Quadratic forms

Definition: A homogeneous polynomial of the second degree in any

number of variables is called quadratic forms.

Examples: 3𝑥12 + 4𝑥1 𝑥2 + 𝑥22 , 𝑥12 + 𝑥22 − 𝑥32 + 2𝑥1 𝑥2 − 4𝑥1 𝑥3 + 𝑥3

are quadratic forms in 2 and 3 variables respectively .
The general quadratic form may be written as
𝑛 𝑛

𝑞 = 𝑋 / 𝐴𝑋 = ∑ ∑ 𝑎𝑖𝑗 𝑥𝑖 𝑥𝑗
𝑖=1 𝑗=1

The coefficients aij are the elements of the symmetric matrix

A = [𝑎𝑖𝑗 ] The coefficients of the square terms which arise when i =j are
𝑎11 , 𝑎22 , 𝑎33 , − − 𝑎𝑛𝑛 respectively. The coefficients of product terms
1
𝑥𝑖 𝑥𝑗 which arise when 𝑖 ≠ 𝑗 are 𝑎𝑖𝑗 = 𝑎𝑗𝑖 = (coefficients of 𝑥𝑖 𝑥𝑗 .)
2

The symmetric matrix A is called matrix of quadratic form q.

The rank 𝜌(𝐴) = 𝑟 is called rank of quadratic form q.
The determinant of A is called determinant of quadratic form q.
If rank 𝜌(𝐴) is r=n (number of variables) , the quadratic form q is
called non-singular and if rank 𝜌(𝐴) is r <n the quadratic form q is
called singular..
Linear transformation of quadratic form
Congruent matrix:
The linear transformation 𝑋 = 𝑃𝑌 , where P is a non-singular matrix
carries the quadratic form 𝑞 = 𝑋 / 𝐴𝑋 with symmetric matrix A into the
quadratic form
 𝑋 / 𝐴𝑋 = (𝑃𝑌)/ 𝐴𝑃𝑌 = 𝑌 / 𝑃/ 𝐴𝑃𝑌 = 𝑌 / (𝑃/ 𝐴𝑃)𝑌 = 𝑌 / 𝐵 𝑌,
where 𝐵 = 𝑃/ A𝑃 is symmetric matrix. Two square matrices A and
B are called congruent if there exists a nonsingular matrix P such
that 𝐵 = 𝑃/ A𝑃

The form 𝑌 / 𝐵 𝑌 is called a linear transformation of 𝑋 / 𝐴𝑋. A

quadratic form

𝑞 = 𝑋 / 𝐴𝑋

= 𝑎11 𝑥12 + − − +𝑎𝑟𝑟 𝑥𝑟2 + − − + 𝑎12 𝑥1 𝑥2 + − −

of rank r can be reduced by a non-singular linear transformation 𝑋 = 𝑃𝑌
to the canonical form (B is diagonal)

𝑌 / 𝐵 𝑌 = ℎ1 𝑦12 + ℎ2 𝑦22 + ℎ3 𝑦32 + − − +ℎ𝑟 𝑦𝑟2

Canonical form: B is diagonal matrix , then the transform quadratic is a
sum of square terms , knows as canonical form
Index: The number of positive terms (p) in canonical form of a
quadratic form is called index of the form q.
Signature: The difference of positive terms (p) and negative terms ( r-p)
i,e p-(r-p)=2p-r is called signature of the quadratic form q.
Elementary congruent transformations:
A and B are congruent if A is transformed into B by a sequence of pairs
elementary transformations where each pair involves an elementary row
transformation followed by the same elementary column
transformation.
Example; Reduce the quadratic form

𝑞 = 𝑥12 + 2𝑥22 + 𝑥32 + 2 𝑥1 𝑥2 + 4𝑥1 𝑥3 + 6𝑥2 𝑥3 to the canonical form

and find rank, index and signature of the form.
Solution:
The matrix of the quadratic form is
1 1 2
𝐴 = [1 2 3]
2 3 1
Write A=I/AI
1 1 2 1 0 0 1 1 2 1 0 0
[1 2 3] = [0 1 0] [1 2 3] [0 1 0]
2 3 1 0 0 1 2 3 1 0 0 1
Applying elementary congruent transformation, Performing R21 (-1),
R31 (-2), we get
1 1 2 1 0 0 1 1 2 1 0 0
[0 1 1 ] = [−1 1 0] [1 2 3] [0 1 0]
0 1 −3 −2 0 1 2 3 1 0 0 1
Performing C21 (-1), C13 (-2), we get

1 0 0 1 0 0 1 1 2 1 −1 −2
[0 1 1 ] = [−1 1 0] [[1 2 3]] [0 1 0]
0 1 −3 −2 0 1 2 3 1 0 0 1

Performing R32 (-1), we get

 1 0 0 1 0 0 1 1 2 1 −1 −2
[0 1 1 ] = [−1 1 0] [[1 2 3]] [0 1 0]
0 0 −4 −1 −1 1 2 3 1 0 0 1

Performing C32 (-1), we get

1 0 0 1 0 0 1 1 2 1 −1 −1
[0 1 0 ] = [−1 1 0] [[1 2 3]] [0 1 −1]
0 0 −4 −1 −1 1 2 3 1 0 0 1
1
Performing , 𝑅3 ( ) , we get
2

1 0 0
[0 1 0]=
0 0 −2
1 0 0 1 1 2 1 −1 −1
[ −1 1 0 ] [[1 2 3]] [0 1 −1]
−1/2 −1/2 1/2 2 3 1 0 0 1
1
Performing , 𝐶3 ( ) , we get
2

1 0 0 1 0 0 1 1 2 1 −1 −1/2
[0 1 0 ] = [ −1 1 0 ] [[1 2 3]] [0 1 −1/2]
0 0 −1 −1/2 −1/2 1/2 2 3 1 0 0 1/2

Which is the form diag(1, 1, -1). Thus the given quadratic q reduced to
the canonical form
1 0 0 𝑤1
[𝑤1 𝑤2 𝑤3 ] [0 1 0] [ 𝑤2 ]
0 0 −1 𝑤3
 𝑤1
= [𝑤1 𝑤2 𝑤3 ] [ 𝑤2 ]
−𝑤3
= 𝑤12 + 𝑤22 − 𝑤32
So the rank = r= 3, index = p=2 and signature =2p-r=4-3=1

System of linear equations

System of first order linear equations are

dx1
= a11 (t ) x1 + a12 (t ) x2 + − − −a1n (t ) xn + f1 (t )
dt
dx2
= a21 (t ) x1 + a122 (t ) x2 + − − −a2n (t ) xn + f 2 (t )
dt

dxn
= an1 (t ) x1 + an2 (t ) x2 + − − −ann (t ) xn + f n (t )
dt

When fi (t ) = 0 , the system is said to be homogeneous ,

otherwise it is nonhomogeneous.
 Matrix form of a system .

If X , A(t ) and F (t ) denotes the respective matrices

 x1 (t )   f1 (t ) 
 x (t )  f (t )
 2   2 
X = − −  F (t ) = − − 
− −  ,
− − 
   
 xn (t )  f n (t )

a11 (t ) a11 (t ) − − a11 (t ) 

a (t ) a21 (t ) − − a21 (t ) 
 21 
A(t ) = − − −− − − −− 
− − −− − − − − 

an1 (t ) an1 (t ) − − ann (t )

Then the system of first order linear form can be written as

  x1  a11 (t ) a11 (t ) − − a11 (t )   x1   f1 (t ) 
 x  a (t ) a21 (t ) − − a21 (t )   x2   f 2 (t )
d  2  21    
−  = − − −− − − − −  −  +  − − 
dt   
−  − − −− − − − −  −  − − 
 xn  an1 (t ) an1 (t ) − − ann (t )  xn   f n (t )

Or X / = AX + F (t ) -------(1)

If the system is homogeneous , its matrix fomr is then

X / = AX -------(2)

Solution vector: A solution vector on an interval is any

column matrix

 x1 (t ) 
 x (t )
 2 
X = − − 
− − 
 
 xn (t )
Whose entities are differentiable functions satisfying (1)
on the interval.

General solution of homogeneous system

Let 1, 2 3 − − − n be a n distinct eigenvalues of the

coefficient matrix A of the homogeneous system (2) and
let K1, K2 K3 − − − Kn be the corresponding eigenvectors

and X1 = K1e t , X 2 = K2e t ,− − − X n = Kne t are the

1 2 n

fundamental solution of (2) . Then general solution of

the system on the interval (−, ) is given by
X = c1X1 + c2 X 2 + c3 X3 − − − +cn X n ,

= c1K1e1t + c2 K2e2t + − − − + cn Knent , where

c1 , c2 , c3 − − − cn are arbitrary constants.

General solution of non-homogeneous system

The general solution of the non-homogeneous system

on the interval (−, ) is X = Xc + X p
Where X c = c1X1 + c2 X 2 + c3 X3 − − − +cn X n is the
complementary function and

X p = G(t ) G −1 (t ) F (t )dt is the particular integral of

non-homogeneous system respectively. Here

X1 = K1e1t , X 2 = K2e2t ,− − − X n = Knent are the

fundamental solution of (2), K1, K2 K3 − − − Kn

eigenvectors corresponding to eigenvalues
1, 2 3 − − − n of the coefficient matrix A and

G(t ) is the n  n matrix whose columns consist of the

entities of the solution vectors X1, X 2 ,− − − X n .
Example; Using matrix method to find the currents
𝑖1 (𝑡) and 𝑖2 (𝑡) in a electrical networks shown in Fig.
given below

where R1= 24 , , R2= 9 , , L1 =3 H , L2 = 3 H and E=300

V and the currents 𝑖1 (𝑡) and 𝑖2 (𝑡) are initially zero

Solution:
The given system
𝑑𝑖1
𝐿2 = −(𝑅1 + 𝑅2 )𝑖1 + 𝑅2 𝑖2 + 𝐸
𝑑𝑡
𝑑𝑖2
𝐿1 = 𝑅2 𝑖1 − 𝑅2 𝑖2
𝑑𝑡

Putting the values of R1= 24 , R2= 9 , L1 = 3 , L2 = 3 and

E=300 in above equations , we get
𝑑𝑖
3 1 = −33𝑖1 + 9 𝑖2 + 300
𝑑𝑡
 𝑑𝑖1
=> = −11𝑖1 + 3 𝑖2 + 100 ---(1)
𝑑𝑡
𝑑𝑖2
3 = 9𝑖1 − 9𝑖2
𝑑𝑡
𝑑𝑖2
=> = 3𝑖1 − 3𝑖2 --- (2)
𝑑𝑡

subject to 𝑖1 (0) = 𝑖2 (0) = 0 --(3)

Matrix form of the system is
𝐼/ = 𝐴𝐼 + 𝐹(𝑡) ----(4) , where

𝑖1 (𝑡) −11 3 100

𝐼=[ ],𝐴 = [ ] 𝐹(𝑡) = [ ]
𝑖2 (𝑡) 3 −3 0

Now, the characteristic equation is ,

|𝐴 − 𝐼𝜆| = |−11 − 𝜆 3
|=0
3 −3 − 𝜆

⇒ 𝜆2 + 14𝜆 + 24 = 0 ⇒ 𝜆 = −12, −2

Eigenvector 𝐾1 corresponding to eigen value 𝜆 = −2

−9 3 𝑘1 0
[ ] [ ] = [ ] ⇒ −9𝑘1 + 3𝑘2 = 0 => 3𝑘1 −
3 −1 𝑘2 0
𝑘2 = 0
 𝑘2
Thus 𝑘1 = , when 𝑘2 = 3 , then 𝑘1 = 1 and hence
3
1
eigenvector is 𝐾1 = [ ] and hence the solution vector
3
is

1 −2𝑡 −2𝑡
𝑋1 = 𝐾1 𝑒 −2𝑡
= [ ]𝑒 𝑒
= [ −2𝑡 ]
3 3𝑒

Eigenvector 𝐾2 corresponding to eigen vector 𝜆 = −12

1 3 𝑘1 0
[ ][ ] = [ ]
3 9 𝑘2 0

⇒ 𝑘1 + 3𝑘2 = 0

3𝑘1 + 9𝑘2 = 0

Thus 𝑘1 = −3𝑘2 , when 𝑘2 = −1,then 𝑘1 = 3 ,

3
therefore the eigenvector is 𝐾2 = [ ] and hence the
−1
3
solution vector is 𝑋2 = 𝐾2 𝑒 −12𝑡 = [ ] 𝑒 −12𝑡 =
−1
−12𝑡
3𝑒
[ −12𝑡 ]
−𝑒

The complementary function is

 1 3
𝐼𝑐 = 𝑐1 𝑋1 + 𝑐2 𝑋2 = 𝑐1 [ ] 𝑒 −2𝑡 + 𝑐2 [ ] 𝑒 −12𝑡
3 −1
−2𝑡
We get, 𝑒
𝐺(𝑡) = [ −2𝑡 3𝑒 −12𝑡 ]
3𝑒 −𝑒 −12𝑡
−12𝑡
∴𝐺 −1
(𝑡) = −
𝑒 14𝑡
[ −𝑒 −3𝑒 −12𝑡 ] =
10 −3𝑒 −2𝑡 𝑒 −2𝑡
2𝑡
1
[ 𝑒 3𝑒 2𝑡 ]
10 3𝑒 12𝑡 −𝑒12𝑡

Therefore the particular integral is

𝐼𝑝 = 𝐺(𝑡) ∫ 𝐺 −1 (𝑡)𝐹(𝑡)𝑑𝑡

−2𝑡
𝑒
= [ −2𝑡 3𝑒 −12𝑡 ] × 1 𝑒 2𝑡
3𝑒 2𝑡 ] [100] 𝑑𝑡
−12𝑡 ∫ 10 [ 12𝑡
3𝑒 −𝑒 3𝑒 −𝑒 12𝑡 0
−2𝑡
𝑒
= 10 [ −2𝑡 3𝑒 −12𝑡 ] [ 𝑒 2𝑡 ] 𝑑𝑡
∫
3𝑒 −𝑒 −12𝑡 3𝑒 12𝑡

𝑒 2𝑡
−2𝑡
𝑒
= 10 [ −2𝑡 3𝑒 −12𝑡 ] 2
3𝑒 −𝑒 −12𝑡 3𝑒 12𝑡
[ 12 ]
 1 9 5 25
+
= 10 [23 12
3] = 10 [45] =[25
2
]
−
2 12 4 2

The general solution is

25
𝐼1 (𝑡) 1 3 2
⇒[ ] = 𝑐1 [ ] 𝑒 −2𝑡 + 𝑐2 [ ] 𝑒 −12𝑡 + [25 ]
𝐼2 (𝑡) 3 −1
2

25
0 1 3 2
From (3), we get [ ] = 𝑐1 [ ] + 𝑐2 [ ] + [25 ]
0 3 −1
2
25
⇒ 𝑐1 + 3𝑐2 + = 0 − − − (5)
2
25
3𝑐1 − 𝑐2 + = 0 − − − (6)
2
5
From (5) and (6), 𝑐1 = −5 and 𝑐2 = −
2

Hence the general solution is

25
𝐼1 (𝑡) 1 5 3 2
[ ] = −5 [ ] 𝑒 −2𝑡 − [ ] 𝑒 −12𝑡 + [25 ] Ans
𝐼2 (𝑡) 3 2 −1
2