1 Linear Algebra

Objective é ____________________ ù
A = [aij ]m´n = êê ú
ú
Upon completion of this chapter, you will be a a
ëê m1 m2 ___________ a mn ûú
able to determine:
Horizontal lines are called rows and vertical
y Determinant of a matrix lines are called columns.
y Adjoint and inverse of a matrix Types of matrices:
y Rank of a matrix
The square matrix: The matrix in which the
y Solution of homogeneous and
number of rows is same as the number of
non-homogeneous linear equations.
columns is called as square matrix. If it has
y Eigen values and eigen vectors of a matrix
n rows, then it is called as n-rowed square
y Higher powers of matrix using Cayley
matrix. The elements aij/i=j i.e., a11, a22 ... are
Hamilton theorem
called diagonal elements and the line along
Introduction which the elements lie is called principal
diagonal of matrix. Elements other than a11,
Linear algebra is a branch of mathematics a22, etc. are called off – diagonal elements
concerned with the study of vectors, with i.e. aij/i=j.
families of vectors called vector spaces or
é 1 9 4ù
linear spaces and with functions that input ê ú
one vector and output another, according Example: A = êê7 2 8 úú is a square matrix
to certain rules. These functions are called ê 10 78 45 ú
ëê ûú 3´3
linear maps or linear transformations and The diagonal elements of this matrix are {1,
are often represented by matrices. 2, 45}
Matrices are rectangular arrays of numbers or
Note:
symbols, and matrix algebra or linear algebra
provides the rules defining the operations A square sub-matrix of a square matrix A is
that can be formed on such an object. An called a “principle sub-matrix” if its diagonal
elementary application of linear algebra is to elements are also the diagonal elements
é1 9ù
the solution of a system of linear equations in the matrix A. So êê ú is a principle sub
ú
several unknowns, which often result when êë7 2úû
linear mathematical models are constructed matrix of the matrix A given above, but
to represent physical problems. é9 4 ùú
ê is not.
Matrix ê2
êë 8úúû
A system of mn numbers arranged in the
form of a rectangular array having m rows Diagonal matrix: A square matrix in which
and n columns is called a matrix of order all off-diagonal elements are zero is called a
m×n diagonal matrix. The diagonal elements may
or may not be zero.
If A = [aij]m×n be any matrix of order m×n then
é 11 0 0 ù
it is written in the form: ê ú
Example: A = êê0 15 0 úú is a diagonal matrix
éa a ______________a ù ê0 0 89ú
ê 11 12 1n ú
ëê ûú
êa a ______________a ú
ëê 21 22 2n ûú The above matrix can also be written as
A = diag [11 , 15 ,89]

Linear Algebra 5
 Properties of diagonal matrix: Upper triangular matrix: An upper triangular
matrix is a square matrix whose lower
diag [x, y, z] + diag [p, q, r] = diag [x + p, y +
q, z + r] off-diagonal elements are zero, i.e. aij = 0
diag [x, y, z] × diag [p, q, r] = diag [xp, yq, zr] whenever i > j
é ù It is denoted by U. The diagonal and upper
(diag êëéx, y, zúûù ) = diag êê x1 , y1 , z1 úú
-1
off-diagonal elements may or may not be
ë û zero.
é ù
t
(diag éêëx, y, zùúû ) = diag éêëx, y, zùúû ê79 45 - 76 ú
Example : U = êê 0 54 87úú
n ê 0 0 0 úúû
(diag [x, y, z]) = diag éêëxn , yn , zn ùúû êë
Eigen values of diag [x, y, z] = x, y andz Lower triangular matrix: A lower triangular
determinant of diag [x, y, z] = [x, y, z] = xyz matrix is a square matrix whose upper off-
diagonal triangular elements are zero. i.e.
Scalar matrix: A scalar matrix is a diagonal
aij = 0 whenever i < j . The diagonal and lower
matrix with all diagonal elements being
equal. off-diagonal elements may or may not be
é5 zero. It is denoted by L.
ê 0 0ùú
é 17 0 0 ù
Example : A = êê0 5 0úú is a scalar matrix ê ú
ê0 Example : L êê85 0 0 úú
êë 0 5úúû
ê42 34 62ú
ëê ûú
Unit matrix or identity matrix: A diagonal
matrix whose each diagonal element is 1 is Idempotent matrix: A matrix A is called
called as identity matrix. Idempotent if and only if A2 = A .
Thus a square matrix A = [aij] is a unit matrix é 2 - 2 - 4ù
if aij = 1 when i = j and aij = 0 when i ≠ j é 1 0ù é0 0ù ê ú
Example : êê úú , êê ú , ê-1 3 4 ú
ú ê ú
é 1 0 0ù é1 0ùú ëê0 1ûú ëê0 0ûú ê 1 - 2 - 3ú
ê ú ê êë úû
Example : I3 = êê0 1 0úú ,I2 = êê ú
ú
ê0 0 1ú ê0 are examples of idempotent matrices.
ëê úû ëê 1úûú
Involutory matrix: A matrix A is called
Properties of identity matrix: involutory if and only if A2 = I .
1. AI = IA = A é 1 0ù
2. In = I Example : êê úú isInvolutory
ëê0 1ûú
3. I-1 = I
4. I = 1 é 4 3 3ùú
ê
Null matrix: The m×n matrix whose elements Also êê-1 0 - 1 úú isInvolutory since A 2 = I.
are all zero is called null matrix. Null matrix ê-4 - 4 - 3ú
is Denoted by null matrix need not be square. êë úû
é0 0ù é0ù Nilpotent matrix: A matrix A is said to be
é0 0 0ù ê ú ê ú nilpotent of class x or index x if Ax = O and Ax-1
Example : O3 = êê ú ,O = ê ú ,O = ê ú
ú 2 ê ú 2´1 ê ú ≠ O i.e. x is the smallest index which makes
ëê0 0 0ûú ê0 0ú ê0ú
ëê ûú ëê ûú Ax = O
Properties of null matrix:
Matrix Algebra
1. A + O = O + A = A
So, O is additive identity. Equality of matrices:
if
2. A + (-A) = O Two matrices A = [aij] and B = [bij] are said to
be equal if.
1. They are of same size.

6 Linear Algebra
2. The elements in the two corresponding obtained by multiplying every element of the
places of two matrices are the same i.e., matrix A by k is called scalar multiple and is
aij = bij for each pair of subscripts i and j. denoted by kA.
éx y ù é42 55ù If A = [aij]m×n then Ak = kA = [kA]m×n
Example : Let êê ú=ê
ú ê
ú
ú
ëêz w ûú ëê ûú é5 6 1 ùú é20 24 4 ùú
ê ê
⇒ x = 42 , y = 55 , p = 31 and q = 19 If A = êê2 - 5 3 úú then, 4A = êê8 - 24 12úú
ê1 1 6úûú ê4 4 24 úûú
Addition of matrices: ëê ëê
Two matrices A and B are compatible for Properties of multiplication of a matrix by a
addition only if they both have the same size scalar:
say m×n . Then their sum is defined to be the 1. Scalar multiplication of matrices distributes
matrix of the type m×n obtained by adding over the addition of matrices i.e., k (A + B)
corresponding elements of A and B = kA + kB
i.e. adding the elements that lie in the same 2. If p and q are two scalars and A is any m×n
position. matrix then, (p + q) A = pA + qA
Thus if, A = [aij]m×n & B = [bij]m×n then A + B = [aij 3. If p and q are two matrices and A = [aij]-m×n
+ bij]m×n then, p (qA) = (pq) A
é1 2ùú é4 6ù 4. If A = [aij]-m×n be a matrix and k be any
Example : A = êê B = ê ú
ëê3 5 úûú ê7 8ú
ëê ûú
scalar then, (-k) A = -(kA) = k (-A)
é 1 2 ù é4 6ú ê 5 8 ùú
ù é Multiplication of two matrices:
A + B = êê ú+ê =
ú ê
êë3 5úû êë7 - 8 úúû êêë 10 13úúû Let A = (aij)-m×n; B = [bjk]-n×p be two matrices
such that the number of columns in A is
Properties of matrix addition: equal to the number of rows in B.
1. Matrix addition is commutative A+B = B+A Then the matrix C = [Cik]-m×p such that Cik =
2. M
 atrix addition is associative (A+B) + C = n

A + (B+C) åa b
j= 1
ij jk is called the product of matrices A
3. C
 ancellation laws holds good in case of and B.
addition of matrices, which is X = -A Properties of matrix multiplication:
A+ X =B + X ⇒A=B 1. M
 ultiplication of matrices is not
X +A= X +B ⇒A= B commutative. If AB exits, then it is not
4. The equation A + X = 0 has a unique guaranteed that BA also exists.
solution in the set of all m×n matrices. For example, A2×2 × B2×7 = C2×7 but B2×7 ×
Subtraction of two matrices: A2×2 does not exist since these are not
If A and B are two m×n matrices, then we compatible for multiplication.
define, A - B = A + (B) 2. M
 atrix multiplication is associative. i.e., A
Thus, the difference A - B is obtained (BC) = (AB)C where A, B , C are m×n, n×p,
by subtracting from each element of A p×q matrices respectively.
corresponding elements of B. 3. M
 ultiplications of matrices are distributive
i.e., A (B + C) = AB + AC
Note:
4. The equation AB = O does not necessarily
Subtraction of matrices is neither commu­
imply that at least one of matrices A and
tative nor associative
B must be a zero matrix
Multiplication of a matrix by a scalar: é 1 1ù é 1 1ùú éê0 0ùú
For example, êê úê
ú=ê
Let A be any m×n matrix and k be any úê ú
êë 1 1úû êë-1 - 1úû êë0 0úû
real number called scalar. The m×n matrix

Linear Algebra 7
 5. AB = AC ⇒ B = C (iff A is non-singular taking the conjugate of each element of the
matrix) matrix. For real matrices,
6. BA = CA ⇒ B = C (iff A is non – singular Conjugate will yield the same.
matrix) é6 8ù
ê ú é6 3 2ù
Trace of a matrix: Let A = êê3 4úú then, A T = A = êê ú
8 4 1 ú
The sum of the diagonal elements of a square ê2 1ú ëê ûú
êë úû
matrix is termed as trace of a matrix.
n If B = éëê 1 2 3ùûú then
Thus if A = [aij]-m then, Tr (A) = åa ji
= a11 + a22
é2 + 3i 4 - 7i 8 ùú
+ ... ann xn j= 1
Example : If A = êê
é 11 êë - i 6 9 + iúúû
ê 2 5ùú
é2 - 3i 4 + 7i 8 úù
Let A = êê-2 - 23 1 úú A êê
ê-1 6 35 úûú êë i 6 9 - iúúû
ëê
Properties of conjugate of a matrix:
Then, trace (A) = tr (A) = 11 + (-23) + 35 = 23
If A & B be the conjugates of A & B respectively.
Properties of trace of a matrix:
Then
Let A and B be two square matrices of order
1. (A ) = A
n and k be a scalar. Then,
1. tr (kA) = k tr A 2. (A + B) = A + B
2. tr (A + B) = tr A + trB 3. (KA ) = K A , k being any complex number
3. tr (AB) = tr (BA)
4. (AB) = AB , if A & B can be multiplied
Transpose of a matrix:
5. A = A if and only if A is real matrix
Let A = [aij]-m×n . Then the n×m matrix obtained
from A by changing its rows into columns and 6. A = -A if and only if A is purely imaginary
its columns into rows is called the transpose matrix
of A and is denoted by A′ or AT .
é6 8ù Transposed conjugate of matrix:
ê ú é6 3 2ù The transpose of the conjugate of a matrix
Let A = êê3 4úú then, A T = A ¢ = êê ú
ú
8 4 1 A is called transposed conjugate of A and is
ê2 1ú ëê ûú
ëê ûú denoted by
If B = éëê 1 2 3ùûú then T
Aθ or A* or (A ) . It is also called conjugate
é1 ù
¢ t ê ú transpose of A.
B¢ = éêë 1 2 3ùúû = éêë 1 2 3ùúû = êê2úú
é2 + i 3 - iùú
ê ú Example : If A = êê
ëê3ûú 1 - iúúû
êë 4
Properties of Transpose of a matrix:
é2 - i 3 + iùú
If AT and BT be transposes of A and B To find Aθ , we first find A = êê
respectively then, ëê 4 1 + iúûú
1. (AT)T = A T é2 - i 4ù
Then A = (A ) = ê ú
2. (A + B) = A + B
T T T
ê3 ú
ë û
3. (kA)T = kAT . k being any complex number Properties:
4. (AB)T = BTAT If Aθ & Bθ be the transposed conjugates of A
5. (ABC)T = CTBTAT and B respectively then,
Conjugate of a matrix: 1. (Aθ)θ = A
The conjugate of a matrix is obtained by 2. (A + B)θ Aθ + Bθ

8 Linear Algebra
 
3. (KA ) = KA, K ® complex number A complex matrix can be resolved into
Hermitian and skew hermitian matrix,
4. (AB)θ = Bθ Aθ
Classification of Real Matrices Symmetric A + A A - A
A= +
Matrices: 2 2
If the transpose of a matrix is identical to the Here, the first term represents a Hermitian
original matrix, then the matrix is called as matrix, and the second term represents
symmetric matrix. Skew Hermitian Matrix.
AT = A Unitary matrix:
If the product of matrix and its conjugated
A + AT transpose is the identity matrix. Thus,
Note: AAT and are always symmetric
matrices. 2 inverse of matrix is identical to its conjugated
transpose of a matrix.
Skew symmetric matrices: Aθ = A-1 or AAθ = I
If the transpose of a matrix is negative of the Let a11, a12, a21, a22 be any four numbers. The
original matrix, then the matrix is termed as
a skew symmetric matrix. a 11 a 12
symbol  = represents the number
AT = -A a 21 a 22
The diagonal elements of a skew symmetric a11 a22 - a12 a21 and is called determinants of
matrix are always zero. order 2.
Here, the first term represents a symmetric a 11 a 12 a 13
matrix, and the second term represents skew Similarly,  = a a 22 a 23 represent a
21
symmetric matrix. a 31 a 32 a 33
Orthogonal matrices:
determinant of order 3.
If the product of a matrix and its transpose
Minors and cofactors: a 11 a 12 a 13
is equal to the identity matrix or the inverse
of a matrix is identical to its transpose, then Consider the determinant a 21 a 22 a 23
the matrix is termed as orthogonal matrix. a 31 a 32 a 33
AT = A-1 or AAT = I Leaving the row and column passing through
the elements aij , then the second order
AA T = I = 1
determinant thus obtained is called the
2
Thus, A A T = ( A ) = 1 minor of element aij and we will be denoted
by Mij.
A = ±1
As for example, the minor of the element
Classification of complex matrices:
a 22 a 23
Hermitian matrix: a 11 is M11 = = (a 22 a 33 - a 23 a 32 )
a 32 a 33
If the conjugated transpose of a matrix is
identical to the original matrix, then the Similarly, the minor of the element
matrix is termed as Hermitian matrix. a 21 a 23
a 12 is M12 = = (a 21 a 33 - a 23 a 31 )
A =A
θ
a 31 a 33
Skew-hermitian matrix: Minor of the element
If the conjugated transpose of a matrix a 12 a 13
is negative of the original matrix, then the a 31 is M31 = = (a 12 a 23 - a 22 a 13 )
a 22 a 23
matrix is termed as skew Hermitian matrix.
Aθ = -A

Linear Algebra 9
 Cofactors: 2. The value of determinant is multiplied
The minor Mij multiplied by (-1) is called i+j by -1 if two rows or two columns are
the cofactor of element aij . We shall denote interchanged.
the cofactor of an element by corresponding 3. The value of determinant can be zero in
capital letter. the following cases.
Example: Cofactor of aij = Aij = (-1)i+j Mij a) The elements in two rows or two
columns are identical
Cofactor element
1+2 1 2 3
a 12 = A 12 = (-1) M12
A = 1 2 3 =0
a a 23 4 5 6
= - 12 = - (a 21 a 33 - a 23 a 31 )
a 13 a 33
b) The elements in two rows and two
Cofactor element columns are proportional to each other.
3+ 1
a 31 = A 31 = (-1) M31 = M31 1 2 3
a 12 a 13 1 5 3 =0  C3 = 3C1
= = (a 12 a 23 - a 22 a 13 )
a 22 a 23 2 5 6

We define for any matrix, the sum of the c) All elements in any row or any column
products of the elements of any row or are zeros.
column with corresponding cofactors is 1 2 3
equal to the determinant of the matrix. A = 4 5 6 =0
Thus, the determinant of a 3x3 matrix is, ∆ = 0 0 0
a11 × A11 + a12 × A12 + a13 × A13
é 1 2 0ù d) If the elements in the determinant are
ê ú of consecutive order (valid for 3rd and
Example : If A = êê-1 6 1 úú higher order)
ê 2 0 2ú
êë úû 1 2 3
é 4 - 12úù
ê 12 A = 4 5 6 =0
ê
Cof (A ) = ê-4 2 4 úú 7 8 9
ê 2 -1 8 úúû
êë
e) The first row of each element starts from
A = (1 ×12) + (2 × 4) + (0 × -12) = (-1 × -4) + the 2nd element of previous row such
(6 × 2) + (1 × 4) = (2 × 2) + (0 × -1) + (2 × 8) = 20 that the elements in that determinant
Note: The determinant can only be are of consecutive order.
computed for square matrices, and it can 1 2 3
be calculated along any row or column. A = 2 3 4 =0
Properties of determinants: 3 4 5
1. The value of the determinant will not 4. The determinant of upper triangle, lower
be changed if rows and columns are triangle, diagonal, scalar or identity matrix
interchanged. is the product of its diagonal elements.
i.e. A = A 10 7 5
5 4 Upper triangle = A = 0 13 1
A = = 10 - 4 = 6
1 2 0 0 7
T 5 1 = 10 ´ 13 ´ 7 = 910
A = = 10 - 4 = 6
4 2

10 Linear Algebra
 11 0 0 y If A is n × n matrix then
Lower triangle = A = 21 7 0 KA = Kn A
54 5 9 4 2
Consider the matrix A =
= 11 ´ 7 ´ 9 = 693 2 4 2´2
3 0 0 4 2
\ 3A = 32 ´ = 9 ´ 12 = 108
Diagonal matrix = 0 6 0 = 3 ´ 6 ´ 1 = 18 2 4
0 0 1
y If each element of a determinant
Scalar matrix: A diagonal matrix with same contains sum of two elements, then that
diagonal elements. determinant should be expressed as sum
2 0 0 of two determinants of the same order.
A = 0 2 0 = 2´ 2´ 2 = 8 a a2 a3 + 3 a a2 a3 a a2 3
0 0 2 Eg : Aa 11= b
a 12b................
2
b3 + 3 a= m b b
2
b3 + b b2 3
a 12 a 22 ................ a
c c2 c3 + 3 m c c2 c3 c c2 3
A = :
: ................ am
Solved Examples
am1 am2 ................ amn
Example: If A = (a)ij 3×3, B = (b)ij 3×3 such that bij Since, aij = i + j Vi, j
é ù
= 2i+jaij Vi, j and A = 2 then B = _______? ê 1 + 1 1 + 2 ............. 1 + nú
ê2 + 1 2 + 2 ............. 2 + n ú
Solution: Given, bij = 2 aij V ij
i+ j
ê ú
A = êê: ú
ú
22 a 11 23 a 12 24 a 13 ê: ú
ê ú
\ B = 23 a 21 24 a 22 25 a 23 = 22 ´ 23 ´ 24 ê ú
êëm+ 1 m+ 2 ............. m+ nûú
24 a 31 25 a 32 26 a 33 m´n

a 11 2a 12 22 a 13
1 1 1 ........... 1 1 2 3 .......... n
a 21 2a 22 22 a 23
2 2 2 .......... 2 1 2 3 .......... n
a 31 2a 32 22 a 33
= : +:
a 11 a 12 a 13 : :
B = 2 ´ 2 ´ 2 ´ 2 ´ 2 a 21 a 22 a 23
2 3 4 2
m m .......... m 1 2 3 .......... n
m´n
a 31 a 32 a 33 Taking the sum of each column in first matrix.
A = 2 Þ B = 2 ´2 = 2 12 13 It will be equal to sum of first ‘m’ natural
numbers
Example: If A = aij m×n such that aij = i + j V i, j
m (m + 1)
then sum of all elements of A is ?
2
Solution: The determinant of A can be
expressed as, Sum of each column = 2
mn (m + 1)
a 11 a 12 ................ am Sum of ‘n’ columns =
2
a 12 a 22 ................ am
Taking sum of each row in second matrix.
A = : Each row sum is equal to sum of ‘n’ natural
: ................ am numbers.
n (n + 1)
am1 am2 ................ amn Sum of each row =
2
Since, aij = i + j Vi, j
é 1 + 1 1 + 2 ............. 1 + nù
ê ú
Linearê2Algebra
+ 1 2 + 2 ............. 2 + n ú 11
ê ú
A = êê: ú
ú
 Sum of ‘m’ rows = mn (n + 1) a 22 a 23 a 24 a 21 a 22 a 24
2 A = a 11 a 32 a 33 a 34 - a 12 a 31 a 33 a 34
mn é a 42 a 43 a 44 a 41 a 43 a 44
Total Sum = m + n + 2ûúù
2 ëê
a 21 a 22 a 24 a 21 a 22 a 23
Example: Given A = (aij) such that aij = i2 - j2 V +a 13 a 31 a 32 a 34 + a 14 a 31 a 32 a 33
i, j . Find sum of the elements of the matrix.
a 41 a 42 a 44 a 41 a 42 a 43
Solution: The elements of matrix are,
é0 - 3 - 8 .......... 1 - n2 ù By using either row or column or both
ê ú operations make all elements above and
ê3 0 - 5 .......... 2 2
- n2ú

A = êê ú below or to the left and right of the selected

2 ú
ê8 5 0 ............... - n ú element are zeros. And then expand the
ê 2 ú
ên - 12
n2 - 22 .......... 0 úû determinant along the row or column.
ë
a 11 a 12 a 13 a 14
⇒ Skew symmetric matrix
0 b22 b23 b24
∴ sum of elements = 0 A=
0 b32 b33 b34
Example: Matrix A has m rows and (m +
0 b42 b43 b44
5) columns and B has n rows and (11 - n)
columns, the orders of A & if AB and BA are b22 b23 b24
defined is? A = a 11 b32 b33 b34
Solution: Since AB is defined m + 5 = n Since b42 b43 b44
BA is defined, 11 - n = m
bij are the new elements after row or column
Solving above two equations, m = 3 and n = 8
transformation.
Example: The following represents equation
Example: Determine the determinant of the
x 2 4 matrix
of a straight line. y 8 0 = 0. The line é ù
ê 1 2 - 1 5ú
1 1 1 ê0 4 5 7 ú
passes through, ê ú
ê1 1 1 3 ú
ê ú
(A) 0, 0 (B) 3, 4 ê2 3 4 9 ú
(C) 4, 3 (D) 4, 4 ëê ûú
Solution: Applying the row transformations,
Solution: The determinant is given by 8x + 2y
R3 → R3 - R1
- 32 = 0 8x + 2y = 32 é 5úù
R4 → R4 - 2R1 ê1 2 -1
Substituting the options we get (3, 4) ê0 4 5 7 úú
The resultant matrix is, A = êê
Solution: 2 - 2 úú
ê0 - 1
A = 5 ´ (3 - 0) + 2 ´ (0 - 2 ´ 3) = 15 - 12 = 3 ê0 - 1 6 - 1 úúû
êë

Note: A 4x4 determinant can be calculated Determinant is given by,

as, é4 5 7 ùú
ê
a 11 a 12 a 13 a 14 A = 1 ´ êê-1 2 - 2úú = 17
ê-1 6 - 1 ú
a a 22 a 23 a 24 ëê úû
A = 21
a 31 a 32 a 33 a 34 Example: Find the value of the determinant
a 41 a 42 a 43 a 44

12 Linear Algebra
 1+b b 1 2. A and B inverse of each other iff AB = BA = I
A= b 1+b 1 3. (AB)-1 = B-1A-1
1 2b 1 4. (ABC)-1 = C-1B-1A-1
5. If A be an n×n non-singular matrix, then
Solution: Applying the column transformation,
(AT)-1 = (A-1)T
C1 → C1 + C2
6. If A be an n×n non-singular matrix, then
1 + 2b b 1
(A-1)θ = (Aθ)-1
1 + 2b 1+b 1
7. 
For a 2 × 2 matrix there is a short-cut
1 + 2b 2b 1 formula for inverse as given below
Since, C1 = (1 + 2b) C3 i.e. both columns are éa bùú 1 é d - bù
proportional ê = ê ú
êc dúûú (ad - bc) êëê-c a úûú
\ A =0 ëê

Adjoint of a matrix: 8. Assuming A is matrix of order “n x n”

Let B be the cofactor element of the matrix A A–1 adj A
A, then adj of A = BT
A A–1 = adj A
Singular matrix:
A matric is said to be singular matrix if |A| =0. Since, KA = kn A
n
It is called as nonsingular if |A| ≠ 0 Thus, A A–1 = adj A
The inverse of only non-singular matrix can n –1
A A = adj A
be computed. Thus, singular matrices are
n–1
non-invertible. A = adj A
A matrix A is used to be invertible if we can
Replacing A by adj A
find some other matrix B such that AB=BA=I,
then B is called Inverse of the matrix A. ⇒ adj (adj A) = adj A
n–1
= A
n–1

adjA
Inverse of a matrix A is given by, A-1 =
A (n – 1)2
i.e., adj (adj A) = A
Properties of inverse of a matrix:
(n – 1)3
1. AA = A A = I
-1 -1 Similarly adj(adj(adj A)) = A

Solved Examples

é3 4 ùú Example: If A = (aij)5×5 such that aij = i - j . Find

ê
ê 5 úú and mT = A-1 in each case.
Example: For the matrix m ê 5
m-1. Find x? ê 3ú Solution: aij = i - j = -(j - i)
ê´ ú
êë 5 úû aij = -aji
Solution: If m is an orthogonal matrix, its rows Thus, A is a skew symmetric matrix
and columns must be pairwise orthogonal
and periodic orthogonal. A T = -A

3 4 3 A T = -A
x+ ´ =0 5
5 5 5 A = (-1) A = - A
3 4 3 A + A =0
x=- ´
5 5 5
4 2A =0
x=-
5 A =0

Linear Algebra 13
 Hence, A is a singular matrix and thus A-1 \ B = 0 Þ adj B = Bn-1 = 0
does not exist n-1 2
(B) adj A = A = A
Note: Every odd order skew symmetric
2k - 1 2 k 2 k
matrix is a singular matrix A = 0 Every
n´n
A = 2 k 1 - 2k
even order skew symmetric matrix is a non-
-2 k 2k -1
singular matrix A ¹0
n´n
The determinant of even order skew
2k - 11
2k - 2
2 kk 2
2 kk
symmetric matrix is always perfect square.
R ®R
R22 ® +R
R22 + R33 Þ
Þ 0
0 + 2k - (1 + 2k ))
11 + 2k - ( 1 + 2k
0 -3 -2
Eg : = 9 = 32 ¹ 0 - 2 k
k 2k
2k - 11
-
3 0
2k - 11
2k - 4
4 k
k 2
2 k
k
Example: If x and y are two non-zero matrices C ®C
C22 ® ®C
C22 ® C33 Þ
Þ 0
0 0
0 - ((11 +
- 2k ))
+ 2k
of the same order, such that xy=O, then
(A) x ¹ 0, y = 0 -2
- 2 k
k 2k - 11
2k - - 11
-

(B) x = 0, y ¹ 0 Expanding the determinant along the second

(C) x ¹ 0, y ¹ 0 row,

(D) x = 0, y = 0 A = (1 + 2k ) éê(2k - 1) + 8ùú

ë û
Solution : Let x = 0 and y ¹ 0 Þ y-1 exists. = (1 + 2k ) ê4k - 4k + 1 + 8k ùú
é 2
ë û
xy = 0 = (1 + 2k ) = (2k + 1)
2
xyy-1 = 0y-1 2 é 3ù
\ adj A = A ê(2k + 1) ú = (2k + 1)
6

x = O is false since it is given x and y are non- ë û

6
zero matrices. (C) (2k + 1) = 106
\ y ¹ 0 is wrong Þ y = 0 9
ie. 2k + 1 = 10 Þ 2k = 9 Þ k =
2
Similarly, x = 0
Example: Given an orthogonal matrix A
Note: Product of two nonzero matrices is a é1 1 1 1 ùú
ê
null matrix if both of them are singular. ê1 1 - 1 - 1 úú -1
= êê ú . The (A ´ A T ) is ______?
Example: Let k be a positive real number and ê1 - 1 0 0ú
ê0 0 1 - 1 úúû
let ëê
é 2k - 1 2 k 2 k ùú
ê( ) Solution: (A × AT) = I
ê ú
A=ê 2 k 1 - 2k ú , ∴ (A × AT)-1 = I-1 = I
ê ú
ê-2 k 2k - 1 úú Example: Find the inverse of matrix.
ê
ë û é ù
é 0 2k - 1 ù êcos - sin 0ú
ê ú A = êêsin cos 0 úú ?
ê ú
B = ê 1 - 2k 0 2 kú ê 0
ê ú ëê 0 1 úûú
ê- k -2 k 0 k úû
ë
Solution:
Find A) |adj B|   B) |adj A|   C) |adj A|
Since its rows and columns are pairwise
= 106 , the value of k is ________?
orthogonal. Thus, A is an orthogonal matrix.
Solution: (A) Here B is an odd order skew AT = A-1
symmetric matrix.

14 Linear Algebra
 é os sin 0ùú Some of the submatrices of A are
ê
\A -1
= êê- sin cos 0 úú é1 2 3 ù é 3 4ù
B1 = ê ú , B2 = ê ú
ê 0 0 1 úúû êë4 5 6úû êë6 7 úû
êë 2´3 2´2

é 5 6 7ù
Rank of a Matrix B3 = ê ú
êë-1 0 2úû
2´3
Submatrix:
Thus, minor can also be defined as the
A matrix is obtained after deleting some rows determinant of a square sub-matrix.
or columns is called a submatrix.
If the determinant of at-least one highest
é1 2 3 4ùú possible square sub matrix is nonzero then
ê
Eg : A = êê4 5 6 7 úú order of the determinant is called rank of a
ê1 - 1 0 2 úûú matrix.
ëê

Solved Examples

Example: Find the rank of the matrix, y Rank (A) = rank (AT)
é1 3 - 2 1 ùú y If rank (A)n × n = n then rank (adj A) = n
ê
ê
A = ê2 - 1 4 5 úú ? y If rank (An × n) = n - 1 then rank (adj A) = 1
ê 1 - 11 Example: Determine the rank of adjoint of
êë 14 5úúû
3´4
1 2 3
Solution: Highest possible square matrix is the matrix A = 0 5 -1
3 × 3.
0 0 0
1 3 -2
Solution: Adjoint of A is given by,
B = 2 -1 4
é 0 - 17úù
1 - 11 14 ê0
ê
adj (A ) = ê0 0 1 úú
= (44 - 14) - 2 (42 - 22) + (12 - 2) = 0 ê0
êë 0 5 úúû
3 -2 1
B = -1 4 5 Applying row transformations, R1 → R1 → 17R2
, R3 → R3 → 5R2 as rank is invariant under row
-11 14 5
transformation.
= 3 (20 - 70) + 1 (-10 - 14) - 11 (-10 - 4) ¹ 0 é0 0 0ù
ê ú
\ Rank (A ) = 3 adj A = êê0 0 1 úú
Properties of rank of a matrix: ê0 0 0 ú
ëê ûú
y Rank (On × n) = 0 ∴ rank of (adj A) = number of non-zero rows
y Rank (In×n) = n =1
y Rank (aij (In×n)) = n If rank (An×n) ≤ n - 2 , then rank of (adj A) = 0
y Rank (A + B) ≤ rank (A) + rank (B)
Echelon form:
y Rank (A - B) ≥ rank (A) - rank (B)
If A is an m × n matrix, then rank (A) ≤ min Row Echelon Form: A matrix is said to be
(m,n) in echelon form if it satisfies the following
conditions.
y Rank (AB) ≤ min {rank(A), rank(B)}
y Rank (AB) ≥ rank (A) + rank (B) - n. If A and y All zero rows should occupy last rows if any.
B are n × n matrices y The number of zeros before a non-zero
entry of each row is less than the number
of such zeros before a non- zero entry of

Linear Algebra 15
 the next row. i.e., zeros increases row by R3 → 2R3 – R1
row. é2 1 1 0 ù
ê ú
y Rank of a matrix in Echelon form = number A = êê0 1 1 5 úú
of non-zero rows. ê0 1 3 6ú
y To reduce any matrix to row echelon form êë úû
we use only row operations. R3 ® R3 - R2
y The number of non-zero rows in row é2 1 1 0ù
ê ú
echelon Form are also called linearly A = êê0 1 1 5úú
independent vectors or LI rows. ê0 0 2 1ú
êë úû
Note: Every upper triangular matrix will be Since, there are three linearly independent
a row Echelon matrix, but converse may or rows, rank (A) = 3
may not be true. Echelon form as it satisfies
the conditions. Note: If rank(A) = n then it has n linearly
independent rows and linearly independent
0 -2 5 6 columns.
0 0 5 2
A= ® Echelon form Example: If A = (aij) such that aij = i × j V i, j
0 0 0 3 then rank(A) = ?
0 0 0 0 Solution:
But not upper triangular matrix as the The matrix A when expanded is,
diagonal elements are all zero. é1 2 3 ...... nùú
ê
Example: Determine the rank of matrix A ê2 4 6 ...... 2n úú
é ù A = êê ú
ê 1 2 2 - 1ú ê: ú
A = êê2 1 2 0 úú êm 2m 3m ...... nmú
êë úû
ê2 2 1 3 ú é 1 2 3 ...... nù
ëê ûú ê ú
ê 1 2 3 ...... nú
Solution: Applying row transformation
= 1 ´ 2 ´ 3 ´ ................ ´ m êê ú
ú
R2 → R2 → 2R1 :ê ú
ê 1 2 3 ...... nú
R3 → R3 → 2R1 êë úû
é 2 - 1úù R2 → R2 - R1
ê1 2
\ A = êê0 - 3 - 2 2 úú R3 → R3 - R1
ê0 - 2 - 3 5 úúû Rm → Rm - R1
êë
Now R3 → 3R3 – 2R2 é 1 2 3 .......... n ù
ê ú
é1 2 2 - 1ùú ê0 0 0 .......... 0ú
ê A = m! êê ú
ú
A = êê0 - 3 - 2 2 úú ê: ú
ê0 0 - 5 ê0 0 0 .......... 0ú
ëê 11úúû êë úû
There are three linearly independent rows Number of non-zero rows = 1
and thus rank(A) = 3 ∴ Rank (A) = 1 as there is only 1 linearly
Example: Determine the rank of independent vector.
é2 1 1 0 ùú Example: If x = (x1 x2 ... xn)T is an n-tuple
ê
A = êê0 1 1 5 úú nonzero vector. Then find rank (xxT )
ê1 1 2 3úûú Solution: Since x is a n-element column
ëê
vector. Then, xT is an n-element row vector.
Solution: Applying row transformation

16 Linear Algebra
Thus, both have rank of 1. Linearly dependent vectors:
A set of r n-vectors (vectors having n-element
{
(XXT ) £ min (X)(XT ) } Here, ρ represents each), x1, x2 … xr are said to be linearly
P (XXT ) £ min {1, 1} = 1. dependent if there exists r scalars K1, K2 … kr
such that K1x1 + K2x2 + … +Krxr = 0 where K1,
rank of matrix. K2, ….. Kr not all zeros. (At least 1 K should be
Example: The rank of 5 × 6 matrix is 4. Then a non-zero number)
which of the following statement is true? Linearly independent vectors:
(A) Q has 4 linearly independent rows and 4 ‘r’ vectors x1, x2 … xr are linearly independent
linearly independent columns. vectors if there exist ‘r’ scalars. K1, K2 … Kr
(B) Q has 4 LI rows and 5 LI columns such that K1X1 + K2X2 + … + KrXr = 0 where K1,
(C) Q . QT is invertible K2 … Kr are all zeros.
(D) QT . Q is invertible y If the rank of matrix is less than the
Solution: Option (A) is correct dimension of the matrix, then the vectors
are said to be linearly dependent. Thus, if
If rank of a matrix is 4 then it have 4 LI rows
the determinant of the matrix is zero the
and 4 LI columns. These 4 LI rows can be
vectors are said to be linearly dependent.
obtained by reducing the matrix into echelon
form and 4 LI columns can be obtained by y If the rank of matrix is equal to the
reducing the matrix into echelon form. dimension of the matrix, then the vectors
are said to be linearly independent. Thus, if
Linearly dependent and independent vectors:
determinant of the matrix is non-zero the
If the elements are written in a horizontal vectors are said to be linearly independent.
line or in vertical line is called as vector. This y If the number of components of the
means row matrix, or a column matrix can be vectors is more than the number of
termed as a vector. vectors, then the vectors are said to be
Example: (1 2 3) → vector linearly independent but if the number of
Two vectors x1 and x2 are said to be linearly components is less than the number of
dependent if it is possible to express one of vectors then the vectors are said to be
the vectors as multiple of another vector. linearly dependent.
é1 ù Example: Consider vectors (1 2 3 4)
Suppose, one vector is x 1 = êê úú (2 1 -1 7)
ëê2ûú
é2 ù 1 (2 3 4 5)
X2 = êê úú = 2x 1 or x 1 = x
êë4úû 22 Number of vectors r = 3 Number of component
n=4
Two vectors in R2 (two dimensional) are said
to be linearly dependent if they are collinear. Since r < n, the vectors are linearly
independent.
Similarly 3 vectors in R3 (3-D) are said to be
linearly dependent if they are coplanar i.e. If the vectors are expressed in Matrix form,
lying in the same plane. é1 2 3 4 ùú
ê
If it is not possible to express 1 vector as A = êê2 1 - 1 7úú
multiple of other vector the two vectors are ê2 3 4 5 úûú
ëê 3´4
said to be linearly independent.
é1 ù é2ù Applying row transformations,
Suppose, x1 = ê ú and x2 = ê ú R2 → R2 - 2R1
ê2ú ê3ú
ëê ûú ëê ûú
Hence, x1 ≠ kx2 . Thus, x1 & x2 are linearly R3 → R3 - 2R1
independent Vectors.

Linear Algebra 17
 é1 2 3 4ù The set of vectors having at least one vector
ê ú is a null vector, the vectors are said to be
A êê0 - 3 - 7 - 1úú
ê linearly dependent.
êë0 - 1 - 2 - 3úûú
Example: Consider the vectors X1 = (2 3 4);
R33 ® 3R33 - R22 X2 = (1 -1 2); X3 = (0 0 0). Determine whether
é1 2 3 4 ùú these vectors are linearly dependent or
ê
A êê0 - 3 - 7 - 1úú independent.
ê ú Solution: Expressing the vectors in matrix
êë0 0 1 8 úû
(A ) = 3 form. The determinant is given by,
2 3 4
Thus, the rank of matrix is same as number
A = 1 -1 2 =0
of vectors. Thus, the vectors are linearly
Independent. But if we consider the vectors 0 0 0
X1 = (1 2 3); X2 = (1 0 3); X3 = (1 1 - 1); X4 = (0 1 Since |A| =0, rank(A) < 3, but the determinant
2) Number of vectors r = 4 of 2×2 square sub-matrix is nonzero
Number of components n = 3 2 3
Since r > n, so the given vectors are linearly B = = -5 ¹ 0 Thus,(A ) = 2
1 -1
dependent.
Number of given vectors = 3. Thus, ρ(A) < n
Suppose the vectors are linearly dependent,
and the vectors are linearly dependent.
then any 1 of the vectors can be expressed as
a linear combination of other vectors. Orthogonal vectors:
If the given vectors are linearly independent, For orthogonal Vectors dot product or inner
then it is impossible to express any of the product of the vectors must be zero. Each
vectors as a linear combination of other term in the row or column vector can be
vectors. considered as a component of the vector.
The dot product is obtained as sum of the
Example: Determine whether the given
product of corresponding components of
vectors are linearly independent or Dependent
both vectors.
X1 = (1 0 0); X2 = (1 0 0); X3 = (0 0 1)
In matrix form it can be expressed as, X1TX2 =
Solution: Expressing the linear combination
X1 X2T = 0
of vectors and equating them to zero, K1X1 +
K2X2 + K3X3 = 0 The set of orthogonal vectors in Rn (R vectors
with n elements are linearly independent).
Substituting the vectors, K1(1 0 0) + K2(0 1 0)
+ K3(0 0 1) = (0 0 0) Dimension and basis:
i.e. (K1 K2 K3) = (0 0 0) K1 = K2 = K3 = 0 y The number of linearly independent
Since, all coefficients are zero. Thus, all vectors in a vector space is called as
vectors are linearly independent. dimension of vector space.
y For a three-dimensional space, there are
Alternate method:
only 3 linearly independent vectors and
Expressing the vectors in matrix form, any fourth vectors can be expressed as
é1 0 0ùú a linear combination of the other three
ê
A = êê0 1 0úú = 1 ¹ 0 vectors.
ê 1ûúú y Number of linearly independent vectors
ëê0 0 can be determined by expressing the
(A ) = number of non-zero rows = 3 matrix in Row Echelon Form and the
number of non-zero rows is equal to the
Since, rank is same as number of vectors.
number of linearly independent vectors.
The vectors are linearly independent.

18 Linear Algebra
y The set of linearly independent Vectors is same as the order since the determinant of
called as basis of vector space. Any vector the matrix is non-zero.
in vector space can be expressed as a Connection between rank and span:
linear combination of its basis vectors.
A set of n vectors X1 , X2 , X3 ... Xn spans Rn iff
Nullity of a matrix: they are linearly independent which can be
Nullity of a matrix is denoted by N(A) and is checked by constructing a matrix with X1 , X2 ,
defined as the difference between order and X3 ... Xn as its rows (or columns) and checking
rank of the matrix. that the rank of such a matrix is indeed n. If
i.e. N(A)=n(A) -ρ(A) however the rank is less than, n say m, then
Nullity of a non-singular matrix of any order the vectors span only a subspace of Rn
is always zero as the rank of the matrix is

Solved Examples

Example: Test whether the following vectors In row echelon form, the number of non-zero
are linearly dependent or linearly independent. rows is 3. Thus, rank of the matrix is 3. Thus,
Also find dimensions and basis? the dimension is 3.
(1 1 -1 0) (4 4 -3 1) (-6 2 2 2) (9 9 -6 3) Basis is {(1 -1 0 1), (0 8 -4 2), (0 0 1 1)}
Solution: Expressing the vectors in matrix Example: Determine the value of k if nullity
form, ék
ê 1 2ùú
é 0úù
ê 1 1 -1 of A = êê 1 - 1 - 2úú is 1?
ê 4 4 -3 1 úú ê2
A = êê ëê 1 1 úúû
ê-6 2 2 2 úú
ê 9 Solution: The order of matrix A is n(A) = 3
êë 9 -6 3 úúû
nullity of matrix A is, N(A)=n(A)-ρ(A) 1=3 - ρ(A)
Applying row transformation, R2 → R2 – 4R1 ; ∴ ρ(A) = 3 - 1 = 2
R3 → R3 + 6R1 ; R4 → R4 - 9R1
Since, Rank is 2. Thus, A = 0
1 1 -1 0
K 1 2
0 0 1 1
A= A = 0 Þ 1 -1 -2 = 0
0 8 -4 2
2 1 1
0 0 3 3
K (-1 + 2) - 1 (1 + 4) + 2 (1 + 2) = 0
Interchanging row 2 and row 3.
K = -1
é 1 - 1 0úù
ê1 Example: The rank of a matrix is 5 and nullity
ê0 8 -4 2 úú
R2 Þ R3 Þ A = êê of the matrix is 3. Then what is the order of
ê0 0 1 1 úú the matrix.
ê0 0 3 3úûú
ëê Solution: N(A) = n(A) - ρ(A)
Applying, R4 → R4 → 3R1 3 = n(A) - 5
é 1úù n(A) = 8
ê1 - 1 0
ê0 8 - 4 2 úú Example: Check if the vectors [1 2 -1], [2 3 0],
A = êê [-1 2 5] span R3 é ù
ê0 0 1 1 úú ê 1 2 - 1ú
ê0 0
êë 0 0úúû Solution: Constructing A = êê 2 3 0úú
ê-1 2 5 úûú
êë
Finding determinant to determine the rank,

Linear Algebra 19
 1 2 -1 Thus, the following system has a unique
2 3 0 = 1 (15 - 0) - 2 (10 - 0) - 1 (4 + 3) solution.
-1 2 5 System of Equations with ‘n’ variables
= 15 - 20 - 7 = -12 ¹ 0
Consider a system of m equations and n
So, rank = 3
variables.
∴ The vectors are linearly independent and
a11x1 + a12x2 + a13x3 + ... + a1nxn = b1
hence span R3
a21x1 + a32x2 + a21x3 + ... + a2nxn = b2
System of Non-Homogenous and :
Homogeneous Linear Equations :
Consider a non-homogenous system of am1x1 + am2x2 + am3x3 + ...+ amnxn = bn
equations, containing two equations and two The above system of equation can be put in
variables. matrix form
ax + by = e éa ùé ù é ù
ê 11 a 12 .......... a 1n ú êX 1 ú êb1 ú
cx + dy = f êa úê ú ê ú
ê 21 a 22 .......... a 2n ú êX2 ú = êb2 ú
The above system of equations has ê: ú ê: ú ê: ú
ê úê ú ê ú
a b e êa úê ú ê ú
y No solution if = ¹ . Lines parallel to êë 21 am2 .......... amn úû ëêXn ûú ëêbn ûú
each other. c d f
AX = B
a b
y Unique solution if = . Lines intersect Where X → solution matrix.
c d
at 1 particular point only. If we write the elements of matrix B in the
a b e last column of matrix A, the resulting matrix
y Infinite number of solutions if = =
c d f is called augmented matrix and is denoted
The given system of equations can be by (A | B)
brought to the form of number of equations éa a 12 .......... a 1 b1 ùú
ê 11
ê an b2 úú
< number of variables. In this case, both éA Bù = êa 22 a ..........
the lines are identical and hence overlap ëê ûú ê: ú
ê ú
or intersect at infinite number of points. êam am .......... a b ú
êë 1 2 mn m úû (m+ 1)´n
Example: The system of equations 4x + 2y =
7, 2x + y = 6 have? Procedure to determine the solution:
Solution: For the given set of equations, 1. Reduce (A | B) to row echelon form
a b e 4 2 7 2. Find (A| B) and ρ(A)
= ¹ as = ¹
c d f 2 1 6 3. I f ρ(A) < ρ(A | B) or ρ(A | B) ≠ ρ(A) the given
∴ No solution exists for the following system system of equations has no solution,
of equations. system is inconsistent.
Example: The system of equations x + 3y = 5, 4. I f ρ(A | B) = ρ(A) = number of unknowns,
2x + 5y= -3 have. the given system of equations has unique
Solution: For the following system of solution i.e. system is consistent.
equations, 5. I f ρ(A | B) = ρ(A) < number of unknowns,
the given system of equations has infinite
1 3 a b
¹ as ¹ number of solutions.
2 5 c d

20 Linear Algebra
6. 
If the number of equations < number In matrix form
of variables (r < n), the given system of éa 11 a 12 ... a 1n ù éX1 ù é0ù
ê úê ú ê ú
equations will have infinite number of non- êa 21 a 12 ... a 2n ú êX2 ú ê0ú
ê úê ú=ê ú
zero solutions. This non-zero solution can ê: ú ê: ú ê: ú
ê úê ú ê ú
be found by assigning (n – r) variables or êa úê ú ê ú
ë m1 am2 ... amn û ëXn û ë0û
arbitrary constants. These n – r solutions are
The following system can be represented as,
said to be linearly independent solutions.
AX=O.
r=rank of matrix & n=order of matrix
The solution of this system can have following
Inconsistent: The given system of equations is cases:
said to be inconsistent if they have no solution.
Case-1: Inconsistency: This is not possible
Consistent: The given system of equations is in a homogeneous system since such a
said to be consistent if they have a solution system is always consistent since the trivial
(unique or infinite). solution C [0, 0, 0.......]t always exists for a
Note: If the number of equations < number homogeneous system.
of variables (or) number of variables exceeds Case-2: Consistent unique solution: If r = n ;
the number of equations, the system of the equation AX = O will have only the trivial
equations have ∞ number of solutions. unique solution X = [0, 0, 0 ...]
Homogeneous system of linear equation: Note: That r = n ⇒ |A| ≠ 0 i.e., A is non-singular.
Consider the following homogenous system of Case-3: Consistent infinite solution: If r
linear equations of m equations in n variables. <n we shall have n - r linearly independent
a11x1 + a12x2 + ... + a1nxn = 0 non-trivial infinite solutions. Any linear
a21x1 + a22x2 + ... + a2nxn = 0 combination of these (n - r) solutions will
also be a solution of AX = 0
:
Thus in this case, the equation AX = O will
:
have infinite solutions.
am1x1 + am2x2 + ... + amnxn = 0
Note: If r < n ⇒ |A| 0 i.e, A is a singular matrix.

Solved Examples

Example: How many solutions does the So rank (A|B) = 2 = rank (A) = number of
following system of equations have? variables. Thus, this system has unique
-x + 5y = -1, x - y = 2, x + 3y = 3 solution.
Solution: The matrix A and B are given as, Example: Find the value of λ and µ for the
é-1 5 ù é-1 ù system of equation.
ê ú ê ú
A = êê 1 - 1úú ; B = êê 2úú x + y + z = 6, x + 4y + 6z = 20, x + 4y λ z = µ
ê 1 3ú ê 3ú to have
ëê ûú ëê ûú
(A) No solution
The augmented matrix is given by,
(B) Unique solution
é-1 5 - 1ù
ê ú (C) ∞ number of non-zero solution
A B = êê 1 - 1 2úú Solution: The augmented matrix is given by,
ê 1 3 3úúû
êë é1 1 1 6 ùú
ê
Determinant of augmented matrix is, éA Bù = ê 1 4 6 20úú
êë úû ê
A B = -1 (-3 - 6) - 5 (3 - 2) - 1 (3 + 1) = 0 ê1 4  úúû
êë

Linear Algebra 21
 Applying row transformation, é1 1
ê 1 5 ùú
R2 → R2 → R1 and R3 → R3 → R1 éA Bù = ê0 2
êë úû ê 2 4 úú
é1 1 1 6 ùú ê0 0 - 2 - 10ú
ê êë úû
éA Bù = ê0 3 5 14 úú
ëê ûú ê α = 2, β = 10
ê0 3 - 1 - 6ú
ëê ûú This ensures that ρ(A | B) = ρ(A) = 2
R3 → R3 → R2 Example: Find the condition on a, b, c for
é1 1 1 6 ùú which the following system of equations
ê
éA Bù = ê0 3 5 14 úú 3x + 4y + 5z = a
ëê ûú ê
ê0 0 - 6 - 20ú 4x + 5y + 6z = b
ëê ûú
(A) For No solution 5x + 6y + 7z = c
ρ(A) < ρ(A | B) do not have a solution?
i.e. λ - 6 = 0 so that last row of A is reduced Solution:
to zero. Then, ρ(A) = 2 The augmented matrix is given by,
µ - 20 ≠ 0 é3 4 5 a ùú
ê
Thus, last row of [A|B] is non-zero and ρ(A | éA Bù = ê4 5 6 búú
êë úû ê
B) = 3 ê5 6 7 c úúû
êë
λ = 6 and µ ≠ 20
Applying row transformations, R2 → 3R2 - 4R1
(B) Unique solution and R3 → 3R3 - 5R1
ρ(A | B) = ρ(A) = number of unknowns. é3
ê 4 5 a ùú
ρ(A | B) = ρ(A) = 3 éA Bù = ê0 - 1 - 2 3b - 4a ú
êë úû ê ú
λ - 6 ≠ 0, µ - 20 ≠ 0 or µ - 20=0 ê0 - 2 - 4 3c - 5a ú
ëê ûú
∴ λ ≠ 6, µ = any value
R3 ® R3 - 2R2
Thus, last row will be non-zero for both A
é3 4 5 a ù
and A | B. ê ú
éA Bù = ê0 - 1 - 2 3b - 4a úú
(C) Infinite number of solutions ëê ûú êê ú
ρ(A | B) = ρ(A) < number of unknowns. êë0 0 0 3 (a + c - 2b)ú
û
λ - 6 = 0 and µ - 20=0 From above matrix (A ) = 2
∴ λ=6 and µ = 20 If 3(a + c - 2b) ≠ 0, then ρ(A | B) = 3 ≠ ρ(A).
Example: For what value of α and β, the Thus, no solution.
following system of equations Example: For what values of λ does the system
x + y + z = 5, x + 3y + 3z = 9, 2x + 2y + αz = β of equations have two linearly independent
have infinite number of solutions. solutions.
Solution: For infinite number of solutions ρ(A x + y + z = 0, (λ + 1)y + (λ + 1)z = 0, (λ2 - 1)z = 0
| B) = ρ(A) < number of unknowns Solution: Since there are two linearly
é1 1 1 5ùú independent solutions n - r = 2
ê
éA Bù = ê 1 3 3 9 ú 3 -r= 2
êë úû ê ú
ê2 2  ú
êë úû r=3-2=1
Apply Row Transformations, R2 → R2 -R1; R3 If rank is 1, determinant of the matrix should
→ R3 - 2R1 be zero.

22 Linear Algebra
 1 1 1 Example: Find the value of k for which the
A = 0 + 1 + 1 =0 system of equations.
0 0  - 1 3´3
2 (3k - 8) × 3y + 3z = 0
3x + (3k - 8) y + 3z = 0
1 éê(+ 1) (2 - 1) - 0ùú = 0 3x + 3y + (3k - 8) z = 0
ë û
(+ 1)(+ 1)(- 1) = 0 has nontrivial solutions?
 = +1, -1 Solution: For non-trivial solutions, rank (A) <
For λ = -1, rank = 1 as the determinant of 3. Thus |A| = 0
order 2 square sub-matrix goes to 0. If λ = 1, 3k - 8 3 3
rank = 2 3 3k - 8 3 =0
∴ λ = -1 is correct. 3 3 3k - 8
Example: The rank of A3×3 is 1. The system
R1 → R1 + R2 + R3, R2 → R2 - R3
of equations AX=0 has how many linearly
independent Solutions? 1 1 1
Solution: Order of Matrix n=3 (3k - 8 + 6) 0 3k - 8 - 3 3 - (3k - 8) = 0
Rank of Matrix r =1 3 3 3k - 8
Number of linearly independent Solutions = i.e. (3k - 8 + 6)(3k - 8 - 3)(3k - 8 - 3) = 0
n-r = 2 2 11 11
k= ,k = ,
Example: The system of equations x + 3y - 3 3 3
2z = 0, 2x - y + 4z = 0, x - 11y + 14z = 0
trivial solutions? Eigen Values and eigen Vectors
Solution: The coefficient matrix A is given by, Let A be an n × n matrix and λ is a scalar.
é1 3 - 2ùú The matrix (A - λI) is called a characteristic
ê
A êê2 - 1 4 úú equation matrix.
ê 1 - 11 14ú |A - λI| is called characteristic determinant
ëê ûú or characteristic polynomial. The equation |A
Determinant of A is - λI| = 0 , is called as characteristic equation.
A - 1 (-14 + 44) - 3 (28 - 4) - 2 (-22 + 1) = 0 The roots of the characteristic equation are
called characteristic roots of eigen value or
Thus, rank(A) < 3 = number of unknowns
proper values or latent roots.
Hence, this system has infinite non-trivial
The set of eigen values of a matrix ‘A’ is called
solutions.
as spectrum of A.

Solved Examples

Example: Determine the eigen values of the 4 - 2

é4 2ù A - =
matrix A = êê ú 2 4 -
ú
êë2 4úû Characteristic equation is A - = 0
Solution: The characteristic matrix is
4 - 2 2
é4 2ù  é 0ùú éê4 - 2 ùú = (4 -) - 4 = 0
A -I = êê ú-ê 2 4 -
ú ê ú=ê ú
êë2 4úû êë0 úû êë 2 4 -úû
16 + λ2 - 8λ - 4 = 0
Characteristic polynomial is λ2 - 8λ + 12 = 0
λ = 6,2

Linear Algebra 23
 Eigen Vectors: If λ is an eigen value of the Equation (2) is obtained from Equation (1)
matrix A, then there exists a non-zero vector by multiplying a factor of ‘-2’. Thus, unique
X such that AX = λX. The X is called an eigen solution is impossible.
vector corresponding to the eigenvalue λ. x1 2
=
From Previous Example, consider the matrix x2 1
é4 2ù é 1ù é4 + 2ù é6ù é ù
A = êê úê ú=ê ú = ê ú = 6 ê1 ú Therefore the eigen vector corresponding to
úê ú ê ú ê ú ê ú
ëê2 4ûú ëê 1ûú ëê ûú ëê ûú ëê ûú eigen value λ = 5
é x ù é2ù
é 1ù X 1 = êê 1 úú = êê úú
\ êê úú is an eigen vector corresponding to ëêx2 ûú êë 1 úû
ëê 1ûú
Case-2: When λ = -5
the eigen value λ = 6.
AX - λIX = 0
é4 2ù é 1 ù é4 + 4ù é8 ù é4ù
Similary, êê úê ú=ê ú ê ú ê ú
ú ê2ú ê8 + 2 ú = ê 10ú = ê5 ú
é3 - 4 ùú éêx 1 ùú éê0ùú
2 4 ê
ëê úû ëê ûú ëê úû ëê ûú ëê úû ê 4 - 3 -ú êx ú = ê0ú
ëê ûú ëê 1 ûú ëê úû
é1 ù For  = -5
\ êê úú is not an Eigen vector.
êë2úû é8 4ù éx ù é0ù
ê úê 1ú ê ú
ê 4 2ú ê x ú = ê ú
ëê ûú ëê 2 ûú ëê ûú
Example: Find the eigen values and eigen 8x1 + 4x2 = 0
é3 4 ù 4x1 + 2x2 = 0
vectors of matrix A = êê ú
ú
êë4 - 3úû The first equation in this case can be obtained
Solution: The characteristic equation is, from second by multiplying a factor of ‘2’
3 - 4 i.e. 4x1 = -2x2
A - = 0 Þ =0
4 - 3 - -x 2
x1 =
- (3 -)(3 +) - 16 = 0 2
x1 -1
2 - 9 - 16 = 0 \ =
x2 2
2 - 25 = 0 éx ù é-1ù
\ X2 = êê 1 úú = êê úú
2 = 25 êëx2 úû êë2 úû
 = ±5 é 3 4ù
A T = êê ú = A. Thus, the matrix A is symmetric.
Case-1: When λ = 5 ëê4
ú
ûú
AX = λX For the two eigenvectors obtained,
AX = λXI
x 1xT1 = 0 & T1 x2 = 0. Thus, both eigen vectors
AX - λXI = 0
are orthogonal.
(A - λI) X = 0
Remark:
é3 - 4 ù éx ù é0ù
ê úê 1ú ê ú The eigen vectors corresponding to distinct
ê 4 - 3ú êx ú = ê0ú
ëê ûú ëê 2 ûú ëê ûú eigen values of a real symmetric matrix are
For  = 5 always orthogonal.
é-2 4ù éx ù é0ù 2 -1
ê úê 1ú = ê ú x1 x2 = = 4 - (-1) = 5 ¹ 0 Þ L.I.
ê4 úê ú ê ú
êë ûú ëêx2 ûú ëê ûú 1 2
-2x1 + 4x2 = 0 → (1) The eigen vectors corresponding to distinct
4x1 - 8x2 = 0 → (2) eigen values of any square matrix are always

24 Linear Algebra
linearly independent. The eigen vectors é1 ù
ê ú
corresponding to repeated eigen values may
x 1 = êê0úú
be linearly independent or linearly dependent ê0ú
vector. êë úû
Example: Find the eigen values and corres­ Example: Find the eigen values and corres­
ponding eigen vectors of the following matrix: ponding eigen vectors of the following matrix:
é1 1 0ù é 1 1 0ù
ê ú ê ú
A = êê0 1 1úú A = êê0 1 0úú
ê0 0 1ú ê0 0 1ú
ëê ûú êë úû
Solution: The characteristic matrix is, Solution: For any upper triangular matrix,
é 1 - 1 0 ùú diagonal elements represent the eigen values
ê as seen in previous example.
A -I = êê 0 1 - 1 úú
ê 0 Thus, λ = 1,1,1
ëê 0 1 -úûú
To determine the eigen-vectors, (A - λI) x = 0
Characteristic equation is,
é 1 - 1 0 ùú
1 - 1 0 ê
ê 0 1 - 0 ú éxù = 0
3 ê ú ëê ûú
A - = 0 1 - 1 = (1 -) = 0 ê 0 ú
ëê 0 1 - ûú
0 0 1 -
é0 1 0ù éx ù
λ = 1,1,1 are the eigenvalues of A. ê úê 1ú
Since, λ = 1; êê0 0 0úú êêx2 úú = 0
To determine the eigen-vectors, (A - λI) X = 0 ê0 0 0ú êx ú
é 1 - 1 0 ùú éêx 1 ùú éê0ùú ëê ûú ëê 3 ûú
ê In this case, x2 = 0 but x1 & x3 can have any
ê 0 1 - 1 ú êx ú = ê0ú
ê ú ê 2ú ê ú values. r < n, put n – r = 3 - 1 = 2
ê 0 0 1 - ú êx ú ê0ú
ëê ûú ëê 3 ûú ëê ûú Two variables can be assumed as arbitrary
Since,  = 1 values.
é0 1 0 ù é x ù é0ù Put x1 = c1, x2 = 0, x3 = c2
ê úê 1ú ê ú
ê0 0 1 ú ê x ú = ê0ú éc ù é1 ù é0ù
ê ú ê 2ú ê ú ê 1ú ê ú ê ú
ê0 0 0ú ê x ú ê0ú
êë úû êë 3 úû êë úû X = êê0 úú = c êê0úú + c2 êê0úú
êc ú ê0ú ê1 ú
x2 = x3 = 0, x1 can be any value. ëê 2 ûú ëê ûú ëê ûú
Number of equations < number of variables The two linearly independent eigen-vectors
Infinite number of nonzero solutions, r = 2, n are,
=3. If r < n, put n – r = 3 - 2 = 1 é 1 ù é0ù
This means one variable can be assigned any ê ú ê ú
ê0ú & ê0ú
arbitrary value. ê ú ê ú
ê0ú ê 1 ú
ëê ûú ëê ûú
Note: Zero vectors cannot be eigen vectors.
Both these vectors cannot be expressed as
éc ù é 1 ù
ê ú ê ú multiple of one another. Thus, both are linearly
∴ Eigen vectors = êê0úú = êê0úú c can be 1, 2 independent. Therefore, corresponding to 3
ê ú ê ú
ëê0ûú ëê0ûú equal eigen values, there exist two linearly
independent eigen vectors.
Corresponding to 3 repeated eigen values, Example: Determine the eigen values and
there exists only 1. Linearly independent eigen vectors of the matrix
eigen vector.

Linear Algebra 25
 é 1 0 0ù P = Number of linearly independent eigen
ê ú
ê0 1 0ú vectors. n = order of matrix or number of
ê ú variables.
ê0 0 1ú
êë úû r = ρ (A - λI) i.e. Rank of characteristic Matrix.
Solution: The characteristic equation is, m is the number of times an eigen value is
é 1 - 0 0 ùú repeated.
ê
A -I = êê 0 1 - 0 úú = (1 -)
3 If some eigen values are repeated and some
ê 0 are non-repeated, then the corresponding
ëê 0 1 -úûú
eigen vectors may be linearly independent or
λ = 1,1,1 linearly dependent.
Thus, eigen values of a diagonal matrix are Example: The number of linearly independent
same as its diagonal elements. eigen vector of
To determine eigen vectors, (A - λI) X = 0 é2 1 ù
A = êê ú
ú
λ=1 ëê0 2ûú
é0 0 0ùú éêx 1 ùú Solution: Since the matrix is upper triangular,
ê
i. e . êê0 0 0úú êêx2 úú = 0 diagonal elements represent eigenvalues λ =
ê0 0 0úûú êëêx3 úûú 2,2
ëê
To determine eigen vectors, (A - λI) X = 0
r = 0, n = 3
é2 - 1ù
r < n, put n - r = 3 - 0 = 3 ê ú = A -I
ê 0 ú
Three variables can assume arbitrary values. êë ûú
x1 = c1, x2 = c2, x3 = c3 Since, λ = 2
é0 1 ùú
éc ù é1 ù é0ù é0ù Characteristic Matrix becomes, êê
ê 1ú ê ú ê ú ê ú êë0 0úúû
x = êêc2 úú = c1 êê0úú + c2 êê 1 úú + c3 êê0úú P = n – r, where 1 ≤ P ≤ 2
êc ú ê0ú ê0ú ê1 ú Since, there is only one non-zero row, r =1
ëê 3 ûú ëê ûú ëê ûú ëê ûú
¯ ¯ ¯ P = 2 - 1 = 1 linearly independent eigen vector
X1 X2 X3 Example: The number of LI eigen vector of
é3 0ù
1 0 0 A = êê ú
ú
ëê0 3ûú
0 1 0 = 1¹0
0 0 1 Solution: Since, this is a diagonal matrix, the
diagonal elements represent eigen values.
The vectors are linearly independent. λ = 3,3
There exists 3 linearly independent eigen To determine eigen vectors, (A - λI) X = 0
vectors.
é3 - 0 ù
é1 ù é0ù é0ù ê ú
ê ú ê ú ê ú ê 0 3 -ú
êë úû
x 1 = êê0úú , x2 êê 1 úú , x3 = êê0úú
ê0ú ê0ú ê1 ú Since λ = 3
êë úû êë úû êë úû é0 0ùú
Characteristic Matrix becomes, ê
Corresponding to 3 equal eigen values. ê0 0úûú
P = n-r = 2 - 0 = 2 ëê
Note: If an eigen value λ is repeated m times, Thus, there are 2 linearly independent eigen
then the corresponding number of linearly vectors.
independent eigen Vectors is given by
Properties of eigen values and eigen vectors:
P = n-r where 1 ≤ P ≤ m
Let A be on n × n matrix. The eigen values of
A are λ1, λ2 … λn. Trace (A) = λ1 + λ2 + λ3 + ... + λn

26 Linear Algebra
Product of the eigen values = Determinant of y If λ is an eigen value of a matrix A then
a matrix. Kλ is an eigen value of kA.
y The eigen values of a upper triangle or
⚪ 1 is an Eigen value of A-1
lower triangle or diagonal or scalar or 
identity matrix is its diagonal elements.
⚪ λ2 is an eigen value of A2
y The eigen values of A and AT are same.
⚪ λm is an eigen value of Am
y The eigen values of A and P-1AP (P is a non-
singular matrix) are same. A
⚪ is an Eigen value of Adj A
y The eigen values of orthogonal matrix are 
unit modular i.e. they have a magnitude ⚪ λ ± K is an eigen value of A ± KI
of 1.
⚪ (λ ± K)2 is an eigen value of (A ±KI)2
y If λ is an eigen value |λ| = 1 of an orthogonal
1 -1
matrix, then 1/λ is also one of its eigen ⚪ is an Eigen value of (A ± KI)
value. ± K
y The eigen values of real symmetric matric
Note:
are real.
y A matrix A is said to be singular. i.e. Det
y The eigen value of skew symmetric matric
A = 0; if one of its eigen value is zero. Its
are purely imaginary or zeros.
converse is also true.
y For a real matrix if a+ib is an eigen value
y If one of the eigen values of a matrix A
then a-ib is also another eigen value of
is zero, then the homogenous system of
the same matrix.
equations has infinite number of non-
y The eigen vector of A and A-1 are same.
zero solutions.
y The eigen vectors of A and Am are same.

Solved Examples

Example: The eigen value of The matrix is

é-1 3 5ùú éé66 2
2ùù éé44 6ùù
6
ê (A) êê
(A) úú (B) êê
(B) úú
A = ê-3 - 1 6úú are
ê êêë2 6úúû
2 6 êêë6
6 4úúû
4
ë û ë û
ê 0 0 3 úúû éé2 2 4
4ùù éé8 8 8ùù
êë 8
(C) êê
(C) úú (D) êê
(D) úú
êëêë4 2úúû
4 2 êëêë8
8 4úúû
4
(A) 3, 3+5j, -6 -j û û
(B) -6 + 5j, 3 - j, 3+j Solution: Since Trace of matrix is same as
(C) 3 - j, 5 + j, 3 + j sum of eigen values Trace = 8+4 = 12
(D) 3, - 1 + 3j, - 1 - 3j Options (A) and (D) have same trace.
Solution: Tr [A] = λ1 + λ2 + λ3 = (-1) + (-1) + 3 = 1 Product of eigen values is same as
Verifying the options, determinant Product = 8 x 4 =32
Only option (d) satisfies the above equation Only option (A) has the determinant 32.
3, - 1 + 3j, - 1 - 3j é1 ù
ê ú
Example: The eigen values and eigen vectors Example: The vector êê2 úú is an Eigen vector of
of a 2 × 2 matrix are given by eigen value. ê-1ú
é-2 ù ëê ûú
λ1 = 8 and λ2 = 4 ê 2ú
é 1ù é1 ù A = êê 2 1 úú . The eigen value corresponding
Eigen vectors 1 = êê úú and2 = êê úú ê-1 - 2 ú
êë 1úû êë-1úû êë úû
to the eigen vector is ?

Linear Algebra 27
 Solution: If X is the eigen-vector of matrix A A = 1´ 2 ´ 3 = 6
then (A - λI)x = 0
Eigen Values of A-1 = 1 = 1, 1 , 1
é-2 - 2 - 3 ùú éê 1 ùú éê0ùú  2 3
ê
ê
ê 2 1 - - 6 úú êê2 úú = êê0úú Eigen Values of A = λ = 1, 4, 9
2 2

ê -1 - 2 0 -úûú êëê-1úûú êëê0úûú A

ëê Eigen Values of Adj(A)= = = 6, 3, 2
-1 - 2 ´ 2 + -´-1 = 0 
-1 - 4 + = 0 (A) EigenValues of A2 + A-1 + adj A
= 5 A
1
= 2 + +
é4 2ù  
Example: For the matrix êê ú , the eigen
ú Eigen Values = 8 + 7.5 +11.33 = 26.83
êë2 4úû
é 101ù Thus, Tr(A2 + A-1 + adj A) = 26.83
value corresponding to the eigen vector êê úú
êë 101úû 15 34
(B) Det (A 2 + A-1 + adj A) = 8 ´ ´ = 680
is _________________ ? 2 3
Example: The eigen vectors of a 2 × 2 matric
Solution: For any eigenvector X,
é1 2 ù é ùé ù
éA -Iù X = 0 ê ú are given by ê 1 ú ê 1 ú What is a + b?
ëê ûú ê0 2ú êa ú êbú
êë úû êë úû êë úû
é4 - 2 ù é 101ù
ê úê ú=0 Solution: Since the matrix is upper triangular,
ê 2 4 úê ú
ëê -ûú ëê 101ûú the diagonal values represent eigen values.
2 × 101 + (4 - λ) 101 = 0 λ = 1, 2
101 (2 + 4 - λ) = 0 Since, (A - λI)X = 0
é ùé ù
6-λ=0 ê 1 - 2 ú êx 1 ú = 0
é ê 0 2 -ú ê x ú
λ=6
ê1 2 - 3 úù êë úû êë 2 úû
Example: Given A = êê0 3 2úú The eigen For λ = 1
ê0 0 - 2úûú é0 2ù éx ù é0ù
êë ê úê 1ú ê ú
values of 3A3 + 5A2 - 6A + 2I is? ê0 1ú êx ú = ê0ú
ëê ûú ëê 2 ûú ëê úû
Solution: Since the matrix is upper triangular, 2x2 = 0 ...(1)
diagonal elements represent eigen values.
x2 = 0...(2)
λ = 1,3, -2
Thus, there is only one independent variable
If λ is eigen value of A, then eigen value of 3A3 x1
+ 5A2 - 6A + 2I is 3λ3 + 5λ2 - 6λ + 2
Since only one non-zero column, r = 1, n = 2
λ = 1 ⇒ 3(1)3 + 5(1)2 - 6 (1) + 2 = 4 If r < n, n-r = 2 - 1 =1
λ = 3 ⇒ 3(3)3 + 5(3)2 - 6 × 3 + 2 = 110 Thus, only one variable can 1 have arbitrary
λ = -2 ⇒ 3(-2) + 5(-2) - 6 × -2 + 2 = 10
3 2
value.
Example: The eigen value of a 3 × 3 matrix éc ù é 1 ù é 1 ù é 1 ù
are given by 1, 2, 3. x 1 = êê úú = êê úú = êê úú or êê úú a = 0 or b = 0
êë0úû êë0úû êëa úû êëbúû
Find
For λ = 2, (A - λI)X = 0
(A) Tr(A2 + A-1 + adj A)
é ùé ù é ù
(B) Det(A2 + A-1 + adj A) ê-1 2ú êx 1 ú = ê0ú
ê 0 0ú êx ú ê0ú
Solution: eigen Values of A are λ = 1, 2, 3 êë úû êë 2 úû êë úû
-x1 + 2x2 = 0

28 Linear Algebra
 a=0
If  = 2
é 1 2ù é 1 ù é ù
ê ú ê ú = 2 ê1 ú
ê0 2 ú êbú ê ú
ëê ûú ëê ûú ëê ûú
x1 2 é 1 + 2bù é 2 ù
= ê ú ê ú
x2 1 ê 2b ú = ê2bú
ëê ûú ëê ûú
1 1 + 2b = 2
If x 1 = 1, then x2 =
2 2b = 1
éx ù é 1 ù
ê 1ú = ê ú 1
êx ú ê 1 / 2ú b=
êë 2 úû êë úû 2
1 1
i.e. b = \ a +b =
2 2
1 1
\ a +b = 0 + =
2 2 Cayley Hamilton Theorem
Method 2: Every square matrix satisfies its own
If X is the eigen vector then, AX = λX characteristic equation using Cayley Hamilton
λ = 1, 2 theorem, we can find Inverse of a matrix or
é1 2 ù é1 ù é ù Higher powers of a matrix
If  = 1 Þ êê ú ê ú = 1 ê1 ú é4 2ù
úê ú ê ú Example: Consider matrix A = êê ú
ú Its
ëê0 2ûú ëêa ûú ëê ûú 2 4
é 1 + 2a ù é 1 ù ëê ûú
ê ú ê ú characteristic equation is λ - 8λ + 12 = 0
2
ê 2a ú = ê ú
ëê ûú ëê ûú By Cayley Hamilton theorem we have A2 - 8A
1 + 2a = 1 + 12I = 0
a=0
If  = 2 Note: The constant term of any characteristic
é 1 2ù é 1 ù é ù equation of any order square matrix is equal
ê ú ê ú = 2 ê1 ú to the determinant of a matrix.
ê0 2 ú êbú ê ú
êë úû êë úû êë úû We can compute positive and negative
é 1 + 2bù é 2 ù powers of A as,
ê ú ê ú
ê 2b ú = ê2bú
ëê ûú ëê ûú
1 +Positive
2b = 2 Power of A Negative Power of A
2b = 1
Consider A2 - 8A + 12I = 0 A2 - 8A + 12I = 0
1
b=A2 = 8A - 12I 12I = 8A - A2
2
A3 = 8A21 - 12A 1 é 2
\ a 4+ b = 3 I= -A + 8A ùú
A = 8A2 - 12A2 12 êë û
1 é-A + 8Iù
............................... A -1 =
12 ëê ûú
............................... 1 é
A -2 = -I + 8A-1 ùú
............................... 12 êë û
1 é-A-1 + 8A-2 ù
A -3 =
12 ëê ûú
...............................
Characteristic equation is A - = 0
...............................
a - b
i.e.
Table 1.1 =0
c d -
éa bùú (a -)(d -) - bc = 0
Consider A 2´2 = êê
êëc dúúû ad - aλ - λd + λ2 - bc = 0
Characteristic equation is A - = 0 λ2 - λ(a + d) + (ad - bc) = 0
a - b 2 -êéTr (A )úù - A = 0 which is the
i.e. =0 ë û
c d - characteristic equation
(a -)(d -) - bc = 0
Linear Algebra 29
 According to Cayley Hamilton Theorem we Characteristic equation:
have We can replace λ → A
3 -2 éêTr (A )ùú +
ë û
A - A (Tr (A )) + A I = 0
2
éa a 12 a 13 ùú ïìï a 11 a 12
ê 11 a a a a ïü
Consider a 3 × 3 matrix A = êêa 21 a 22 a 23 úú í + 22 23 + 11 13 ïý + A = 0
ïï a 22 a 22 a 32 a 33 a 31 a 33 ïï
êa a 32 a 33 úûú î þ
ëê 31

Solved Examples

é1 2 ù Solution: By cayley theorem theorem

Example: Find A8 if A = ê ú
ê2 - 1ú using Cayley λ→P
Hamilton theorem ëê úû
P3 + P2 + 2P + I = 0
Solution: |A| = -1 -4 = -5
I = -(P3 + P2 + 2P)
Characteristic equation:
P-1I = -(P2 + P + 2I)
2 -(1 + (-1)) + -5 = 0
P-1 = -(P2 + P + 2I)
é-5 - 3ù é ù
λ2- 5 = 0 Example: Given A = ê ú and I = ê 1 0ú
ê2 0 úúû ê0 1ú
By Cayley Hamilton theorem êë êë úû
A2 - 5I = 0 Express A3 as a linear polynomial?
A2 = 5I Solution: Characteristic equation is λ2 + 5λ
é 1 0ù é625 0 ù + 6I = 0
(A2)4 = (5I)4 = 625I = 625 êê ú=ê
ú ê
ú
ú
êë0 1úû êë 0 625úû A2 + 5A + 6I = 0
Example: The characteristic equation of a 3 A2 = -(5A + 6I) ........ (1)
× 3 matrix P is given by α(λ) = |λI - P| = λ3 + λ2 A3 = -(5A2 + 6A) = -5(-5A - 6I) -6A from (1)
+ 2λ + 1 = 0 , where I  Identity matrix. The A3=19A+30I
inverse of matrix P will be ______________ ?

Practice Questions

1 -2 -1 (D) Orthogonall
1. If A = 2 3 1 and 3. 3. If a matrix A is m x n and B is n x p, the
0 5 -2 number of multiplication operations and
addition operations needed to calculate
-11 - 9 1 the matrix AB, respectively, are:
adj (A ) = 4 - 2 - 3 then k = ? (A) mn2p, mpn
10 k 7 (B) mpn ( n-1 )
(C) mpn, mp (n-1)
(A) -5
(D) mn2p, (m + p) n
(B) 3
écos - sinúù é ù
(C) -3 4. The matrices êê & êa 0ú
êësin cosúúû ê0 bú
êë úû
(D) 5
commute under multiplication.
2. If AT = A-1 ,, where A iis a reall matriix ,,
then A is (A) If a = b or  = n, n is an integer
(A) Normall (B) Always
(B) Symmetriic (C) Never
(C) Hermiitiian (D) If a cos ¹ b sin

30 Linear Algebra
 é5 1ù é 1 5ù
ê - ú ê- ú
ê 13 13 ú ê 13 13 ú
(A) ê ú (C) ê ú
ê2 3 ú ê 2 3ú
ê ú ê ú
êë 13 13 úû êë 13 13 úû
5. The determinant of the matrix é2 5ù é 1 5ù
ê ú ê - ú
ê 13 13 ú ê 13 13 ú
é 1 úù (B) ê ú (D) ê ú
ê6 - 8 1
ê -1 3ú ê2 3ú
ê0 2 4 6úú ê ú ê ú
ê êë 13 13 úû êë 13 13 úû
ê0 0 4 8úú
ê 8. If A and B are real symmetric matrices
ê0 0 0 - 1 úúû
êë of order n then which of the following is
true.
(A) 11 (A) AAT = I
(B) -48 (B) A = A-1
(C) 0 (C) AB = BA
(D) -24 (D) (AB)T = BT AT
6. If A and B are two matrices and if AB é 1 - 4ù
9. The matrix êê ú is an inverse of the
ú
exists, ten BA exists ëê 1 - 5 ûú
(A) Only if A has an many rows as B has é5 - 4ù
columns matrix êê ú
ú
ëê 1 - 1 ûú
(B) 
Only if both A and B are square
(A) True
matrices
(B) False 2 1
(C) Only if A and B are skew symmetric 3 2
matrices 10. Give matrix L = 3 2 and M = then
L × M is 0 1
(D) Only if both A and B are symmetric 4 5
é-3 5 ù 8 1 6 5
7. The inverse of the matrix A = êê ú
ú
2 1 (A) 13 2 (B) 9 8
ëê úû
22 5 12 13
é5 1ù é 1 5ù
ê - ú ê- ú 2 13
ê 13 13 ú ê 13 13 ú 9 4
(A) ê ú (C) ê ú (C) 5 22 (D)
ê2 3 ú ê 2 3ú 0 5
ê ú ê ú 6 2
êë 13 13 úû êë 13 13 úû
é2 5ù é 1 5ù
ê ú ê - ú
ê 13 13 ú ê 13 13 ú
(B) ê ú (D) ê ú
ê -1 3ú ê2 3ú Answer Key
ê ú ê ú
êë 13 13 úû êë 13 13 úû
1 - (A) 2 - (D) 3 - (C) 4 - (A) 5 - (B)

6 - (A) 7 - (C) 8 - (D) 9 - True 10 - (B)

Linear Algebra 31