UNIT-I

MATRICES
A matrix is defined as a rectangular array(or arrangement in rows or columns) of
scalars.
If mn numbers (real or complex) or functions are arranged in the form of a
rectangular array A having m rows and n columns then A is called an m × n matrix.
Each of the mn numbers is called an elements of the matrix. An m × n matrix is
also called a matrix of order m × n.

a
a12 a13
 11

 a21 a22 a23


An m × n matrix is usually written as  a31 a32 a33


 ...
...
...

am1 am2 am3
An m × n matrix A may be written as
A = [aij ]m×n

... a1n
...
...
...
...



a2n 


a3n 


... 

amn

or A = [aij ], where i = 1, 2, 3, ..., m; j = 1, 2, 3, ..., n.

APPLICATION OF MATRICES

In Algebra, the matrices have their largest application in the theory of simultaneous equations and linear transformation. e.g., the set of simultaneous equations
a11 x1 + a12 x2 + a13 x3 = b1
a21 x1 + a22 x2 + a23 x3 = b2
a31 x1 + a32 x2 + a33 x3 = b3
may be symbolically represented by the equation
AX = B
1

a11 a12 a13



where A =  a21 a22 a23

a31 a32 a33

x1

b1









 , X =  x 2  , B =  b2




x3
b3





The theory of matrices has been found of great utility in many branches of higher
mathematics such as algebraic and differential equations, astronomy, mechanics, theory of electrical circuits, quantum mechanics, nuclear physics and aerodynamics.

1.1 CHARACTERISTIC EQUATION

The characteristic equation is the equation which is solved to find a matrix’s
eigenvalues, also called the characteristic polynomial.
Let A = [aij ] be a square matrix of order n. the characteristic equation in
variable λ, is defined by
|A − λI| = 0.
The equation |A − λI| = 0 is called characteristic equation of matrix A.
In a 2 ×
2 case, where:

a11 a12

A=
a21 a22
then:


a11 − λ
a12

A − λI = 
a21
a22 − λ
so the characteristic equation is:
|A − λI| = (a11 − λ)(a22 − λ) − a21 a12 = 0
or simply:
|A − λI| = λ2 − (a11 + a22 )λ + (a11 a22 − a21 a12 ) = 0.

2

Solution: The characteristic equation of A is |A − λI| = 0.1.1 Find the characteristic equation of the matrix A =  5 4 1 2  . SOLVED EXAMPLES  Example.2 CHARACTERISTIC POLYNOMIAL The determinant |A − λI| when expanded will give a polynomial of degree n in λ which is called characteristic polynomial of matrix A. .

.

.

.

.

5−λ 4 .

.

. = 0.

.

.

.

1 2−λ .

Example. λ2 − 7λ + 6 = 0.2  Find the characteristic polynomial of the matrix A =  Solution: The characteristic polynomil of A is .

.

.

1−λ 2 |A − λI| = .

.

.

2 1−λ λ2 − 2λ − 3. . Exercise: 1 2 2 1  .

.

.

.

.

.

  2 −2 2     1.   1 3 −1   1 −1 1     2. Find the characteristic polynomial of the matrix A =  0 1 0  .   1 −1 1 3 . Find the characteristic equation of the matrix A =  1 1 1 .

we could take x2 = 1 or impose x1 +x2 = 1 or x21 + x22 = 1 as a normalization device. Corresponding to each characteristic root λ there corresponds non-zero vector X which satisfies the equation (A − λI)X = 0.3 EIGEN VALUES AND EIGEN VECTORS The roots λ1 .1. If we take the first eigenvalue λ1 then our equation system is (A − λ1 I)x = 0 For the 2 × 2 case. we can write this out in a systems of two equations: (a11 − λ1 )x1 + a12 x2 = 0 a21 x1 + (a22 − λ1 )x2 = 0 Although it might not be obvious. The non-zero vectors X are called eigen vectors or characteristic vectors. This x is the eigenvector associated with the eigenvalue λ1 . In an n-dimensional system. If we then took the second eigenvalue λ2 .e. e.g.λn of the characteristic equation are called eigen values or characteristic roots. we have to normalize the system. the first equation is a linear transformation of the second equation. λ2 . i.. From this we would thus obtain the vector x = [x1 x2 ]T . we would also find another eigenvector x associated with that by the same means.. 4 . −a12 /(a11 − λ1 ) = x1 /x2 = (a22 − λ1 )/a21 Once x1 /x2 is obtained. thus the ratio x1 /x2 will be the same regardless of how we solve it. we would have n eigenvalues with associated eigenvectors. the only thing that remains to obtain some levels of x1 and x2 .

Example. .1  Find the eigenvalues and eigenvectors of the matrix A =  1 1 3 −1  . Solution: The characteristic equation of A is |A − λI| = 0.

.

.

.

.

1−λ .

1 .

.

. = 0.

.

.

3 −1 − λ .

−2 i. The eigenvector corresponding to any λ is given by (A − λI)X = 0 . −2. λ = 2. λ2 − 4 = 0.e. The eigenvalues of A are 2.

.

  .

.

.

1−λ .

x1 1 .

 =0 i.e.

.

.

.

3 −1 − λ .

which are one and the same. Taking x1 = 1we get  x2 = 1. which are one and the same. 1  The eigenvector is  −3 5 . the eigenvector is given by the equations −x1 + x2 = 0 and 3x1 − 3x2 = 0. Solving. Taking x1 = 1 we get  x2 = −3. 1 The eigenvector is   1 Case(ii) When λ = −2. x1 = x2 . Solving. the eigenvector is given by the equations 3x1 + x2 = 0 and 3x1 + x2 = 0. 3x1 = −x2 . x2 Case(i) When λ = 2.

.2  1 1 3   Find the eigenvalues and eigenvectors of the matrix A  1 5 1  3 1 1 Solution: The characteristic equation of A is |A − λI| = 0.Example.

.

.

1−λ 1 3 .

.

.

1 5−λ 1 .

.

.

3 1 1−λ .

.

.

.

.

.

=0 .

.

.

we have x1 20 = x2 0 = x3 −20 Thus the eigenvector corresponding to λ = −2 is   1     X1 =  0    −1 6      . The characteristic equation of A is λ3 − 7λ2 + 36 = 0. 3.  x1        x2  = 0   x3 x1 + 7x2 + x3 = 0 3x1 + x2 + 3x3 = 0 Solving these equation by rule of cross-multiplication.e. Case(i) When λ = −2. The eigenvector is given by  3 1 3    1 7 1  3 1 3 i. 6. The eigenvalues of A are λ = −2.

The eigenvector is given by  −5 1 3    1 −1 1  3 1 −5 x1 1 =   x  1      x2  = 0   x3 x2 2 = x3 1 Thus the eigenvector corresponding to λ = 6 is   1     X3 =  2    1 7 .Case(ii) When λ = 3. The eigenvector is given by  −2 1 3    1 2 1  3 1 −2 x1 1 =   x  1      x2  = 0   x3 x2 −1 = x3 1 Thus the eigenvector corresponding to λ = 3 is   1     X2 =  −1    1 Case(iii) When λ = 6.

1/λn are the eigenvalues of the inverse matrix A−1 .. then (i) kλ1 .λpn are the eigenvalues of the matrix Ap . ..λn are the eigenvalues of a matrix A. The sum of the eigenvalues of a matrix A is equal to the sum of the principal diagonal elements of A (The sum of the principal diagonal elements is called the Trace of the matrix)..4 PROPERTIES OF EIGENVALUES: 1.. 6. If λ1 .... (ii) λp1 . kλ2 . A is non-singular. The eigenvalues of a real symmetric matrix are real. . 3. A square matrix A and its transpose AT have the same eigenvalues. The product of the eigenvalues of a matrix A is equal to |A|..1 Find the sum and product of the eigenvalues of  8 −4 2 2  . 2. . 1/λ2 .1. 5. 8 . where k is a non-zero scalar. λp2 . The eigenvectors corresponding to distinct eigenvalues of a real symmetric matrix are orthogonal. 4. Solution: Sum of the eigenvalues of a matrix A = Sum of the principal diagonal elements of A. Sum of the eigenvalues of a matrix A = 8+2=10. where p is a positive integer.kλn are the eigenvalues of the matrix kA. . (iii) 1/λ1 . SOLVED EXAMPLES  Example. provided λr 6= 0 i. λ2 .e.

using this theorem. + cn−1 λ + cn = 0 is the characteristic equation of a square matrix of order n. Solution: The product of the eigenvalues of a matrix A = |A|..A 6= 0.5 CAYLEY-HAMILTON THEOREM: Statement: Every square matrix satisfies its own characteristic equation. If c0 λn + c1 λn−1 + . If 3 and 15 are two eigenvalues of A =  −6 7 −4  . Find the eigenvalues   2 −4 3 2 of A − 5I and A . 1.   6 −2 2     Example. + cn−1 A + cn I = 0. Exercise:  1. then c0 An + c1 An−1 + .. Multiply (1) by A−1 . Find the eigenvalues of 2A2 if A =  4 1 3 2    without finding A2 . The product of the eigenvalues of a matrix A = 24.2 The product of two eigenvalues of the matrix  −2 3 −1  is 16.The product of the eigenvalues of a matrix A = |A|.. we can find A−1 as follows. 16λ3 =32. (1) Corollary: (i). λ3 =2.e. 9 .   2 −1 3 Find the third eigenvalue.. If A is non-singular.  8 −6 2     2. i.

.... (ii). −2    0  and hence  1 Solution: The characteristic equation of A is λ3 − 5λ2 + 9λ − 1 = 0. + cn−1 I + cn A−1 = 0. Example:  1 2   Verify Cayley-Hamilton theorem for the matrix A =  −1 3  0 −2 −1 find A .. Multiply (1) by A. A = −11 9 10    10 −22 −3 ∴ A3 − 5A2 + 9A − I= 0     −13 42 −2 5 −60 20 9 18 −18 −1 0 0               = −11 9 0 + 0 −1 0 10 + 20 −35 −10  + −9 27        10 −22 −3 −10 40 −5 0 −18 9 0 0 −1   0 0 0     = 0 0 0    0 0 0 10      . + cn−1 A2 + cn A = 0.c0 An−1 + c1 An−2 + . + cn−1 A2 + cn A]. 2 To verify A3 − 5A + 9A − I = 0.  −1 12 −4     A2 =A.A. A2 = −4 7 2    2 −8 1   −13 42 −2     3 2 3 A =A . c0 An+1 + c1 An + . The higher positive integral powers of A can be computed by the following way. + cn−1 I].. A−1 = − c1n [c0 An−1 + c1 An−2 + . An+1 = − c10 [c1 An + ..A..

to find the value of the matrix  given by  2 1 1     (A8 − 5A7 + 7A6 − 3A5 + A4 − 5A3 + 8A2 − 2A + I).  1 0 −2     2. when A =   5 3 1 3  . using Cayley-Hamilton theorem.   1 1 2 11 . 4. Find An . A −1 A2 − 5A + 9I − A−1 = 0 A−1 = A2 − 5A + 9I      −1 12 −4 −5 −10 10 9 0 0           =  −4 7 2  +  5 −15 0  +  0 9 0      2 −8 1 0 10 −5 0 0 9  ∴ A−1       3 2 6     =  1 1 2 . Find the characteristic equation of the A =  2 1 −1 . Verify Cayley-Hamilton theorem for the matrix A =  2 2 4  and hence   0 0 2 −1 4 find (i) A and (ii) A . if the matrix A =  0 1 0  . Hence find A4 .   2 2 5 Exercise:  1. Verify that the   1 −1 1 matrix satisfies its own characteristic equation. Also calculate (i) A−1 and (ii) A4 .To Find A−1 : A3 − 5A2 + 9A − I = 0 Multiply by A−1 . Use Cayley-Hamilton theorem.   1 0 3     3.

The transformation A to P −1 AP is called similarity transformation. where D is a diagonal matrix. Then the matrix B is said to be similar to the matrix A if there exists a non-singular matrix P such that B = P −1 AP . the a matrix M can be found such that M −1 AM = D. Proof: Let A and B be two similar matrices. is 12 . B − λI = P −1 AP − λI =P −1 AP − P −1 λIP =P −1 (A − λI)P |B − λI| = |P −1 ||A − λI||P | =|A − λI||P −1 P | =|A − λI||I| =|A − λI| Thus A and B have the same Characteristic polynomials and hence the same characteristic equations. ∴ A and B have the same eigenvalues.6 DIAGONALISATION OF A MATRIX: Similarity of Matrices: Let A and B be two square matrices of the same order.1. B = P −1 AP Then by definition. DIAGONALISATION OF A MATRIX BY MEANS OF A SIMILARITY TRANSFORMATION: A square matrix A of order n has n linearly independent Eigenvectors. Property: Two similar matrices have the same eigenvalues.

A square matrix A is orthogonal if and only if A−1 = AT . Note: Orthogonal transformation is possible only for a real symmetric matrix.A = I .A−1 = A−1 . If A is an orthogonal matrix then AT and A−1 are also an orthogonal matrix. The Normalized eigenvectors of A is formed by divide each element of the vector X.AT = AT . DIAGONALISATION OF A MATRIX BY ORTHOGONAL TRANSFORMATION (OR) ORTHOGONAL REDUCTION: If A is a real symmetric matrix. the matrices A and D are similar and hence A and D have the same eigenvalues. If A and B are orthogonal. 2. PROPERTIES OF ORTHOGONAL MATRIX: 1. Definition: A square matrix A is said to be orthogonal if A.A = I . then there exists an orthogonal matrix such that N T AN = N −1 AN = D is known as Orthogonal Reduction (or) Orthogonal Transformation. 3.A = I. by the square-root of the sum of the squares of all the elements of X.called diagonalisation of the matrix A. If A be a real symmetric matrix. Let N be the Normalized modal matrix whose columns are the normalized eigen vectors of A. −→(2) From (1) and (2) if A is orthogonal ⇒ AT = A−1 . As M −1 AM = D is a similarity transformation. −→(1) we know that A. then the eigen vectors of A are not only Linearly independent but also pairwise orthogonal. then AB is also orthogonal. Note: A is said to be orthogonal if A. 13 .AT = AT . Then N ia as orthogonal matrix and by the property N T = N −1 .

4.Example:  3 1 1      Diagonalize the matrix A  1 3 −1  by means of an orthogonal transformation.   c X3 is orthogonal to X1 ⇒ −a + b + c = 0 X3 is orthogonal to X2 ⇒  −b + c = 0 2     Solving we get X3 =  1    1 Now X1T X2 = X2T X3 = X3T X1 = 0    The Normalized Model matrix is N =   −1 √ 3 √1 3 √1 3 The orthogonal transformation is N T AN = D. The characteristic equation of A is λ3 − 9λ2 + 24λ − 16 = 0. Case(i) When λ = 1.   1 −1 3 Solution: The characteristic equation of A is |A − λI| = 0.   0     The eigenvector X2 is X2 =  −1    1   a     Let X3 =  b  be the third eigenvectors which is orthogonal to X1 and X2 . 14 0 −1 √ 2 √1 2 √2 6 √1 6 √1 6      . The eigenvalues of A are λ = 1.  −1      The eigenvector X1 is X1 =  1    1 Case(ii) When λ = 4. 4.

x21 + 2x22 + 3x23 + 4x1 x2 + x2 x3 is a Q. Diagonalize the matrix A  −2 2 3  by means of an orthogonal transformation. −1 √ 3   N T AN = D =  0  √2 6  √1 3 −1 √ 2 √1 6 √1 3 √1 2 √1 6  3 1 1        1 3 −1     1 −1 3 −1 √ 3 √1 3 √1 3 0 −1 √ 2 √1 2 √2 6 √1 6 √1 6      1 0 0     =  0 4 0  = D(1. Note 1: The matrix corresponding to the quadratic form is      co-effi x21 1 2 co-effi x2 x1 1 2 co-effi x3 x1 1 2 co-effi x1 x2 co-effi x22 1 2 co-effi x3 x2 1 2 co-effi x1 x3    x2 x3   co-effi x23 1 2 co-effi 15 . Exercise:  10 −2 −5      1.7 QUADRATIC FORM: A homogeneous polynomial of the second degree in any number of variable is called a quadratic form.   2 −1 3 1.F in the variable x1 .F in the variable x and y. x3 . x2 . x2 + y 2 + 2xy is a Q. 4)   0 0 4 . 2. 4.   −5 3 5   6 −2 2     2. Diagonalize the matrix A  −2 3 −1  by means of an orthogonal transformation. For examples: 1.

    . . x3 . + λn yn2 . . . . . 0    0    . .     .. where P is a non-singular matrix.    .Note 2:  x1      Quadratic form corresponding to the matrix A is X T AX where X =  x2    x3 CANONICAL FORM OF A QUADRATIC FORM: Let Q = X T AX be a quadratic form in n variables x1 .. NATURE OF THE QUADRATIC FORM: Index of the Quadratic Form: The number of positive terms in the canonical form is called the index and it is denoted by s.. 0 λ2 0 . y1    y2    . Q = (P Y )T A(P Y ) = Y T (P T AP )Y = Y T DY        = [y1 . . . ... . Now.    . . .   yn Q = λ1 y12 + λ2 y22 + .   0 λn . . . Which is the Canonical form of the given Quadratic Form. 0 .. x2 . .. y2 . . 16 . xn . yn ]        λ1 0 . . . . . . Let X = P Y be a linear transformation..

Signature of the Quadratic Form: The difference of number of positive and negative square terms is called the signature of the Quadratic Form. if all the eigenvalues are negative numbers. • Indefinite in all other cases. • Negative semi definite. Rules for finding nature of Quadratic form using principal sub-determinants: In this method we can determine the nature of the quadratic form without reducing it to the canonical form. Let A be square matrix of order n. • Positive semi definite. if all the eigenvalues os A ≥ 0 and atleast one eigenvalue is zero. • Negative definite. if all the eigenvalues os A ≤ 0 and atleast one eigenvalue is zero. The Quadratic Form Q = X T AX in n variables is said to be • Positive definite. if all the eigenvalues are positive numbers. D1 = |a11 | .

.

.

a11 D2 = .

.

.

a21 .

.

.

a11 .

.

D3 = .

a21 .

.

.

Dn = |A|. a31 . 17 . . .

.

a12 .

.

.

a22 .

.

.

a12 a13 .

.

.

a22 a23 .

.

.

a32 a33 .

.

. • A Quadratic form is said to be Negative semi definite.. Dn are called the principal sub determinants of A. if D1 . . if D1 .e) Dn > 0 for all n.. • A Quadratic form is said to be Positive semi definite. if Dn ≥ 0 and atleast one Di = 0.. The characteristic equation of A is λ3 − 10λ2 + 12λ + 72 = 0.. D6 .. if (−1)n Dn ≥ 0 and atleast one Di = 0. Dn are all positive. in all other cases. The eigenvalues of A are λ = −2. D3 . (i. (i.e) (−1)n Dn > 0 for all n. D4 . D3 . Example: Reduce the quadratic form 2x21 + 6x22 + 2x23 + 8x1 x3 to canonical form by orthogonal reduction. D2 .. • A Quadratic form is said to be Positive definite. are all negative and D2 . 6. 6. are all positive. Case(i) When λ = −2... Solution: The matrix of quadratic form is   2 0 4     A= 0 6 0    4 0 2 The characteristic equation of A is |A − λI| = 0. D5 . .  1      The eigenvector X1 is X1 =  0    −1 18 .. .Here D1 . • A Quadratic form is said to be Indefinite.. D2 . D3 . • A Quadratic form is said to be Negative definite.

Case(ii) When λ = 6.  1      The eigenvector X2 is X2 =  0    1   a     Let X3 =  b  be the third eigenvectors which is orthogonal to X1 and X2 . which is of the form Q = Y T DY 19 . 4)   0 0 6 .   c X3 is orthogonal to X1 ⇒ a − c = 0 X3 is orthogonal to X2 ⇒  a+c=0 0     Solving we get X3 =  1    0 Now X1T X2 = X2T X3 = X3T X1 = 0  √1 2   The Normalised Model matrix is N =  0  −1 √ 2 The orthogonal transformationisN T AN −1 √1 0 √ 2 0 2   2   N T AN = D =  √1 0 √1   0 6 2   2 4 0 0 1 0   −2 0 0     =  0 6 0  = D(1. 4. √1 2 0 √1 2 = D.  4 0    1   0 √1 2   0  0  −1 √ 2 2 √1 2 0 √1 2 0    1   0 The orthogonal transformation X = N Y reduces quadratic form to the canonical form.

Reduce 2x21 + 2x22 + 3x23 + 2x1 x2 − 4x2 x3 − 4x3 x1 into canonical form by an orthogonal reduction and find the rank. y3 )  0  0 2 2 = −2y1 + 6y2 + y32  y1      6 0   y2    0 6 y3 Therefore the Quadratic form is indefinite in nature. since canonical form contains both positive and negative terms. Reduce 8x2 + 7y 2 + 3z 2 − 12xy + 4xz − 8yz into canonical form by orthogonal reduction. y2 . 3. 2. Reduce 6x21 + 3x22 + 3x23 − 4x1 x2 − 2x2 x3 + 4x3 x1 into canonical form by an orthogonal reduction and find the rank. index. index. Exercise: 1. signature and the nature of the quadratic form. −2 0 0   = (y1 . 20 . signature and the nature of the quadratic form.