annihilator of W if
∅ ( w )=0 for every vector w ∈ W .
¿
i.e if ∅ ( w )={ 0 } a set of such matrix denoted by W 0 , the annihilator of W is a subspace/subset of V (a
dual space). This clearly shows that 0 ∈W 0 and that the subspace ∅ , δ ∈W 0 for every scalar a , b ∈ K
and every w ∈ W we then have:

( a ∅ +bδ ) W =a ∅ ( W )+ bδ(W )

¿ a .0+b .0

¿0

The concept of annihilator enables us to give another interpretation of a homogeneous system of linear
equations:

a 11 x 1 +a 12 x 2+ ⋯ +a 1n x n=0

a 21 x 1+ a22 x 2 + ⋯ + a2 n x n=0

⋯ ⋯⋯⋯ ⋯⋯ ⋯⋯⋯ ⋯⋯⋯ ⋯

a m 1 x 1+ am 2 x2 + ⋯ +a mn x n=0
Each row of coefficient matrix A=( aij ) is an element of K n and each solution of vector∅ =( x1 , x2 , ⋯ x n )
is viewed as an element of the dual space.

Example

Let W be the subspace of R4 , spanned by ( 1 , 2,−3 , 4 ) , ( 1 , 3 ,−2 , 6 )and ( 1 , 4 ,−1 ,8 ). Find the basis of
the annihilator W .

Solution

To find the basis, we shall have a set of linear functionals in the form

∅ ( x , y , z , w )=ax +by +cz + dwwherea , b , c , d ∈ K

Where ∅ ( v 1 ) =0 , ∅ ( v 2 )=0 ∅ ( v3 ) =0

s.t ∅ ( 1 ,2 ,−3 , 4 )=a+2 b−3 c+ 4 d =0

∅ ( 1 ,3 ,−2 ,6 )=a+3 c−2 c+6 d=0

∅ ( 1 , 4 ,−1 , 8 ) =a+ 4 b−c +8 d=0

Solving the above system of equations through echelon steps we obtain:

a+ 2b−2 c +4 d=0

b+ c+ 2d =0

2 b+2 c +4 d=0

Which further gives

∅ 1 ( x , y , , z , w )=5 x− y+ z

∅ 2 ( x , y , z , w ) =2 x + y

∅ 3 ( x , y , z , w )=2 y−w

Question

Given that W is a subspace of R3 spanned by (1 , 1, 0)and (0 , 1 ,1). Find the basis of the annihilator of
W.

Ans: ∅ ( x , y , z ) =x− y + z
Transposing a Linear mapping

Let T :V → U be an arbitrary mapping from a vector space V into a vector space U for any linear
¿
functional ∅ ∈U , the composition ∅ oT is a linear mapping from a vector space V into a field of scalars
K.

V U K

∅ oT
¿
⟹ ∅ oT ∈V corresponding to ∅ → ∅ oT which is a mapping T t →U ¿ → K ¿ and is defined by
t
T ( ∅ )= ∅ oT .

Therefore every vector T t ( ∅ )( v )= ∅ T (v ) for every vector v ∈ V

Note:

(i) The transformation [T ¿¿ t ( ∅ )] ( v )= ∅ [T ( v ) ]¿ is linear for each and every v ∈ V .

Proof

¿
For any scalars a , b ∈ K and any linear functionals∅ , δ ∈U

We have T t ( a ∅ +bδ )=( a ∅ +bδ ) oT

¿ a ( ∅ oT ) +b( δoT )

t t
¿ aT ( ∅ ) +b T ( δ )

t t t
Thus T ( a ∅ +bδ )=aT ( ∅ ) +b T (δ)

Hence T t is linear.

Example

Let ∅ be a linear functional on R2 defined by ∅ ( x , y )=x−2 y for each of the following linear operators
T on R2. Find [T ¿¿ t ( ∅ )](x , y )¿

(i) T ( x , y )=(x , 0)
(ii) T ( x , y )=(2 x−3 y , 5 x +2 y )
Solution

(i) Since ∅ ( x , y )=x−2 y

T ( x , y )=(x , 0)

By definition: [T ¿¿ t ( ∅ )](x , y )= ∅ T (x , y )¿

¿ [x−(2)(0)]

¿x

Such that [T ¿¿ t ( ∅ )](x , y )=x ¿

(ii) Since ∅ ( x , y )=x−2 y

T ( x , y )=(2 x−3 y , 5 x +2 y )

Then[T ¿¿ t ( ∅ )]= ∅ T ( x , y )¿

¿ ∅ (2 x−3 y ,5 x +2 y)

¿ 2 x−3 y−2(5 x +2 y)

¿−8 x−7 y=−(8 x+ 7 y)

Activity

Let ∅ be a linear functional on R2 defined by ∅ ( x , y )=3 x−2 y .

For each linear mapping T : R 3 → R 2. Find [T ¿¿ t ( ∅ )](x , y , z)¿ where

(i) T ( x , y , z )=( x+ y , y + z ) Answer: [3 x + y −2 z ]

(ii) T ( x , y , z )=( x+ y+ z , 2 x− y) Answer: [−x+5 y +3 z ]

Solution

(i) T t ( ∅ )( x , y , z )= ∅ [ T ( x , y , z ) ] = ∅ [ x+ y , y + z ] =3 x+ 3 y −2 y−2 z =3 x + y −2 z
(ii) ∅ [ T ( x , y , z ) ] =∅ [ x + y + z ,2 x− y ]=3 x +3 y +3 z−4 x+2 y=−x +5 y +3 z

Note: We can emphasize that if T is a linear mapping from V into U then T t is a linear mapping from
¿ ¿
U into V i.e

U T V andV ¿ T t U ¿

Theorem
Let T :V → U be linear and let A be a matrix representation of A relative dual to the bases { v i } ∈ V and
{ ui } ∈ U .

Then the transpose matrix At is the matrix representation of T t :U ¿ → V ¿ relative dual to the bases{ ui }
and { v i }.

Practice exercise

LetT :V → U be linear and let T t :U ¿ → V ¿ be its transpose. Show that the kernel of T t is annihilator of
the image of T .

Solution

Supposing∅ ∈ker T t

Let T t ( ∅ )=( ∅ oT ) v=0

Now if u ∈imT , then u=T ( v ) ∀ v ∈V

Hence ∅ ( u )= ∅ (T (v)) by definition

¿ ( ∅ oT )( v )

¿ 0(v )

¿0

Now that ∅ ( u )=0 ∀ u ∈imT

Then ∅ ∈ ( imT )0 ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ (1)

Secondly δ ∈ ( imT )0

i.eδ (imT )= { 0 } ∀ v ∈ V

⟹ [ T t ( δ ) ] ( v )=(δoT )v by definition

¿ δ (T (v ))

¿ 0(v )
 [ T t ( δ ) ] ( v )=0
Hence T t ( δ )=0 ∴ δ ∈ ker T t

0 t
⟹ ( imT ) ⊂ ker T ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ (2)

Both (1) and (2) give the required equality

Thusker T t =( imT )0 as required

BILINEAR QUADRATIC AND HARMITIAN FORMS

Let V be a vector space of definite and finite dimension of a field over K . A bilinear form on V is a
mapping f :V ×V → K and this satisfies

(i) f ( a u1+ b u2 ,V ) =af ( u1 , V ) +bf (u 2 , V )

(ii) f ( u ,a v 1+ b v 2 ) =af ( u , v 1 ) +af ( u , v 2 ) ∀ a , b ∈ K and ∀ ui , v i ∈ V

Note:

The function (i) f is linear in the field variable and

(ii) f is linear in the second variable

Example

1. Let f be the dot product on Rn

i.e f ( u , v )=u . v where u=(a i) and v=bi
f ( u , v )=a1 b1 +a 2 b 2+ ⋯ + an bn

Then in this case f is a bilinear form of Rn

2. When ∅ and δ are arbitrary linear functionals on a vector space V

Let f :V ×V → K be defined by f ( u , v )=∅ (u)δ(v )

The f is bilinear because ∅ and δ are each linear. Such bilinear form is called Tensor Product of ∅ and δ is
denoted by f = ∅ ⨂ δ

3. Let matrix A=( aij ) be an m ×n matrix over a field of scalars K , then A is seen as a bilinear
form on K n by defining
 f ( x , y )=X t AY

( )( )
a 11 a12 ⋯ a 1n y1
¿( X 1 X 2 ⋯ X n ) a21 a22 ⋯ a 2n y2
⋯ ⋯ ⋯ ⋯ ⋮
a n 1 a n2 ⋯ ann y n
¿ a 11 X 1 y 1 +a 22 X 2 y 2+ ⋯ + ann X n y n
n
¿ ∑ aij X i y i
i , j=1

Examples

1. Let u=( x 1 , x 2 , x 3 ) v=( y 1 , y 2 , y 3)

Then when f ( u , v )=3 x1 y 1−2 x 1 y 2 +5 x2 y 1 +7 x 2 y 2−8 x 2 y 3+ 4 x 3 y2 −x3 y 3
Express f in matrix form.

Solution

Let A be a 3 ×3 matrix whose (i , j) entry is the coefficient of X i Y j then

t
f ( u , v )=X AY

( )( )
3 −2 0 y 1
¿( x1 x 2 x 3 ) 5 7 −8 y 2
0 4 −1 y 3
2. Let A be an n × n matrix over a field of scalars K . show that f is a bilinear form on
n t
K :f ( X , Y )= X AY

Solution
n
Let any a , b ∈ K and X i Y i ∈ K
t
By definition f ( a X 1 +b X 2 , Y )=(a X 1+b X 2) AY
t t
¿(a X 1 +b X 2 ) AY
t t
¿ a X 1 AY + b X 2 AY ( By definition)
¿ af ( X 1 , Y ) +bf (X 2 ,Y )
Here f is linear in the first variable
t t t
Also f ( X , a Y 1+ b Y 2 )=X A ( a Y 1+ b Y 2 )=a X A Y 1 +b X A Y 2 (By definition)
¿ af ( X , Y 1 ) +bf (X ,Y 2)
f is linear in the second variable so f is in bilinear form on K n

BILINEAR FORMS AND MATRICES

When f is a bilinear form on a vector space V and when { e 1 , e2 , ⋯ , e n } is the basis of V . Supposing
u , v ∈V and u=a1 e1 + a2 e 2 + ⋯ + an e n

v=b1 e 1+ b2 e2 + ⋯ +bn e n

Then f ( u , v )=f (a1 e 1+ a2 e2 + ⋯ +an e n ,b 1 e 1 +b2 e 2+ ⋯ +b n e n )

¿ a 1 b1 ( e 1 ,e 1 ) +a 2 b 2 ( e 2 , e2 ) + ⋯ +an b n (e n , e n)

n
¿ ∑ a1 b j f (e i ,e j )
i , j=1

Here f is determined by n2 values f (e i , e j )

The matrix A=(aij ) where a ij=f (ei , e j) is the matrix that

n
f ( u , v )= ∑ a1 b j f ( ei , e j )
i , j=1

()
b1
¿ ( a 1 a2 ⋯ an ) A b2
⋮
bn

t
¿ [u] e A v e ∀ u , v ∈V

Remember [u]e =¿ column vector of u ∈V in the basis {ei }

Example

Let f be the bilinear form on R2 defined by f [ ( x1 , x 2) ,( y 1, y 2) ]=2 x 1 y 1 −3 x 1 y 1+ x 2 y 2

(a) Find matrix A of f in the basis [u1=( 1 , 0 ) ,u 2=(1 , 1)]

(b) Find the matrix B of f in the basis { v 1=( 2 ,1 ) , v 2=(1 ,−1) }
(c) Find the transition matrix P from the basis { ui } to the basis { v i } and verify that B=Pt AP

Solution

(a) Let A=(aij ) where a ij=f (u i , u j)

Since f [ ( x1 , x 2) ,( y 1, y 2) ]=2 x 1 y 1 −3 x 1 y 1+ x 2 y 2
 ⟹ a11 =f ( u 1 , v 1 ) =f [ ( 1 , 0 ) , ( 1 , 0 ) ] =2

a 12=f ( u1 , v 2 )=f [ ( 1, 0 ) , ( 1 ,1 ) ] =−1

a 21=f ( u1 ,u 2 )=f [ ( 1 , 1 ) , ( 1 , 0 ) ] =2

a 11=f ( u2 ,u2 ) =f [ ( 1 ,1 ) , ( 1 ,1 ) ] =0

Then A= [ 22 −10 ] is the matrix of f in the basis { u1 , u2 }

(b) B= [ 30 69] is the matrix of f in the basis { v 1 , v2 }

(c) Write v 1 and v 2 in terms of ui in order to get the transition matrix P from one basis to another.
For instance v 1=( 2 ,1 ) =( 1, 0 )+ (1 , 1 )=u1 +u2
v 2=( 1 ,−1 )=2 ( 1 , 0 )−( 1 , 1 )=2 u1−u2
From above

u1 +u2
2 u1−u2
=
}[
1 1
2 −1
=P ]
P=
t
[ 11 −12 ]
SYMETRIC BILINEAR FORMS

A bilinear form f on a vector space V is referred to a Symmetric if f ( u , v )=f ( v ,u ) ∀ u , v ∈ V

If A is a matrix representation of f we then write F ( X , Y ) =X t AY =(X t AY )t =Y t A t X

Note:

t t t
X AY and therefore equals to its transpose if f is symmetric then Y A X=f ( X ,Y )=f ( Y , X ) =Y AX
t

Conversely if matrix A is symmetric then f is symmetric.

For symmetric bilinear forms we have

Theorem

Let f be the symmetric bilinear form of vector space V over a field of scalars K and where 1+1 ≠0 .
Then V has a basis { v 1 , v2 , ⋯ , v n } in which f is represented by a diagonal matrix

i.e f ( v i , v j ) =0 for i≠ j

In other words let A be a symmetric matrix over K (Where 1+1 ≠0 ) then there exists nonsingular

Matrix P such that Pt AP is diagonal.i.e A is congruent to a diagonal matrix.

Note: One way of obtaining the diagonal form of Pt AP is by a sequence of elementary row operations
and the same sequence of the elementary column operations. These operations on I , an identity matrix
will yield Pt .

( )
1 2 −3
For instance: Let M = 2 5 −4 , a symmetric matrix forming a block matrix (M , I ) gives
−3 −4 8

( | )
1 2 −3 1 0 0
( M , I )= 2 5 −4 0 1 0
−3 −4 8 0 0 1

( | )
1 2 −3 1 0 0
Applying R2 →2 R1 + R2∧R 3 → 3 R 1+ R 3 we get 0 1 2 −2 1 0
0 2 −1 3 0 1

Then applying corresponding column operations as C 2 →−2 C 1+C 2∧C 3 → 3 C1 +C 3 gives

( | )
1 0 0 1 0 0
0 1 2 −2 1 0
0 2 −1 3 0 1

Now applying R3 →−2 R2 + R3 gives

( | )
1 0 0 1 0 0
0 1 2 −2 1 0
0 0 −5 7 −2 1

The corresponding column operation C 3 →−2C 2+ C3 gives

( | )
1 0 0 1 0 0
0 1 0 −2 1 0
0 0 −5 7 −2 0
 [ ]
1 0 0
The matrix c has been diagonalizedsosetting P= −2 1 0
7 −2 1

[ ]
1 −2 7
t
P = 0 1 −2
0 0 1

[ ]
1 0 0
t
P AP= 0 1 0
0 0 −5

QUADRATIC FORM

Definition: A mapping q :V → K is a quadratic form if q ( v )=f ( v , v) for some symmetric bilinear form
on V . q is called quadratic form associated with symmetric bilinear form f when 1+1 ≠0 in K then f
1
can be obtained from q in accordance to the identity f ( u , v )= q ( u+ v )−q (u )−q (v ).
2

This is the polar form of f .

For instance: let the mapping f be represented by a symmetric matrix A=(aij ). Then the mapping q can
be represented by q ( x )=f ( X , X )= X t AX

( )( )
a11 a12 ⋯ a1 n X 1
¿ ( x1 x2 ⋯ xn ) a 21 a22 ⋯ a2 n X 2
⋯ ⋯ ⋯ ⋯ ⋮
an 1 an 2 ⋯ a nn X n

n
¿ ∑ aij X i X j
i , j=1

¿ a 11 X 1+ a22 X 2+ ⋯ + ann X n+ 2 ∑ aij X i X j

2 2 2

This expression is called A quadratic polynomial corresponding to the symmetric matrix A

Note: When A has diagonal then q has a diagonal representation

t 2 2 2
q ( x )= X AX=a11 X 1+ a22 X 2+ ⋯ +a nn X n

Example

Find the symmetric matrix corresponding to each of the following quadratic polynomials

1. q ( x , y )=4 x 2−6 xy +7 y 2
 2. q ( x , y )=xy + y 2
3. q ( x , y , z )=3 x 2 +4 xy − y 2+ 8 xz−6 yz + z 2

Solution

Note: The symmetric matrix A=( aij ) which represents q (x 1 , x 2 , ⋯ , x n ) has a diagonal entry a ii equal to
2
the coefficient X 1 has the entries a ij and a ji each equal to half the coefficient of X i X j.

1. Thus q ( x , y )=4 x 2−6 xy +7 y 2 can be represented by the diagonals of 4 and 7

⟹ The required matrix A= (−34 −3

7 )
2. q ( x , y )=xy + y 2
⟹ The diagonal elements are 0 and 1

1 1
Other entries are and
2 2

( )
1
0
2
⟹ The required matrix is A=
1
1
2

3. q ( x , y , z )=3 x 2 +4 xy − y 2+ 8 xz−6 yz + z 2
⟹ The diagonal elements are 3 ,−1 and 1
The entries for a 12 and a 21=¿ half the coefficient of 4 xy ; 2 and 2
For a 31 and a 13=¿ half the coefficient of 8 xz ; 4∧4
For a 23 and a 32=¿ half the coefficient of −6 yz ;−3∧−3

[ ]
3 2 4
Therefore the required matrix A= 2 −1 −3
4 −3 1
4. Find the required matrix A , given that q ( x , y , z )=x 2−2 yz + xz
⟹ The diagonal entries are 1 , 0∧0
The entries for a 12 and a 21 are the half coefficients of 0
1
a 31anda 13 are the half coefficients of xz=
2
a 32anda 23 are the coefficients of yz=−1
 [ ]
1
1 0
2
⟹ The required matrix A= 0 0 −1
1
−1 0
2
5. Find a non singular matrix P such that Pt AP is diagonal for each of the following real
symmetric matrices A and also find its signature.

[ ] [ ]
1 −3 2 0 1 1
A= −3 7 −5 and A= 1 −2 2
2 −5 8 1 2 −1
HINT: Signature=No of positive entries−No of negative entries for a diagonal ¿
Answer: Signatureis 1

HERMITIAN FORMS

Let V be a vector space of finite dimension over a complex field C .

By letting f :V ×V → C s.t

(i) f ( a u1+ b u2 , v )=af ( u1 , v )+ bf (u2 , v ). Where a , b ∈ C and ui , v ∈V

Then f is referred to as the Hermitian form of V

(ii) f ( u , v )=f (u , v )

Remembering that K denote a complex conjugate of K ∈C

So from (i) and (ii) above

(iii) f ( u ,a v 1+ b v 2 ) =f (a v 1 +b v 2 , u)
¿ af ( v 1 , u ) + bf (v 2 , u)
¿ a f ( v 1 ,u ) +b f (v 2 , u)
¿ a f ( u , v 1 ) +b f (u , v 2)

Thus f ( u ,a v 1+ b v 2 ) =a f ( u , v 1 ) + b f (u , v 2)

Note

In condition (i) f is linear in the first variable and in condition (ii) f is conjugate linear in the second
variable.

By condition (ii) f ( v , v )=f ( v , v ) so that f (v , v ) is real for each and every vector v ∈ V
For instance Let A=(a¿¿ ij)¿ be an n × n matrix over C we can write A for the matrix obtained by
taking a complex conjugate of every entry of A i.e A=( a¿¿ ij)¿
¿
Also we can write A for A−t= A t

If A is Hermitian form then f ( X ,Y ) =X t A Y defines a Hermitian form on C n where f is defined by

t
f ( X ,Y ) =X A Y

Solution

∀ a , b ∈C and ∀ X 1 , X 2 ,Y ∈C n

We define
t
f ( a X 1 +b X 2 , Y )=(a X 1+b X 2) A Y
t t
¿(a X 1 +b X 2 ) A Y

By right hand distributive property

t t
⟹ a X 1 A Y +b X 2 A Y

¿ af ( X 1 , Y ) +bf (X 2 ,Y ) By definition

Hence f ( a X 1 +b X 2 , Y )=af ( X 1 ,Y ) + bf (X 2 , Y )

Hence f is linear in the first variable

Also f ( X ,Y ) =X t A Y

t t
¿( X A Y )
t t
¿Y A X
t ¿
¿Y A X

Hence f ( X ,Y ) =f (Y , X )

And f is Hermitian form on C n

Remember that X t A Y is a scalar so it is equal to its transpose.

Note: A real matrix is Hermitian if and only if it is symmetric (i.e the symmetric matrices A=A t )

Example
 ( )
2 2+3 i 4−5 i
1. Determine whether A= 2−3 i 5 6+2 i is Hermitian?
4+5 i 6−2i −7

Solution
¿
The condition matrix A=( aij ) is Hermitian if and only if A=A i.ea ij=a ji

( )
2 2+3 i 4−5 i
Since A= 2−3 i 5 6+2 i
4+5 i 6−2i −7

( )
2 2−3 i 4+5 i
Conjugate of A= 2+3 i 5 6+2 i
4−5i 6−2i −7

( )
2 2+3 i 4−5i
Transpose of conjugate 2−3i 5 6+2 i = A
4 +5 i 6−2i −7
¿
Since A = A . Then A is Hermitian (Since it is equal to its conjugate’s transpose)

( )
1
1+i 2i
2. Let H= 1−i 4 2−3i be a Hermitian matrix. Find a non singular such that Pt H P.
−2 i 2+3 i 7
Find its signature.
Solution

( | )
1 1+i 2i 1 0 0
Block matrix(HI)¿ 1−i 4 2−3 i 0 1 0
−2i 2+ 3i 7 0 0 1
Apply row operations R2 → R2 +(−1+i) R1
R3 → R 2+2 i R 1

( | )
1 1+i 2i 1 0 0
¿ 0 2 −5i −i+1 1 0
0 5i 3 2i 0 1
Now applying corresponding Hermitian column operations as
C 2 → (−1−i ) C 1 +C 2
C 3 →−2i C1 +C 3

( | )
1 0 0 1 0 0
Gives 0 2 −5 i −1+i 1 0
0 5i 3 2i 0 1
 Applying R3 →−5 i R2 +2 R3

( | )
1 0 0 1 0 0
0 2 −5 i −1+i 1 0
0 0 19 5+ 9 i −5 i 2

Applying Hermitian column operation, C 3 → 5 iC 2+2 C 3

( | )
1 0 0 1 0 0
0 2 0 −1+i 1 0
0 0 −38 5+9 i −5 i 2

( )
1 0 0
P= −1+i 1 0
5+9 i −5 i 2

( )
1 −1+i 5+ 9 i
t
⟹ H has been diagonalised so setting P = 0 1 −5 i
0 0 2

( )( )( )
1 −1+ i 5+9 i 1 1+i 2i 1 0 0
t
Then P H P= 0 1 −5i 1−i 4 2−3 i −1−i 1 0
0 0 2 −2i 2+ 3i 7 5−9 i 5 i 2

PRACTICE QUESTIONS

For each of the following Hermitian matrices find a non singular matrix P such that Pt H P is diagonal.

(a) H=(−i1 2i )
(b) H=(
2−3i −1 )
1 2+ 3i

[ ]
1i 2+ i
(c) H= −i 2 1−i
2−i 1+i 2

Find the rank and signature in each case.

HINT: Rank is the number of rows with nonzero entries.