Textbook on Matrices for Students
Uploaded by
PrashikTextbook on Matrices for Students
Uploaded by
Prashik[Link]
co/
n a' s
ri s h
K TEXT BOOK on
M atrices
(For B.A. and [Link]. IInd year students of All Colleges affiliated to universities in Uttar Pradesh)
As per U.P. UNIFIED Syllabus
(w.e.f. 2012-2013)
By
A. R. Vasishtha A. K. Vasishtha
Retired Head, Dep’t. of Mathematics [Link]., Ph.D.
Meerut College, Meerut (U.P.) C.C.S. University, Meerut (U.P.)
(Kanpur Edition)
KRISHNA Prakashan Media (P) Ltd.
KRISHNA HOUSE, 11, Shivaji Road, Meerut-250 001 (U.P.), India
[Link]
Jai Shri Radhey Shyam
Dedicated
to
Lord
Krishna
Authors & Publishers
[Link]
P reface
This book on MATRICES has been specially written according to the latest
Unified Syllabus to meet the requirements of the B.A. and [Link]. Part-II
Students of all Universities in Uttar Pradesh.
The subject matter has been discussed in such a simple way that the students
will find no difficulty to understand it. The proofs of various theorems and
examples have been given with minute details. Each chapter of this book
contains complete theory and a fairly large number of solved examples.
Sufficient problems have also been selected from various university examination
papers. At the end of each chapter an exercise containing objective questions has
been given.
We have tried our best to keep the book free from misprints. The authors
shall be grateful to the readers who point out errors and omissions which, inspite
of all care, might have been there.
The authors, in general, hope that the present book will be warmly received
by the students and teachers. We shall indeed be very thankful to our colleagues
for their recommending this book to their students.
The authors wish to express their thanks to Mr. S.K. Rastogi, M.D.
and Mr. Sugam Rastogi, Executive Director and entire team of
KRISHNA Prakashan Media (P) Ltd., Meerut for bringing out this book in
the present nice form.
The authors will feel amply rewarded if the book serves the purpose for
which it is meant. Suggestions for the improvement of the book are always
welcome.
June, 2012 — Authors
[Link]
Syllabus
M atrices
U.P. UNIFIED (w.e.f. 2012-13)
B.A./[Link]. II nd Year–Paper-I st M.M. : 33 / 65
Section-A: Matrices
Unit-1: Symmetric and skew-symmetric matrices, Hermitian and skew-Hermitian
matrices, Orthogonal and unitary matrices, Triangular and diagonal matrices, Rank of a
matrix, Elementary transformations, Echelon and normal forms, Inverse of a matrix by
elementary transformations.
Unit-2: Characteristic equation, Eigen values and eigen vectors of a matrix, Cayley-
Hamilton's theorem and its use in finding inverse of a matrix, Application of matrices to
solve a system of linear (both homogeneous and non-homogeneous) equations,
Consistency and general solution, Diagonalization of square matrices with distinct eigen
values, Quadratic forms.
[Link]
B rief C ontents
Dedication.........................................................................(v)
Preface ...........................................................................(vi)
Syllabus ........................................................................(vii)
Brief Contents ...............................................................(viii)
Section–B: Matrices................................................M-01 — M-184
1. Matrices .................................................................................................M-01 — M-36
2. Rank of a Matrix .................................................................................M-37 — M-82
3. Linear Equations ..............................................................................M-83 — M-124
4. Eigenvalues and Eigenvectors ....................................................M-125 — M-168
5. Quadratic Forms ..............................................................................M-169 — M-184
[Link]
SECTION
B
MATRICES
C hapters
1. Matrices
2. Rank of a Matrix
1.
3. Linear Equations
1.
4. Eigenvalues and Eigenvectors
4.
1.
5. Quadratic Forms
4.
[Link]
M-3
1
Matrices
1.1 Some Basic Concepts
onsider the system of equations
C 2 x + 9 y + 7 z = 4, 3 x + 4 y − 3 z = 5,
6 x + 8 y − 3 z = 1, 4 x − 2 y + z = 2.
Here x, y and z are unknowns and their coefficients are all numbers. Arranging the
coefficients in the order in which they occur in the equations and enclosing them in
square brackets, we obtain a rectangular array of the form
2 9 7
3 4 −3
6 .
8 −3
4 −2 1
[Link]
M-4
This rectangular array is an example of a matrix. The horizontal lines (→ ) are called
()
rows or row vectors, and vertical lines ↓ are called Columns or column vectors of the
matrix. There are 4 rows and 3 columns in this matrix. Therefore it is a matrix of the
type 4 × 3. The numbers 3, 4, − 3, 2 etc. constituting this matrix are called its
elements. The difference between a matrix and a number should be clearly
understood. A matrix is not a number. It has got no numerical value. It is a new thing
formed with the help of numbers. It is just an ordered collection of numbers arranged
in the form of a rectangular array. Simply 7 is a number. But in our notation of
matrices [7] is a matrix of the type 1 × 1and we cannot have 7 = [7]. We cannot have a
relation of equality between a matrix and a number.
We shall use capital letters (in bold type or in italic type) to denote matrices.
0 0 0
5 0 1
Thus A= and B = 0 0 0
6 1 72 × 3 0
0 0 3 × 3
are both matrices. They are of the type 2 × 3 and 3 × 3 respectively.
Sometimes we also use the brackets ( ) or the double bars, , in place of square
brackets [ ] to denote matrices.
2 2 3 + 5 i 9 7 7
Thus A= , B = ,C = ,
2 2 −4 3 − 5 i 7 7
are all matrices each of the type 2 × 2.
1.2 Matrix
Definition: A set of mn numbers (real or complex) arranged in the form of a rectangular array
having m rows and n columns is called an m × n matrix [ to be read as ‘m by n’ matrix. ].
An m × n matrix is usually written as
a11 a12 … a1n
a21 a22 … a2 n
a a32 … a3 n
31
A=
… … … …
… … … …
am1 am2 … amn
[Link]
M-5
In a compact form the above matrix is represented by A = [aij], i = 1, 2,…, m, j = 1, 2,… n
or simply by [aij]m × n. We write the general element of the matrix and enclose it in
brackets of the type [ ] or of the type ( ).
The numbers a11, a12 etc. of this rectangular array are called the elements of the
matrix. The element aij belongs to the ith row and the j th column and is sometimes
called the (i, j)th element of the matrix . Thus in the element aij the first suffix i will
always denote the number of the row and the second suffix j, the number of the column
in which the element occurs. In a matrix, the number of rows and the columns need
not be equal.
1.3 Special Types of Matrices
(i) Square Matrix
Definition: An m × n matrix for which m = n (i. e., the number of rows is equal to the number of
columns)is called a square matrix of order n. It is also called an n-rowed square matrix.
Thus in a square matrix, we have the same number of rows and columns. The elements
aij of a square matrix A = [aij]n × n, for which i = j i. e., the elements a11, a22 , a33 ,…, ann are
called the diagonal elements and the line along which they lie is called the principal
diagonal of the matrix. For example, the matrix
0 1 2 3
2 3 1 0
A=
5 0 1 1
0 0 1 24 × 4
is a square matrix of order 4. The elements 0, 3, 1, 2 constitute the principal diagonal
of this matrix.
(ii) Unit Matrix or Identity Matrix
Definition: A square matrix each of whose diagonal elements is 1 and each of whose
non-diagonal elements is equal to zero is called a unit matrix or an identity matrix and is
denoted by I. In will denote a unit matrix of order n. Thus a square matrix A = [aij] is a unit
matrix if aij = 1 when i = j and aij = 0 when i ≠ j.
1 0 0
1 0
For example, I3 = 0 1 0 and I2 =
0 0 1
0 1
[Link]
M-6
are unit matrices of order 3 and 2 respectively.
(iii) Null Matrix or Zero Matrix
Definition: The m × n matrix whose elements are all 0 is called the null matrix (or zero
matrix) of the type m × n. It is usually denoted by O or more clearly by Om, n . Often a null matrix
is simply denoted by the symbol 0 read as ‘zero’.
0 0 0 0 0 0 0
For example, 0 0 0 0 and 0 0 0
0 0 0 0 3 × 4 0 0 0 3 × 3
are zero matrices of the types 3 × 4 and 3 × 3 respectively.
(iv) Row Matrices and Column Matrices
Definition: Any 1 × n matrix which has only one row and n columns is called a row matrix or
row vector. Similarly any m × 1 matrix which has m rows and only one column is a column
matrix or a column vector.
For example, X = [2 7 − 8 5 11]1 × 5 is a row matrix of the type 1 × 5 while
2
Y = −9 is a column matrix of the type 3 × 1.
11
3 × 1
1.4 Submatrices of a Matrix
Definition: Any matrix obtained by omitting some rows and columns from a given (m × n)
matrix A is called a submatrix of A.
The matrix A itself is a sub-matrix of A as it can be obtained from A by omitting no rows
or columns.
A square submatrix of a square matrix A is called a principal submatrix, if its diagonal
elements are also the diagonal elements of the matrix A. Principal submatrices are
obtained only by omitting corresponding rows and columns.
1 2 3 9
1 2 3
Example: The matrix is a submatrix of the matrix A = 7 11 6 5 as
0 2 1 0
2 1 8
it can be obtained from A by omitting the second row and the fourth column.
[Link]
M-7
1.5 Equality of Two Matrices
Definition: Two matrices A = [aij] and B = [bij] are said to be equal, if
(i) they are of the same size and
(ii) the elements in the corresponding places of the two matrices are the same i.e., aij = bij for
each pair of subscripts i and j.
If two matrices A and B are equal, we write A = B. If two matrices A and B are not equal,
we write A ≠ B. If two matrices are not of the same size, they cannot be equal.
1.6 Addition of Matrices
Definition: Let A and B be two matrices of the same type m × n. Then their sum (to be denoted
by A + B) is defined to be the matrix of the type m × n obtained by adding the corresponding
elements of A and B. Thus if A = [aij]m × n and B = [bij]m × n, then A + B = [aij + bij]m × n.
Note that A + B is also a matrix of the type m × n.
More clearly we can say that, if
a11 a12 … a1n
a21 a22 … a2 n
A=
… … … …
am1 am2 … amn m × n
b11 b12 … b1n
b21 b22 … b2 n
and B=
… … … …
bm1 bm2 … bmn m × n
a11 + b11 a12 + b12 … a1n + b1n
a21 + b21 a22 + b22 … a2 n + b2 n
then A +B=
…… …… … ……
am1 + bm1 am2 + bm2 … amn + bmn m × n
For example, if
3 2 −1 1 −2 7
A= and B = ,
4 −3 12 × 3 3 2 −12 × 3
[Link]
M-8
3 +1 2−2 − 1 + 7 4 0 6
then A+B= = .
4 + 3 −3+2 1 − 1 7 −1 0 2 × 3
Important Note: It should be noted that addition is defined only for matrices which
are of the same size. If two matrices A and B are of the same size, they are said to be
conformable for addition. If the matrices A and B are not of the same size, we cannot
find their sum.
1.7 Properties of Matrix Addition
(i) Matrix addition is commutative: If A, B, C be two m × n matrices, then
A + B = B + A.
(ii) Matrix addition is associative: If A, B, C be three matrices each of the type m × n,then
(A + B) + C = A + (B + C).
(iii) Existence of additive identity: If O be the m × n matrix each of whose elements is
zero, then A + O = A = O + A for every m × n matrix A.
(iv) Existence of the additive inverse.
Negative of matrix. Definition: Let A = [aij]m × n. Then the negative of the matrix A is
defined as the matrix [− aij]m × n and is denoted by –A.
The matrix − A is the additive inverse of the matrix A. Obviously,
− A + A = O = A + ( − A).
Here O is the null matrix of the type m × n. It is identity element for matrix addition.
Subtraction of two matrices. Definition:
If A and B are two m × n matrices, then we define A − B = A + ( − B).
Thus the difference A − B is obtained by subtracting from each element of A the
corresponding element of B.
(v) Cancellation laws hold good in the case of addition of matrices i.e., if A, B, C
are three m × n matrices, then
A + B = A + C ⇒ B =C (left cancellation law)
and B + A =C + A ⇒ B =C (right cancellation law)
(vi) The equation A + X = O has a unique solution X = − A in the set of all
m × n matrices.
[Link]
M-9
1.8 Multiplication of a Matrix by a Scalar
Definition: Let A be any m × n matrix and k any complex number called scalar. The m × n
matrix obtained by multiplying every element of the matrix A by k is called the scalar multiple of A
by k and is denoted by kA or Ak. Symbolically, if A = [aij]m × n, then kA = Ak = [kaij]m × n.
3 2 −1
For example, if k = 2 and A = ,
4 −3 12 × 3
2 × 3 2×2 2 × − 1 6 4 −2
then 2A = = .
2 × 4 2× −3 2 × 1 8 −6 22 × 3
Properties of Multiplication of a Matrix by a Scalar.
Theorem 1: If A and B are two matrices each of the type m × n, then k (A + B) = kA + kB i. e.,
scalar multiplication of matrices distributes over the addition of matrices.
Theorem 2: If p and q are two scalars and A is any m × n matrix, then
( p + q)A = pA + qA.
Theorem 3: If p and q are two scalars and A is any m × n matrix, then
p (qA) = ( pq)A.
Theorem 4: If A be any m × n matrix and k be any scalar, then
(− k )A = − (kA) = k (− A).
Theorem 5: If A be any m × n matrix, then
(i) 1 A = A (ii) (−1)A = − A.
Theorem 6: If A and B are two m × n matrices, then − (A + B) = − A − B.
1.9 Multiplication of Two Matrices
Definition: Let A = [aij]m × n and B = [b jk ]n × p be two matrices such that the number of
columns in A is equal to the number of rows in B. Then the m × p matrix C = [c ik ]m × p such that
n
c ik = ∑ aij b jk [ Note that the summation is with respect to the repeated suffix j ] is called the
j =1
product of the matrices A and B in that order and we write C = AB.
[Link]
M-10
In the product AB, the matrix A is called the pre-factor and the matrix B is called the
post-factor. Also we say that the matrix A has been post-multiplied by the matrix B.
and the matrix B has been pre-multiplied by the matrix A.
Explanation to understand the above definition. The product AB of two matrices
A and B exists if and only if the number of columns in A is equal to the number of rows
in B. Two such matrices are said to be conformable for multiplication. If A is an
m × n matrix and B is an n × p matrix, then AB is an m × p matrix. Further, if A = [aij]m × n
and B = [b jk ]n × p , then AB = [c ik ]m × p , where
n
c ik = ∑ aij b jk = ai1b1k + bi2 b2 k + … ain bnk
j =1
i.e., the (i, k )th element c ik of the matrix AB is obtained by multiplying the
corresponding elements of the ith row of A and the k th column of B and then adding the
products. The rule of multiplication is row-by-column multiplication i. e., in the
process of multiplication we take the rows of A and the columns of B. The element c11
of the matrix AB is obtained by adding the products of the corresponding elements of
the first row of A and the first column of B. The element c12 of the matrix AB is
obtained by adding the products of the corresponding elements of the first row of A
and the second column of B. Similarly the element c21 of the matrix AB is obtained by
adding the products of the corresponding elements of the second row of A and the first
column of B. In this way we multiply two matrices A and B.
a11 a12
b11 b12
For example, if A = a21 a22 ,B=
a b21 b22 2 × 2
31 a32 3 × 2
a11b11 + a12 b21 a11b12 + a12 b22
then AB = a21b11 + a22 b21 a21b12 + a22 b22
a b + a b a31b12 + a32 b22 3 × 2.
31 11 32 21
Important Note: If the product AB exists, then it is not necessary that the product BA
will also exist. For example, if A is a 4 × 5 matrix and B is a 5 × 3 matrix, then the
product AB exists while the product BA does not exist.
2 2 0 1 2 3 4
Example : If A = 3 2 1 and B = 2 0 1 2 then find AB . Does BA exist ?
1 0 1 3 1 0 5
[Link]
M-11
Solution : The matrix A is of the type 3 × 3 and the matrix B is of the type 3 × 4. Since the
number of columns of A is equal to the number of rows of B therefore AB is defined i. e.,
the product AB exists and it will be a matrix of the type 3 × 4.
c11 c12 c13 c14
Let AB = c21 c22 c23 c24 .
c c32 c33 c34
31
Then c11 = the sum of the products of the corresponding elements of the first row of A
and the first column of B.
c12 = the sum of products of the corresponding elements of the first row of A and the
second column of B.
c13 = the sum of the products of the corresponding elements of the first row of A and
the third column of B.
c23 = the sum of the products of the corresponding elements of the second row of A
and the third column of B.
c32 = the sum of the products of the corresponding elements of the third row of A and
the second column of B, and so on.
Therefore by the row by column rule of multiplication (Rows of A multiplied by the
columns of B ), we have
2 1 0 1 2 3 4
AB = 3 2 1 × 2 0 1 2
1 0 13 × 3 3 1 0 53 × 4
2.1 + 1.2 + 0.3 2.2 + 1.0 + 0.1 2.3 + 1.1 + 0.0 2.4 + 1.2 + 0.5
= 3.1 + 2.2 + 1.3 3.2 + 2.0 + 1.1 3.3 + 2.1 + 1.0 3.4 + 2.2 + 1.5
1.1 + 0.2 + 1.3 1.2 + 0.0 + 1.1 1.3 + 0.1 + 1.0 1.4 + 0.2 + 1.5
4 4 7 10
= 10 7 11 21
4 3 3 9 3 × 4.
Since the number of the columns of B is not equal to the number of rows of A, therefore
the product BA does not exist.
1.10 Properties of Matrix Multiplication
(i) Matrix multiplication is associative, if conformability is assured i.e., A (BC) = (AB) C if
A , B , C are m × n, n × p, p × q matrices respectively.
[Link]
M-12
(ii) Multiplication of matrices is distributive with respect to addition of matrices i. e.,
A (B + C) = AB + AC,
where A, B, C are any three, m × n, n × p, n × p matrices respectively.
(iii) The multiplication of matrices is not always commutative.
Whenever AB = BA , the matrices A and B are said to commute.
If AB = − BA the matrices A and B are said to anti-commute.
(iv) If A be any m × n matrix and On,p be an n × p null matrix, then AOn,p = Om,p where Om,p
is an m × p null matrix.
Similarly if Om,n be an m × n null matrix and A be any n × p matrix, then
Om,nA = Om,p .
If A be any n-rowed square matrix and O be an n-rowed null matrix, then AO = OA = O.
(v) The equation AB = O does not necessarily imply that at least one of the matrices A and B
must be a zero matrix.
Or
The product of two matrices can be a zero matrix while neither of them is a zero matrix
(vi) In the case of matrix multiplication if AB = O, then it does not necessarily imply that BA = O.
(vii) If A be an m × n matrix, In denotes the n-rowed unit matrix, it can be easily seen that
AIn = A = Im A.
1.11 Triangular, Diagonal and Scalar Matrices
(i) Upper Triangular Matrix. Definition : A square matrix A = [aij] is called an upper
triangular matrix if aij = 0 whenever i > j.
Thus in an upper triangular matrix all the elements below the principal diagonal are
zero.
a11 a12 a13 … a1n
0 a22 a23 … a2 n
For example 0 0 a33 … a3 n is an upper triangular matrix of the type n × n.
… … … …… 0
0 0 0 … ann
(ii) Lower Triangular Matrix. Definition: A square matrix A = [aij] is called a lower
triangular matrix if aij = 0 whenever i < j.
[Link]
M-13
Thus in a lower triangular matrix all the elements above the principal diagonal are
zero.
a11 0 0 … 0
a21 a22 0 … 0
For example a31 a32 a33 … 0 is a lower triangular matrix of the size n × n.
… … … … …
an1 an2 an3 … ann n × n
A triangular matrix A = [aij]n × n is called strictly triangular if aii = 0 for i = 1, 2,… n.
(iii) Diagonal matrix. Definition: A square matrix A = [aij]n × n whose elements above and
below the principal diagonal are all zero, i. e., aij = 0 for all i ≠ j, is called a diagonal matrix.
Thus a diagonal matrix is both upper and lower triangular. An n-rowed diagonal matrix
whose diagonal elements in order are d1, d2 , d3,,… dn will often be denoted by the
symbol Diag. [d1, d2 ,…, dn].
(iv) Scalar Matrix. Definition: A diagonal matrix whose diagonal elements are all equal is
called a scalar matrix.
k 0 … 0
0 k … M
If S = 0 0 … M
M M … M
0 0 … k
is an n-rowed scalar matrix each of whose diagonal elements is equal to k and A is any
n-rowed square matrix, then
AS = SA = kA.
i. e., the pre-multiplication or the post-multiplication of A by S has the same effect as
the multiplication of A by the scalar k. This is perhaps the motivation behind the name
‘scalar matrix’.
1.12 Idempotent, Involutary, Nilpotent and Periodic Matrices
Indempotent Matrix. A matrix such that A 2 = A is called idempotent matrix.
Example,
2 −2 −4
A = −1 3 4
1 −2 −3
[Link]
M-14
Involutary Matrix. A matrix A is said to be an involutary matrix if A2 = I ( unit
matrix). Example,
−5 −8 0
A= 3 5 0
1 2 −1
Since I2 = I always, we conclude that a Unit matrix is always involutary.
Nilpotent Matrix. A matrix is said to be a nilpotent matrix if A k = 0 if ( null matrix)
where k is a +ve integer; if however k is the least + ve integer for which A k = 0, then k is
called as the index of the nilpotent matrix. Examples,
ab b2
A=
2
is nilpotent of index 2.
− a − ab
1 −3 4
A = −1 3 4 is nilpotent of index 2.
1 −3 −4
Periodic Matrix. A matrix A is said to be a periodic matrix if A k + 1 = A, where k is a
+ve integer. If k is the least +ive integer for which A k + 1 = A then k is said to be period
of A. If we choose k = 1, then A2 = A and we call it to be an idempotent matrix.
1.13 Determinant of a Square Matrix
Let A be any square matrix. The determinant formed by the elements of A is said to be
the determinant of matrix A. This is denoted by| A|or det A. Since in a determinant the
number of rows is equal to the number of columns, therefore only square matrices
can have determinants.
12 0 13 12 0 13
Hence, if A = 15 12 11, then det A = | A| = 15 12 11
13 11 14 13 11 14
Difference between a matrix and a determinant.
(i) A matrix “A” cannot be reduced to a number whereas the determinant can be
reduced to a number.
(ii) The number of rows may or may not be equal to number of columns in a matrix
while in a determinant the number of rows is equal to the number of columns.
[Link]
M-15
(iii) Interchanging the rows and columns, a different matrix is formed while in a
determinant, an interchange of rows and columns does not change the value of
the determinant.
1.14 Non-singular and Singular Matrices
Definition. A square matrix A is said to be non-singular or singular according as
| A| ≠ 0 or | A| = 0
1.15 Transpose of a Matrix
Definition: Let A = [aij]m × n. Then the n × m matrix obtained from A by changing its rows
into columns and its columns into rows is called the transpose of A and is denoted by the symbol A′
or AT .
The operation of interchanging rows with columns is called transposition.
Symbolically if
A = [aij]m × n ,
then A′ = [b ji]n × m , where b ji = aij,
i. e., the ( j, i)th element of A′ is the (i, j)th element of A.
For example, the transpose of the 3 × 4 matrix
1 2 3
1 2 3 4
2 3 4
A = 2 3 4 1 is the 4 × 3 matrix A′ = .
3 3 4 2
4 2 13 × 4
4 1 14 × 3
The first row of A is the first column of A′. The second row of A is the second column of
A′ . The third row of A is the third column of A′.
Theorems: If A′ and B′ be the transposes of A and B respectively, then
(i) (A′ )′ = A;
(ii) (A + B)′ = A ′ + B ′ , A and B being of the same size.
(iii) (kA)′ = kA ′ , k being any complex number.
(iv) (AB)′ = B ′ A ′ , A and B being conformable to multiplication.
The above law (iv) is called the reversal law for transposes i. e., the transpose of the product
is the product of the transposes taken in the reverse order.
[Link]
M-16
1.16 Orthogonal Matrix
Orthogonal Matrix. Definition: A square matrix A is said to be orthogonal if A ′ A = I.
(Kanpur 2009)
If A is an orthogonal matrix, then A ′ A = I
⇒ | A ′ A| = | I | ⇒ | A ′| ⋅| A| = 1 [∵ det (AB) = ( det A). (det B)]
⇒ | A| ⋅| A| = 1 [∵| A ′| = | A|]
⇒ | A|2 = 1 ⇒ | A| = ± 1 ⇒ | A| ≠ 0
⇒ A is invertible.
Also then A ′ A = I ⇒ A ′ = A −1 which in turn implies AA ′ = I.
Thus A is an orthogonal matrix if and only if
A ′ A = I = AA ′ .
Theorem. If A, B be n-rowed orthogonal matrices,AB and BA are also orthogonal matrices.
Proof: Since A and B are both n-rowed square matrices, therefore AB is also an n-rowed
square matrix.
Since | AB | = | A| ⋅| B | and | A| ≠ 0 , also | B|1 ≠ 0 , therefore |AB | ≠ 0. Hence, AB is a
non-singular matrix.
Now (AB)′ = B ′ A ′.
∴ (AB)′ (AB) = (B ′ A ′ ) (AB)
= B ′ (A ′ A) B
= B ′ IB [∵ A ′ A = I]
= B′ B
= I. [∵ B ′ B = I]
∴ AB is orthogonal. Similarly we can prove that BA is also orthogonal.
0 2β γ
Example 1: Determine the values of α, β, γ when α β − γ is orthogonal.
α −β γ
[Link]
M-17
0 2β γ
Solution : Let A = α β − γ .
α −β γ
0 α α
Then A′ = 2β β −β
γ −γ γ
If A is orthogonal, then AA ′ = I.
0 2β γ 0 α α 1 0 0
α β − γ 2β β −β = 0 1 0
α −β γ γ −γ γ 0 0 1
4β2 + γ 2 2β2 − γ 2 −2β2 + γ 2 1 0 0
2 2 2 2 2 2 2 2
or 2β − γ α +β + γ α − β − γ = 0 1 0
0
2
−2β + γ
2 2
α −β − γ 2 2
α 2 + β2 + γ 2 0 1
Now equating the corresponding elements, we get
4β2 + γ 2 = 1 …(1)
2β2 − γ 2 = 0 …(2)
α2 + β2 + γ 2 = 1. …(3)
1 1
From (1) and (2), we get β = ± ,γ = ± .
6 3
1
From (3), we get α = ± .
2
1 1 1
Hence α=± ,β = ± ,γ = ± .
2 6 3
Example 2: Show that the matrix
cos θ sin θ
is orthogonal.
− sin θ cos θ
cos θ sin θ
Solution: Let A= .
− sin θ cos θ
cos θ − sin θ
Then A′ = .
sin θ cos θ
[Link]
M-18
cos θ sin θ cos θ − sin θ 1 0
We have AA′ = = = I2 .
− sin θ cos θ sin θ cos θ 0 1
Hence the matrix A is orthogonal.
Example 3: Verify that the matrix
1 2 2
1
2 1 −2 is orthogonal.
3
−2 2 −1
1 2 2
1
Solution: Let A= 2 1 −2 .
3
−2 2 −1
1 2 −2
1
Then A′ = 2 1 2
3
2 −2 −1
1 2 2 1 2 −2
1
We have AA′ = 2 1 −2 2 1 2
9
−2 2 −1 2
−2 −1
9 0 0 1 0 0
1
= 0 9 0 = 0 1 0 = I3 .
9 0
0 9 0 0 1
Hence, the matrix A is orthogonal.
1.17 Conjugate of a Matrix
If i = (−1 ), then z = x + iy is called a complex number where x and y are any real
numbers. If z = x + iy, then z = x − iy is called the conjugate of the complex number z.
We have zz = ( x + iy) ( x − iy) = x2 + y2 i. e., is real.
Also if z = z , then x + iy = x − iy i. e., 2 iy = 0 i. e., y =0 i. e., z is real.
Conversely, if z is real then z = z .
If z = x + iy, then z = x − iy. ∴ (z ) = x + iy = z .
If z1 and z2 are two complex numbers, then it can be easily seen that
(i) z1 + z2 = z1 + z2 and (ii) z1z2 = (z1) (z2 ).
[Link]
M-19
Conjugate of a Matrix
Definition: The matrix obtained from any given matrix A on replacing its elements by the
corresponding conjugate complex numbers is called the conjugate of A and is denoted by A.
Thus if A = [aij]m × n, then A = [aij]m × n where aij denotes the conjugate complex of aij.
If A be a matrix over the field of real numbers, then obviously A coincides with A.
2 + 3 i 4 − 7i 8 2 − 3 i 4 + 7i 8
Example: If A = , then A = .
−i 6 9 + i i 6 9 − i
Theorem. If A and B be the conjugates of A and B respectively, then
(i) (A) = A
(ii) (A + B) = A + B, A and B being of the same size
(iii) (k A) = k A, k being any complex number
(iv) (AB) = A B, A and B being conformable to multiplication.
1.18 Transposed Conjugate of a Matrix
Definition: The transpose of the conjugate of a matrix A is called transposed conjugate of A and
is denoted byA θ or by A *.
Obviously the conjugate of the transpose of A is the same as the transpose of the
conjugate of A i. e.,
( A ′) = (A)′ = A θ .
If A = [aij]m × n, then A θ = [b ji]n × m
where b ji = aij i. e., the ( j, i)th element of A θ = the conjugate complex of the (i, j)th
element of A.
1 + 2 i 2 − 3i 3 + 4 i
Example: If A = 4 − 5 i 5 + 6i 6 − 7 i ,
8 7 + 8i 7
1 + 2 i 4 − 5i 8
then A′ = 2 − 3 i 5 + 6i 7 + 8 i
3 + 4 i 6 − 7i 7
[Link]
M-20
1 − 2 i 4 + 5i 8
θ
and (A ′ ) = A = 2 + 3 i 5 − 6i 7 − 8 i .
3 − 4 i 6 + 7i 7
Theorem: If A θ and B θ be the transposed conjugates of A and B respectively, then
(i) (A θ )θ = A
(ii) (A + B)θ = A θ + B θ , A and B being of the same size
(iii) (kA)θ = k A θ , k being any complex number
(iv) (AB)θ = B θ A θ , A and B being conformable to multiplication.
1.19 Symmetric and Skew-symmetric Matrices
Symmetric Matrix. Definition: A square matrix A = [aij] is said to be symmetric if its
(i, j)th element is the same as its ( j , i)th element i. e., if aij = a ji for all i, j.
a h g p
1 i −2 i
h b f q 2 4
For example, g , i −2 4 ,
f c r 4 3
−2 i 4 3
p q r s
are symmetric matrices.
Theorem 1: A necessary and sufficient condition for a matrix A to be symmetric is that A and A′
are equal.
Skew-symmetric matrix. Definition: A square matrix A = [aij] is said to be
skew-symmetric if the (i, j)th element of A is the negative of the ( j, i)th element of A i.e., if aij = − a ji
for all i, j.
If A is a skew-symmetric matrix, then
aij = − a ji [by definition]
∴ aii = − aii, for all values of i.
∴ 2 aii = 0 or aii = 0.
Thus the diagonal elements of a skew-symmetric matrix are all zero.
[Link]
M-21
0 h g 0 −3 i −4
For example, the matrices − h 0 f and 3 i 0 8 are skew-symmetric
− g −f 0 4 −8 0
matrices.
Theorem: A necessary and sufficient condition for a matrix A to be skew-symmetric is that
A ′ = − A.
1.20 Hermitian and skew-Hermitian Matrices
Hermitian Matrix. Definition: A square matrix A = [aij] is said to be Hermitian if the
(i, j)th element of A is equal to the conjugate complex of the ( j, i)th element of A i. e., if aij = a ji for all
i and j. (Lucknow 2006)
1 2 − 3i 3 + 4 i
a b + ic
For example, , 2 + 3 i 0 4 − 5 i are Hermitian matrices.
b − ic d 3 − 4 i
4 + 5i 2
If A is a Hermitian matrix, then aii = aii, by definition.
∴ aii is real for all i. Thus every diagonal element of a Hermitian matrix must be
real.
A Hermitian matrix over the field of real numbers is nothing but a real symmetric
matrix.
Theorem: A necessary and sufficient condition for a matrix A to be Hermitian is that A = A θ .
Skew-Hermitian Matrix. Definition: A square matrix A = [aij] is said to be
skew-Hermitian if the (i, j)th element of A is equal to the negative of the conjugate complex of the
( j, i)th element of A i.e.,
aij = − a ji for all i and j.
If A is a skew-Hermitian matrix, then
aii = − aii, by definition.
∴ aii + aii = 0
i. e., aii must be either a pure imaginary number or must be zero.
Thus the diagonal elements of a skew-Hermitian matrix must be pure imaginary numbers or zero.
(Lucknow 2008, 09)
[Link]
M-22
0 −2 − i −i 3 + 4i
For example, the matrices ,
2 − i 0 −3 + 4 i 0
are skew-Hermitian matrices. A skew-Hermitian matrix over the field of real numbers
is nothing but a real skew-symmetric matrix.
Theorem. A necessary and sufficient condition for a matrix A to be skew-Hermitian is that
A θ = − A.
Example 4: If A is a symmetric (skew-symmetric) matrix, then show that kA is also symmetric
(skew-symmetric).
Solution: (i) Let A be a symmetric matrix. Then A ′ = A.
We have (kA)′ = kA ′
= kA. [∵ A ′ = A]
Since (kA)′ = kA, therefore kA is a symmetric matrix.
(ii) Let A be a skew-symmetric matrix. Then A ′ = − A.
We have (kA)′ = kA ′ = k (− A) [∵ A ′ = − A]
= − (kA).
Since (kA)′ = − (kA), therefore kA is a skew-symmetric matrix.
Example 5: If A is a Hermitian matrix, show that iA is skew-Hermitian.
Solution: Let A be a Hermitian Matrix. Then A θ = A.
We have (iA)θ = i A θ [∵(kA)θ = k A θ ]
= (− i)A θ [∵ i = − i]
= − (iA θ )
= − (iA) [∵ A θ = A].
Since (iA)θ = − (iA), therefore iA is a skew-Hermitian matrix.
Example 6: If A and B are symmetric matrices, then show that AB is symmetric if and only if A
and B commute i. e., AB = BA. (Lucknow 2008)
Solution: It is given that A and B are two symmetric matrices. Therefore
A ′ = A and B ′ = B.
[Link]
M-23
Now suppose that AB = BA.
Then to prove that AB is symmetric.
We have (AB)′ = B ′ A ′
= BA [∵ A ′ = A, B ′ = B]
= AB. [∵ AB = BA]
Since (AB)′ = AB, therefore AB is a symmetric matrix.
Conversely suppose that AB is a symmetric matrix. Then to prove that
AB = BA.
We have AB = (AB)′ [∵ AB is a symmetric matrix]
= B ′ A ′ = BA.
Example 7: If A be any matrix, then prove that AA′ and A ′ A are both symmetric matrices.
Solution: Let A be any matrix.
We have (AA ′ )′ = (A ′ )′ A ′ [By the reversal law for transposes]
= AA ′ [∵ (A ′ )′ = A].
Since (AA ′ )′ = AA ′ , therefore AA′ is a symmetric matrix.
Again (A ′ A)′ = A ′ (A ′ )′ = A ′ A.
Since (A ′ A)′ = A ′ A, therefore A ′ A is a symmetric matrix.
Example 8: Show that the matrix B ′ AB is symmetric or skew-symmetric according as A is
symmetric or skew-symmetric. (Lucknow 2005)
Solution: Case I. Let A be a symmetric matrix. Then A ′ = A.
Now (B ′ AB)′ = B ′ A ′ (B ′ )′ , by the reversal law for the transposes
= B′ A ′ B [since (B ′ )′ = B]
= B ′ AB.
Hence B ′ AB is symmetric.
Case II. Let A be a skew-symmetric matrix.
Then A ′ = − A.
Now (B ′ AB)′ = B ′ A ′ (B ′ )′ = B ′ A ′ B = B ′ (− A) B
= − (B ′ A) B = − B ′ AB.
Hence B ′ AB is a skew-symmetric matrix.
[Link]
M-24
Example 9: Show that the matrix
i 3 + 2i −2− i
− 3 + 2 i 0 3 − 4i
2−i − 3 − 4i − 2 i
is skew-Hermitian.
Solution: Let us denote the given matrix byA.
Then A = conjugate of the matrix A
−i 3 − 2i − 2 + i
= − 3 − 2 i 0 3 + 4i
2+i − 3 + 4i + 2 i
−i −3 − 2 i 2+i
∴ A = (A)′ = 3 − 2 i 0 −3 + 4 i
−2 + i 3 + 4i 2 i
i 3 + 2i − 2 − i
= − − 3 + 2 i 0 3 − 4 i = − A.
2−i − 3 − 4i − 2 i
Since A θ = − A, therefore the matrix A is skew-Hermitian.
3 2 − 3i 3 + 5 i
Example 10: If A = 2 + 3 i 5 i , prove that A is a Hermitian matrix.
3 − 5 i −i 7
(Rohilkhand 2008, 10)
Solution: Let us denote the given matrix by A.
3 2 + 3i 3 − 5 i
Then A = 2 − 3 i 5 −i .
3 + 5 i i 7
3 2 − 3i 3 + 5 i
θ
∴ A = (A) ′ = 2 + 3 i 5 i =A
3 − 5 i −i 7
Since A θ = A, therefore the matrix A is Hermitian.
[Link]
M-25
− 2 + 3 i 1− i 2+i
Example 11: Express 3 4 − 5i 5 as the sum of a Hermitian and a
1 1+ i − 2 + 2 i
skew-Hermitian matrix.
Solution. If A is any square matrix, then we can write
1 1
A= (A + A θ ) + (A − A θ ),
2 2
1 1
where (A + A θ ) is a Hermitian matrix and (A − A θ ) is a skew-Hermitian matrix.
2 2
− 2 + 3 i 1− i 2+i
Let A= 3 4 − 5i 5
1 1+ i − 2 + 2 i
Then A = conjugate of the matrix A .
− 2 − 3 i 1+ i 2−i
= 3 4 + 5i 5
1 1− i − 2 − 2 i
− 2 − 3 i 3 1
θ
∴ A = (A)′ = 1 + i 4 + 5i 1− i
2−i 5 − 2 − 2 i
−4 4−i 3 + i
1 θ 1
Now (A + A ) = 4 + i 8 6 − i
2 2
3 − i 6+i − 4
1 3 1
−2 2−
2
i + i
2 2
1 1
= 2 + i 4 3 − i,
2 2
3 − 1 i 3+
1
i −2
2 2 2
which is a Hermitian matrix.
6i −2− i 1 + i
1 θ 1
Again (A − A ) = 2 − i − 10 i 4 + i
2 2
− 1 + i −4+ i 4 i
[Link]
M-26
1 1 1
3i − 1−
2
i + i
2 2
1 1
= 1− i −5 i 2 + i,
2 2
− 1 + 1 i −2+
1
i 2i
2 2 2
which is a skew-Hermitian matrix.
Example 12: Express the following matrix as the sum of a symmetric and a skew-symmetric
matrix :
1 2 4
−2 5 3.
−1 6 3
1 2 4 1 −2 −1
Solution: Let A = −2 5 3; so that A′ = 2 5 6.
−1 6 3 4 3 3
1 2 4 1 −2 −1 2 0 3
∴ A + A ′ = −2 5 3 + 2 5 6 = 0 10 9
−1 6 3 4 3 3 3 9 6
1 0 3 /2
1
or (A + A ′ ) = 0 5 9 / 2, which is a symmetric matrix.
2 3 / 2
9/2 3
1 2 4 1 −2 −1 0 4 5
Again A − A ′ = −2 5 3 − 2 5 6 = −4 0 −3
−1 6 3 4 3 3 −5 3 0
0 2 5/2
1
or (A − A ′ ) = −2 0 −3 / 2, which is a skew-symmetric matrix.
2 −5 / 2
3/2 0
1 0 3 / 2 0 2 5/2
Thus A= 0 5 9 / 2 + −2 0 −3 / 2,
3 / 2 9/2 3 −5 / 2 3/2 0
where the first matrix is symmetric and the second matrix is skew-symmetric.
[Link]
M-27
Example 13: Show that every square matrix is uniquely expressible as the sum of a symmetric
matrix and a skew-symmetric matrix.
(Purvanchal 2006, 09; Rohilkhand 07; Lucknow 11)
Solution. Let A be any square matrix. We can write
1 1
A= (A + A ′ ) + (A − A ′ ) = P + Q, say
2 2
1 1
where P= (A + A ′ ) and Q= (A − A ′ )
2 2
1 1
We have P′ = (A + A ′ ) ′ = (A + A ′ )′ [∵ (kA)′ = kA ′ ]
2 2
1
= { A ′ + (A ′ )′ } [∵(A + B)′ = A ′ + B ′ ]
2
1
= (A ′ + A) [∵ (A ′ )′ = A]
2
1
= (A + A ′ ) = P.
2
Therefore P is a symmetric matrix.
1 1 1
Again Q′ = (A − A ′ ) ′ = (A − A ′ )′ = { A ′ − (A ′ )′ }
2 2 2
1 1
= (A ′ − A) = − (A − A ′ ) = − Q.
2 2
Therefore Q is a skew-symmetric matrix.
Thus we have expressed the square matrix A as the sum of a symmetric and a
skew-symmetric matrix.
To prove that the representation is unique, let A = R + S be another such
representation of A, where R is symmetric and S skew-symmetric. Then to prove that
R = P and S = Q.
We have A ′ = (R + S)′ = R ′ + S ′ = R − S [∵ R ′ = R and S ′ = − S]
∴ A + A ′ = 2 R and A − A ′ = 2S.
1 1
This gives R= (A + A ′ ) and S = (A − A ′ ).
2 2
Thus R = P and S = Q.
Therefore the representation is unique.
[Link]
M-28
1.21 Adjoint or Adjugate of a Square Matrix
Definition: Let A = [aij]n × n be any n × n matrix. The transpose B′ of the matrix B = [ Aij]n × n,
where A ij denotes the cofactor of the element aij in the determinant| A|, is called the adjoint of the
matrix A and is denoted by the symbol adj A.
Thus the adjoint of a matrix A is the transpose of the matrix formed by the cofactors of
A i. e., if
a11 a12 … a1n
a21 a22 … a2 n
A= ,
… … … …
an1 an2 … ann
A11 A12 … A1n
A21 A22 … A2 n
then Adj A = the transpose of the matrix
… … … …
An1 An2 … Ann
A11 A21 … An1
A12 A22 … An2
= the matrix .
… … … …
A1n A2 n … Ann
Note: Sometimes the adjoint of a matrix is also called the adjugate of that matrix.
Theorem: If A is any square matrix of order n, then
A (adj A) = | A| I n = (adj A) A.
1.22 Invertible Matrices
Inverse or reciprocal of a matrix.
Definition: Let A be any n-rowed square matrix. Then a matrix B, if it exists, such that
AB = BA = In is called inverse of A.
Note: For the products AB, BA to be both defined and be equal, it is necessary that A
and B are both square matrices of the same order. Thus non-square matrices cannot
possess inverse.
Existence of the Inverse. Theorem. The necessary and sufficient condition for a square
matrix A to possess the inverse is that | A| ≠ 0 .
[Link]
M-29
1
Important. If A be an invertible matrix, then the inverse of A is Adj. A. It is usual to denote
| A|
the inverse of A by A −1 .
1
Thus, A −1 = Adj . A, provided | A| ≠ 0 .
| A|
Thus the necessary and sufficient condition for a matrix to be invertible is that it is
non-singular.
1.23 Reversal Law for the Inverse of a Product
Theorem 1: If A, B be two n-rowed non-singular matrices, then AB is also non-singular and
(AB) −1 = B −1 A −1, i. e., the inverse of a product is the product of the inverses taken in the reverse
order.
Theorem 2: If A be an n × n non-singular matrix, then (A ′ )−1 = (A −1)′ i. e., the operations of
transposing and inverting are commutative.
Theorem 3: If A be an n × n non-singular matrix, then (A −1)θ = (A θ )−1.
Example : Find the :
(i) Transpose (ii) Adjoint (iii) Inverse of the matrix :
2 −1 3
A = −5 3 1.
−3 2 3
2 −5 −3
T
Solution: (i) We have the transpose of A = A = −1 3 2 .
3 1 3
2 −1 3
(ii) We have | A| = −5 3 1
−3 2 3
= 2 (9 − 2) − (− 1) (− 15 + 3) + 3 (− 10 + 9)
= 14 − 12 − 3 = − 1.
Let A ij denote the cofactor of the element aij of | A| .
[Link]
M-30
3 1 −5 1
Then A11 = = 7, A12 = − = 12,
2 3 −3 3
−5 3
A13 = = − 1,
−3 2
−1 3 2 3
A21 = − = 9, A22 = = 15,
2 3 −3 3
2 −1
A23 = − = − 1,
−3 2
−1 3 2 3
A31 = = − 10 , A32 = − = − 17,
3 1 −5 1
2 −1
A33 = = 1.
−5 3
∴ The matrix B formed of the cofactors of the elements of |A| is,
7 12 −1
B= 9 15 −1 .
−10 −17 1
7 9 −10
Now adj. A (= the transpose of the matrix B) = 12 15 −17 .
−1 −1 1
7 9 −10
−1 1 1
(iii) We have A = adj. A = 12 15 −17
| A| (− 1) −1
1 1
7 9 −10 −7 −9 10
= (−1) 12 15 −17 = −12 −15 17 .
−1 −1 1 1 1 −1
Note: To check that the answer is correct the students should verify that AA −1 = I.
1.24 Unitary Matrix
Definition: A square matrix A is said to be unitary if A θ A = I.
Since | A θ| = | A| and | A θ A| = | A θ|| A|, therefore if A θ A = I, we have | A|| A| = 1.
[Link]
M-31
Thus the determinant of a unitary matrix is of unit modulus.
If A is a unitary matrix, then | A|| A θ| = 1 and so | A| ≠ 0 i. e., A is non-singular and so
invertible.
Hence A θ A = I implies AA θ = I.
Thus A is a unitary matrix iff A θ A = I = AA θ .
Theorem: If A, B be n-rowed unitary matrices, AB and BA are also unitary matrices.
Proof: Since A and B are both n-rowed square matrices, therefore AB is also an n-rowed
square matrix.
Since | AB| = | A|| B| and | A| ≠ 0 , also |B| ≠ 0 , therefore | AB| ≠ 0 .
Hence AB is a non-singular matrix.
Now (AB)θ = B θ A θ .
∴ (AB)θ (AB) = (B θ A θ ) (AB)
= B θ (A θ A) B
= B θ IB [∵ A θ A = I]
= Bθ B
= I. [∵ B θ B = I]
∴ AB is unitary. Similarly we can prove that BA is also unitary.
Example 14: Prove that
1 1 1+ i
B= is unitary.
3 1 − i −1 (Bundelkhand 2006)
1 1 1− i
Solution: We have B= .
3 1 + i −1
1 1 1+ i
∴ B θ = (B)′ = .
3 1 − i −1
[Link]
M-32
1 1 1 + i 1 1+ i
Now Bθ B =
3 1 − i −1 1 − i −1
1 1.1 + (1 + i) (1 − i) 1. (1 + i) + (1 + i) (−1)
=
3 (1 − i).1 + (−1).(1 − i) (1 − i) (1 + i) + (−1) (−1)
1 3 0 1 0
= = = I.
3 0 3 0 1
∴ B is unitary.
α + iγ −β + iδ
Example 15: Show that the matrix A = is a unitary matrix, if
β + iδ α − iγ
α2 + β2 + γ 2 + δ2 = 1.
α + iγ −β + iδ
Solution: We have A=
β + iδ α − iγ
α + iγ β + iδ
∴ A′ =
− β + iδ α − iγ
α − iγ β − iδ
Aθ = A′ = .
− β − iδ α + iγ
A square matrix A is said to be unitary if AA θ = I.
α + iγ −β + iδ α − iγ β − iδ 1 0
∴ =
β + iδ α − iγ − β − iδ α + iγ 0 1
α2 + β2 + γ 2 + δ2
or
αβ − iβγ + iαδ + γδ − αβ − iαδ + iβγ − δγ
αβ − iαδ + iβγ + γδ − αβ − i β α + iαδ − δγ 1 0
=
β2 + δ2 + α2 + γ 2 0 1
α2 + β2 + γ 2 + δ2 0 1 0
or =
0 α2 + β2 + γ 2 + δ2 0 1
if α2 + β2 + γ 2 + δ2 = 1.
Example 16: If A is a unitary matrix, show that A −1 is also unitary.
Solution: We have AA θ = A θ A = 1, since A is a unitary matrix.
[Link]
M-33
⇒ (AA θ )−1 = (A θ . A)−1 = (I)−1 [Taking inverse]
⇒ (A θ )−1. A −1 = A −1 (A θ )−1 = I
⇒ (A −1)θ . A −1 = A −1 (A −1)θ = I.
Hence A −1 is a unitary matrix.
Comprehensive Exercise 1
−1 2 2
1
1. (i) Show that the matrix 2 −1 2 is orthogonal.
3
2 2 −1
(ii) Verify that the matrix.
1 1 1
−
3 6 2
1 2
− 0 is orthogonal.
3 6
1 1 1
3 6 2
1 1 i
2. (i) Show that the matrix = is unitary.
2 − i −1
1 + i − 1 + i
2 is unitary.
(ii) Prove that the matrix 2
1+ i 1− i
2 2 (Rohilkhand 2010)
2 1 −1
3. (i) Let A = −1 2 1, find A + A ′ and A − A ′ and hence express A as the
3 4 1
sum of a symmetric and a skew-symmetric matrix.
(ii) Write the following matrix as the sum of a symmetric and a skew-symmetric
1 2 3
matrix : 4 5 6. (Lucknow 2010)
7 0 0
[Link]
M-34
0 6 7
4. (i) Show that the matrix −6 0 8 is skew-symmetric.
−7 −8 0
1/ 6 −2 / 6 1 / 6
(ii) Show that the matrix A ,where A = −2 / 6 4/6 −2 / 6 is symmetric.
1/ 6 −2 / 6 1 / 6
(Rohilkhand 2010)
1 1− i 2
5. If A = 1 + i 3 i , prove that A is a Hermitian matrix.
2 −i 0
(Luckhnow 2011; Bundelkhand 06)
6. Prove that the matrix A2 is symmetric if either A is symmetric or A is
skew-symmetric.
7. If A be any square matrix then show that A + A ′ is symmetric and A − A ′ is
skew-symmetric.
8. If A is a skew-Hermitian matrix, then show that iA is Hermitian.
9. If A, B are symmetric (skew-symmetric) matrices of the same order, then so is
also A + B.
10. Show that the matrix B θ AB is Hermitian or skew-hermitian according as A is
Hermitian or skew-Hermitian.
11. Show that all positive integral powers of a symmetric matrix are symmetric.
12. If A and B are symmetric matrices of order n, then show that AB + BA is
symmetric and AB − BA is skew-symmetric.
(Luckhnow 2008)
13. If A be any square matrix, prove that A + A θ , AA θ , A θ A are all Hermitian and
A − A θ is skew-Hermitian.
14. Show that every square matrix is uniquely expressible as the sum of a Hermitian
matrix and a skew-Hermitian matrix.
(Purvanchal 2007, 10; Lucknow 09)
15. Show that every square matrix A can be uniquely expressed as P + iQ where P and
Q are Hermitian matrices.
(Lucknow 2007, 10)
[Link]
M-35
A nswers 1
4 0 2 0 2 −4
3. (i) A + A ′ = 0 4 5; A − A ′ = −2 0 −3 ;
2 5 2 4 3 0
4 0 2 0 2 −4
1 1
A = 0 4 5 + −2 0 −3
2 2
2 5 2 4 3 0
1 3 5 0 −1 −2
(ii) A = 3 5 3 + 1 0 3
5 3 0 2 −3 0
Objective Type Questions
Multiple Choice Questions
Indicate the correct answer for each question by writting the corresponding letter from (a),
(b), (c) and (d).
1. Every diagonal element of skew-symmetric matrix is
(a) unity (b) zero
(c) non-zero (d) purely imaginary (Rohilkhand 2005)
2. The necessary and sufficient condition for a matrix to be invertible is that it is
(a) singular (b) non-singular
(c) zero matrix (d) none of these (Rohilkhand 2007)
Fill in the Blank(s)
Fill in the blanks “……” so that the following statements are complete and correct.
1. Every diagonal element of a Hermitian matrix must be ......
2. A square matrix A is said to be unitary if .......
3. A necessary and sufficient condition for a matrix A to be symmetric is that A and
A ′ are ....
[Link]
M-36
True or False
Write ‘T’ for true and ‘F’ for false statement.
1. The diagonal elements of a skew-Hermitian matrix must be pure imaginary
numbers or zero.
2. A necessary and sufficient condition for a matrix A to be skew-symmetric is that
A′ = − A
A nswers
Multiple Choice Questions
1. (b) 2. (b) 3. real
Fill in the Blank(s)
1. Aθ A =I 2. equal
True or False
1. T 2. T
o
[Link]
M-37
2
Rank of a Matrix
2.1 Submatrix of a Matrix
uppose A is any matrix of the type m × n. Then a matrix obtained by leaving some
S rows and columns from A is called a submatrix of A. In particular the matrix A
itself is a sub-matrix of A because it is obtained from A by leaving no rows or columns.
Minors of a matrix. We know that every square matrix possesses a determinant. If A
be an m × n matrix, then the determinant of every square sub-matrix of A is called a
minor of the matrix A. If we leave m − p rows and n − p columns from A, we shall get a
square submatrix of A of order p. The determinant of this square submatrix is called a
p-rowed minor of A.
For example, let
2 4 1 9 1
0 5 2 5 2
A=
1 9 7 3 4
3 −2 8 1 84 × 5.
In a determinant the number of rows is equal to the number of columns. Therefore
there can be no 5-rowed minor of A.
[Link]
M-38
If we leave any columns from A, we shall get a square sub-matrix of A of order 4.
Thus 2 4 1 9 2 4 9 1
0 5 2 5 0 5 5 2
, , etc.
1 9 7 3 1 9 3 4
3 −2 8 1 3 −2 1 8
are 4-rowed minors of A.
If we leave two columns and one row from A, we shall get a square submatrix of A of
order 3.
Thus 2 4 1 4 1 9 5 2 5
0 5 2 , 5 2 5 , 9 7 3 , etc.
1 9 7 9 7 3 −2 8 1
are 3-rowed minors of A.
If we leave three columns and two rows from A , we shall get a square submatrix of A of
order 2.
Thus 2 4 4 1 5 2
, , , etc. are 2-rowed minors of A.
0 5 5 2 9 7
The numbers 2, 4, 1, 9, 1, 0, 5 etc. are all 1-rowed minors of A.
2.2 Rank of a Matrix
(Lucknow 2006)
Definition: A number r is said to be the rank of a matrix A if it possesses the following two
properties :
(i) There is at least one square submatrix of A of order r whose determinant is not equal to zero.
(ii) If the matrix A contains any square submatrix of order r + 1, then the determinant of every
square submatrix of A of order r + 1 should be zero.
In short the rank of a matrix is the order of any highest order non-vanishing minor of the matrix.
Thus the rank of a matrix A is the order of any highest order square submatrix of A whose
determinant is not equal to zero.
We shall denote the rank of a matrix A by the symbol ρ (A).
[Link]
M-39
It is obvious that the rank r of an (m × n) matrix can at most be equal to the smaller of
the numbers m and n , but it may be less.
If there is a matrix A which has at least one non-zero minor of order n and there is no
minor of A of order n + 1, then the rank of A is n. Thus the rank of every non-singular
matrix of order n is n. The rank of a square matrix A of order n can be less than n if and
only if A is singular i. e., |A| = 0.
Note 1 : Since the rank of every non-zero matrix is ≥ 1, we agree to assign the rank, zero, to
every null matrix :
Note 2 : Every (r + 1)-rowed minor of a matrix can be expressed as a linear combination
of its r-rowed minors. Therefore if all the r-rowed minors of a matrix are equal to zero,
then obviously all its (r + 1)-rowed minors will also be equal to zero.
Important : The following two simple results will help us very much in finding the
rank of a matrix:
(i) The rank of a matrix is ≤ r, if all (r + 1) - rowed minors of the matrix vanish.
(ii) The rank of a matrix is ≥ r, if there is at least one r-rowed minor of the matrix which is not
equal to zero.
Example :
1 0 0
(a) Let A = I3 = 0 1 0 be a unit matrix of order 3.
0 0 1
We have |A| = 1. Therefore A is a non-singular matrix. Hence rank A = 3. In
particular, the rank of a unit matrix of order n is n.
0 0 0
(b) Let A = 0 0 0 .
0 0 0
Since A is a null matrix, therefore rank A = 0.
1 2 3
(c) Let A = 2 3 4 .
0 2 2
We have | A| = 1 (6 − 8) − 2 (4 − 6) = 2 ≠ 0 . Thus A is a non-singular matrix. Therefore
rank A = 3.
[Link]
M-40
1 2 3
(d) Let A = 3 4 5 .
4 5 6
We have | A| = 1 (24 − 25) − 2 (18 − 20 ) + 3 (15 − 16) = 0
Therefore the rank of A is less than 3. Now there is at least one minor of A of order 2,
1 2
namely which is not equal to zero. Hence rank A = 2.
3 4
3 1 2
(e) Let A = 6 2 4 .
3 1 2
We have | A| = 0 , since the first two columns are identical.
Also each 2-rowed minor of A is equal to zero. But A is not a null matrix. Hence rank
A = 1.
2 4 3 2
(f) Let A= .
3 5 1 4
2 4
Here we see that there is at least one minor of A of order 2 i. e., which is not
3 5
equal to zero. Also there is no minor of A of order greater than 2. Hence the rank of
A = 2.
Echelon form of a matrix. Definition. A matrix A is said to be in Echelon form if :
(i) Every row of A which has all its entries 0 occurs below every row which has a non-zero entry.
(ii) The number of zeros before the first non-zero element in a row is less than the number of such
zeros in the next row.
Important result. The rank of a matrix in Echelon form is equal to the number of non-zero
rows of the matrix.
0 1 2 3
Example: Find the rank of the matrix 0 0 1 −1 .
0 0 0 0
The matrix A has one zero row. We see that it occurs below every non-zero row.
[Link]
M-41
Further the number of zeros before the first non-zero element in the first row is one.
The number of zeros before the first non-zero element in the second row is two. Thus
the number of zeros before the first non-zero element in any row is less than the
number of such zeros in the next row.
Thus the matrix A is in Echelon form.
∴ rank A = the number of non-zero rows of A = 2.
2.3 Theorem
The rank of the transpose of a matrix is the same as that of the original matrix.
Proof. Let A be any matrix and A′ be the transpose of the matrix A. Let rank A = r and
rank A′ = s. Then to prove that r = s.
We have rank A = r
⇒ there exists at least one r-rowed square submatrix, say R, of A such that
| R ′| = | R| ≠ 0
⇒ there exists at least one r-rowed square submatrix R′ of A ′ such that
| R ′| = | R| ≠ 0
⇒ rank A′ ≥ r ⇒ s ≥ r. ...(1)
Since (A ′ )′ = A, therefore interchanging the roles of A and A′ in the above result (1), we
have
r ≥ s. …(2)
From (1) and (2), we conclude that r = s. Hence the result.
Example 1: Find the ranks of the following matrices :
1 2 3
1 2 3
(i) 2 1 0 , (ii)
0 2 4 5
1 2 (Bundelkhand 2009)
[Link]
M-42
1 2 3
Solution : (i) Let A = 2 1 0 .
0 1 2
We have | A| = 1 (2 − 0 ) − 2 (4 − 0 ) + 3 (2 − 0 ), expanding along the first row
= 2 − 8 + 6 = 0.
1 2
But there is at least one minor of order 2 of the matrix A, namely which is not
2 1
equal to zero. Hence rank A = 2.
1 2 3
(ii) Let A= .
2 4 5
1 3
Here there is at least one minor of order 2 of the matrix A, namely which is not
2 5
equal to 0. Also there is no minor of the matrix A or order greater than 2. Hence rank
A = 2.
Example 2: Under what conditions the rank of the following matrix is 3
2 4 2
A = 2 1 2 ?
1 0 x
Solution: The rank of the given matrix A = 3.
∴ The minor of order 3 of matrix A ≠ 0
2 4 2
i. e., 2 1 2 ≠0
1 0 x
or 2 ( x − 0 ) − 4 (2 x − 2) + 2 (0 − 1) ≠ 0
or 2x − 8x + 8 − 2 ≠ 0
or − 6x + 6 ≠ 0
or x ≠ 1.
[Link]
M-43
1 5 4
Example 3: For which value of ‘b’ the rank of the matrix A = 0 3 2 is 2?
b 13 10
Solution: The rank of the given matrix A = 2.
∴ The minor of order 3 of matrix A = 0
1 5 4
i. e., 0 3 2 =0
b 13 10
or 1 (30 − 26) − 5 (0 − 2 b) + 4 (0 − 3 b) = 0
or 4 + 10 b − 12 b = 0
or 4 − 2b = 0
or b = 2.
Example 4: Find the values of a so that rank (A) < 3, where A is the matrix
3 a − 8 3 3
A= 3 3a − 8 3 .
3 3 3 a − 8
Solution: The rank of the given matrix A < 3
∴ The minor of order 3 of matrix A = 0
3 a − 8 3 3
i. e., 3 3a − 8 3 =0
3 3 3 a − 8
Applying R1 → R1 + R2 + R3 , we get
3a − 2 3a − 2 3a − 2
3 3a − 8 3 =0
3 3 3a − 8
1 1 1
or (3 a − 2) 3 3a − 8 3 =0
3 3 3a − 8
[Link]
M-44
Applying C2 → C2 − C1, C3 → C3 − C1, we get
1 0 0
(3 a − 2) 3 3 a − 11 0 =0
3 0 3 a − 11
or (3 a − 2) (3 a − 11) (3 a − 11) = 0
2 11 11
or a= , , .
3 3 3
Example 5: Prove that the points ( x1, y1), ( x2 , y2 ) and ( x3 , y3 ) are collinear if and only if the
rank of the matrix.
x1 y1 1
A = x2 y2 1 is less than three.
x y3 1
3 (Kanpur 2001; Bundelkhand 07)
Solution: The condition is necessary:
Given that the points ( xi, yi); i = 1, 2, 3 are collinear.
x1 y1 1
We are to prove that the rank of matrix A = x2 y2 1 is less than three.
x y3 1
3
The given points are collinear
⇒ The area of the triangle formed by these points is zero.
x1 y1 1
1
⇒ x2 y2 1 =0
2
x3 y3 1
x1 y1 1
⇒ x2 y2 1 =0
x3 y3 1
x1 y1 1
⇒ The rank of matrix x2 y2 1 is less than 3.
x3 y3 1
Hence the condition is necessary.
[Link]
M-45
The condition is sufficient:
x1 y1 1
The rank of the given matrix x2 y2 1 is less than 3.
x3 y3 1
We are to prove that the points ( xi, yi); i = 1, 2, 3 are collinear.
rank A < 3
x1 y1 1
⇒ x2 y2 1 =0
x3 y3 1
x1 y1 1
1
⇒ x2 y2 1 = 0.
2
x3 y3 1
The given points are collinear.
Hence the condition is sufficient.
Example 6: A is a non-zero column and B a non-zero row matrix, show that rank (AB) = 1 .
a11
a21
Solution : Let A = a31 and B = [b11 b12 b13 … b1n]
…
am1
be two non-zero column and row matrices respectively.
a11b11 a11b12 a11b13 … a11b1n
a21b11 a21b12 a21b13 … a21b1n
We have AB = .
… … … … …
am1b11 am1b12 am1b13 … am1b1n
Since A and B are non-zero matrices, therefore the matrix AB will also be non-zero. The
matrix AB will have at least one non-zero element obtained by multiplying
corresponding non-zero elements of A and B .
All the two-rowed minors of A obviously vanish. But A is a non-zero matrix. Hence
rank A = 1.
[Link]
M-46
Comprehensive E xercise 1
1. Determine the rank of each of the following matrices :
1 0 0 0 1 1 1 1
0 1 0 0 1 1 1 1
(i) (ii)
0 0 1 0 1 1 1 1
0 0 0 1 1 1 1 1
0 0 5 10 (Meerut 2006B)
(iii) (iv)
0 0 3 6
1 2 3 4 1 2 −7 5
(v) 2 4 6 8 (vi) 0 5 0 8
3 6 9 12 0 0 0 −3
1 2 3
(vii) .
2 4 5
2. Show that the rank of a matrix is ≥ the rank of every sub-matrix thereof.
3. Show that the rank of a matrix does not alter on affixing any number of
additional rows or columns of zeros.
0 1 0 0
0 0 1 0
4. If A = , find the rank of A and A2 .
0 0 0 1
0 0 0 0
5. Under what conditions the rank of the following matrix is 3 ? Is it possible for the
2 4 2
rank to be 1? Why ? A = 3 1 4.
1 0 x
(Kanpur 2010)
6. A is an n-rowed square matrix of rank (n − 1) , show that Adj. A is not a null matrix.
A nswers 1
1. (i) 4 (ii) 1 (iii) 0 (iv) 1 (v) 1
(vi) 3 (vii) 2
4. Rank A = 3, rank A2 = 2
[Link]
M-47
7
5. x≠ ; No, because one minor of order 2 of A is non-zero
5
2.4 Elementary Operations or Elementary Transformations
of a Matrix
An elementary transformation (or, an E-transformation) is an operation of any one of
the following types :
1. The interchange of any two rows(or columns).
2. The multiplication of the elements of any row (or column) by any non-zero number.
3. The addition to the elements of any other row (or column) the corresponding elements of any
other row (or column) multiplied by any number.
An elementary transformation is called a row transformation or a column
transformation according as it applies to rows or columns.
2.5 Symbols to be Employed for the Elementary
Transformations
The following notation will be used to denote the six elementary transformations :
1. The interchange of ith and j th rows will be denoted by Ri ↔ Rj.
2. The multiplication of the ith row by a non-zero number k will be denoted by
Ri → k Ri.
3. The addition of k times the j th row to the ith row will be denoted by Ri → Ri + kRj.
The corresponding column transformations will be denoted by writing C, in place of R
i. e., by Ci ↔ C j, Ci → k Ci, Ci → Ci + kC j respectively.
Important. It is quite obvious that if a matrix B is obtained from A by an elementary
transformation, A can also be obtained from B by an elementary transformation of the
same type.
For example, let
1 4 2 9
A = 2 5 1 3
3 7 8 43 × 4.
[Link]
M-48
The elementary transformation R2 → R2 + 2 R3 transforms A into a matrix B,where
1 4 2 9
B = 8 19 17 11
3 7 8 4 3 × 4.
Now if we apply the elementary transformation R2 → R2 − 2 R3 to the matrix B, we
see that the matrix B transforms to the matrix A.
Again suppose we apply the elementary transformation R2 → 3 R2 to the matrix A.
Then A transforms into a matrix C, where
1 4 2 9
C = 6 15 3 9 .
3 7 8 4
1
Now if we apply the elementary transformation R2 → R2 to the matrix C , we see that
3
the matrix C transforms back to the matrix A.
2.6 Elementary Matrices
Definition: A matrix obtained from a unit matrix by a single elementary transformation is
called an elementary matrix (or E-matrix). For example,
0 0 1 1 0 0 1 2 0
0 1 0 , 0 4 0 , 0 1 0
1 0 0 0 0 1 0 0 1
are the elementary matrices obtained from I3 by subjecting it to the elementary operations
C1 ↔ C3 , R2 → 4 R2 , R1 → R1 + 2 R2 respectively.
It may be worthwhile to note that an E-matrix can be obtained from I by subjecting it
to a row transformation or a column transformation. We shall use the following
symbols to denote elementary matrices of different types :
(i) Eij will denote the E-matrix obtained by interchanging the ith and j th rows of a unit
matrix. The students can easily see that the matrices obtained by interchanging
the ith and j th rows or the ith and j th columns of a unit matrix are the same.
Therefore Eij will also denote the elementary matrix obtained by interchanging
the ith and j th columns of a unit matrix.
[Link]
M-49
(ii) Ei (k ) will denote the E-matrix obtained by multiplying the ith row of a unit matrix
by a non-zero number k. It can be easily seen that the matrices obtained by
multiplying the ith row or the ith column of a unit matrix by k are the same.
Therefore Ei (k ) will also denote the elementary matrix obtained by multiplying
the ith column of a unit matrix by a non-zero number k.
(iii) Eij (m) will denote the elementary matrix, obtained by adding to the elements of
the ith row of a unit matrix, the products by any number m of the corresponding
elements of the j th row. It may be easily seen that the E-matrix, Eij (m) can also be
obtained by adding to the elements of the j th column of a unit matrix, the
products by m of the corresponding elements of the ith column.
It can be easily seen that
| Eij| = − 1, | Ei (k )| = k ≠ 0 , | Eij (m)| = 1.
Thus all the elementary matrices are non-singular.
Therefore each elementary matrix possesses inverse.
Now it is very interesting to note that the elementary transformations of a matrix can
also be obtained by algebraic operations on the same by the corresponding elementary
matrices. In this connection we have the following theorem.
2.7 Theorem
Every elementary row (column) transformation of a matrix can be obtained by pre-multiplication
(post-multiplication) with the corresponding elementary matrix.
We shall first prove that every elementary row transformation of a product AB of two
matrices A and B can be obtained by subjecting the pre-factor A to the same
elementary row transformation. Similarly every elementary column transformation
of a product AB of two matrices A and B can be obtained by subjecting the post-factor B
to the same elementary column transformation.
Let A = [aij] and B = [b jk ] be two m × n and n × p matrices respectively so that the
product AB is defined.
Let R1, R2 , R3 ,…, R m denote the row vectors of the matrix A and C1, C2 , …, Cp denote
the column vectors of the matrix B.
[Link]
M-50
R1
R2
We can then write, A = R3 , B = [C1 C2 C3 … Cp ].
…
R m
R1C1 R1C2 R1C3 … R1Cp
R2C1 R2C2 R2C3 … R2Cp
∴ AB = … … … … …
… … … … …
R mC1 R mC2 R mC3 … R mCp
Now if σ denotes any elementary row transformation, it is quite obvious from the
above representation that (σ A) B = σ (AB). For example, if σ denotes the elementary
row transformation R1 ↔ R2 , it is quite obvious that (σA) B = σ (AB).
Similarly it is quite obvious that if the columns C1, C2 ,…, Cp of B be subjected to any
elementary column transformation, the columns ofAB are also subjected to the same
elementary column transformation. Hence the result.
Now to prove our main theorem, if A be an m × n matrix, we can write
A = Im A.
If σ denotes any elementary row transformation, we have
σ A = σ (I m A) = (σ Im ) A = EA,
where E is the E-matrix corresponding to the same row transformation σ.
Similarly, we can write A = A In.
If σ denotes any elementary column transformation, we have
σ (A) = σ (AIn) = Aσ (In) = AE1,
where E1 is the E-matrix corresponding to the same column transformation σ.
1 4 2
Example: Let A = 2 7 1 .
3 8 4
The E-transformation R1 → R1 + 2 R3 transforms A into B, where
[Link]
M-51
7 20 10
B = 2 7 1 .
3 8 4
Also if we apply the row transformation R1 → R1 + 2 R3 to the unit matrix
1 0 0
I3 = 0 1 0 ,
0 0 1
1 0 2
then the E-matrix E thus obtained is E = 0 1 0 .
0 0 1
1 0 2 1 4 2
Now EA = 0 1 0 2 7 1
0 0 1 3 8 4
1⋅ 1 + 0 ⋅ 2 + 2 ⋅ 3 1⋅ 4 + 0 ⋅ 7 + 2 ⋅ 8 1 ⋅ 2 + 0 ⋅ 1 + 2 ⋅ 4
= 0 ⋅ 1 + 1 ⋅ 2 + 0 ⋅ 3 0 ⋅ 4 + 1⋅ 7 + 0 ⋅ 8 0 ⋅ 2 + 1 ⋅ 1 + 0 ⋅ 4
0 ⋅ 1 + 0 ⋅ 2 + 1 ⋅ 3 0 ⋅ 4 + 0 ⋅ 7 + 1⋅ 8 0 ⋅ 2 + 0 ⋅ 1 + 1 ⋅ 4
7 20 10
= 2 7 1 = B.
3 8 4
Similarly we can see that a column transformation of A can be effected by
post-multiplying A with the corresponding elementary matrix.
Non-singularity and Inverses of the Elementary Matrices
(i) The elementary matrix corresponding to the E-operation R1 ↔ Rj is its own inverse.
Let Eij denote the elementary matrix obtained by interchanging the ith and j th rows of a
unit matrix.
The interchange of the ith and j th rows of Eij will transform Eij to the unit matrix. But
every elementary row transformation of a matrix can be brought about by
pre-multiplication with the corresponding elementary matrix. Therefore the row
transformation which changes Eij to I can be effected on pre-multiplication by Eij.
[Link]
M-52
Thus Eij Eij = I or (Eij)−1 = Eij.
Hence Eij is its own inverse.
Similarly, we can show that the elementary matrix corresponding to the E-operation Ci ↔ C j is
its own inverse.
(ii) The inverse of the E-matrix corresponding to the E-operation Ri → kRi, (k ≠ 0 ), is the
E-matrix corresponding to the E-operation Ri → k −1 Ri.
Let Ei (k ) denote the elementary matrix obtained by multiplying the elements of the ith
row of a unit matrix I by a non-zero number k.
The operation of the multiplication of the ith row of Ei (k ), by k −1 will transform E i (k )
to the unit matrix I. This row transformation of E i (k ) can be effected on
pre-multiplication by the corresponding elementary matrix E i (k −1).
Thus Ei (k −1) Ei (k ) = I or { Ei (k )} −1 = Ei (k −1).
Similarly, we can show that the inverse of the E-matrix corresponding to the E-operation
Ci → kCi, k ≠ 0 , is the E-matrix corresponding to the E-operation Ci → k −1 Ci.
(iii) The inverse of the E-matrix corresponding to the E-operation Ri → Ri + kRj is the E-matrix
corresponding to the E-operation Ri → Ri − kRj.
Let Eij (k ) denote the elementary matrix obtained by adding to the elements of the ith
row of a unit matrix I, the products by any number k of the corresponding elements of
the j th row of I.
If we add to elements of the ith row of Eij (k ), the products by −k of the corresponding
elements of its j th row, then this row operation will transform Eij(k ) to the unit matrix I.
Now this row transformation of Eij(k ) can be effected on pre-multiplication by the
corresponding elementary matrix Eij (− k ).
Therefore Eij (− k ) Eij (k ) = I or { Eij (k )} −1 = Eij (− k ).
Similarly, we can show that the inverse of the E-matrix corresponding to the E-operation
Ci → Ci + kC j is the E-matrix corresponding to the E-operation Ci → Ci − kC j.
From the above theorem, we thus conclude that the inverse of an elementary matrix
is also an elementary matrix of the same type.
[Link]
M-53
2.8 Invariance of Rank Under Elementary Transformations
Theorem: Elementary transformations do not change the rank of a matrix.
(Lucknow 2005, 06)
Proof: Let A = [aij] be an m × n matrix of rank r. We shall prove the theorem in three
stages.
Case I. Interchange of a pair of rows does not change the rank.
Let B be the matrix obtained from the matrix A by the E-transformation Rp ↔ Rq . Let r
be the rank of A and s be the rank of B. Then to prove that r = s.
We have rank A = r ⇒ there exists at least one r-rowed square submatrix, say R, of A
such that | R| ≠ 0 .
Let S be the r-rowed square submatrix of B which has the same rows as are in R though
they may be in different relative positions.
Then either |S| = | R| or |S| = − | R|.
∴ | R | ≠ 0 ⇒ |S| ≠ 0 .
∴ rank B ≥ r ⇒ S ≥ r.
Again, as A can also be obtained from B by an interchange of rows, we have r ≥ s.
Hence r = s.
Case II. Multiplication of the elements of a row by a non-zero number does not change the rank.
Let B be the matrix obtained from the matrix A by the E-transformation
Rp → k Rp , (k ≠ 0 ), and let s be the rank of the matrix B .
Now if| B0| be any (r + 1) -rowed minor of B there exists a uniquely determined minor
| A0| of A such that
| B0| = | A0| ( this happens if the pth row of B is one of those rows which
are struck off to obtain B0 from B),
or | B0| = k | A0|
(this happens when pth row is retained while obtaining B0 from B).
[Link]
M-54
Since the matrix A is of rank r, therefore every (r + 1)-rowed minor of A vanishes i. e.,
| A0| = 0 . Hence | B0 | = 0 . Thus we see that every (r + 1)-rowed minor of B also vanishes.
Therefore, s (the rank of B) cannot exceed r ( the rank of A).
∴ s ≤ r.
Also, since A can be obtained from B by E-transformation of the same type i. e.,
Rp → (1 / k ) Rp , therefore, by interchanging the roles of A and B we find that
r ≤ s.
Thus r = s.
Case III. Addition to the elements of a row the products by any number k of the corresponding
elements of any other row does not change the rank .
Let B be the matrix obtained from the matrix A by the E-transformation
Rp → Rp + kRq and let s be the rank of the matrix B.
Let B0 be any (r + 1)-rowed square sub-matrix of B and A0 be the correspondingly
placed sub-matrix of A.
The transformation Rp → Rp + kRq has changed only the pth row of the matrix A. Also
the value of a determinant does not change if we add to the elements of any row the
corresponding elements of any other row multiplied by some number k.
Therefore, if no row of the sub-matrix A0 is part of the pth row of A, or if two rows of A0
are parts of the pth and q th rows of A, then | B0| = | A0|.
Since the rank of A is r, therefore | A0| = 0 , and consequently | B0| = 0 .
Again, if a row of A0 is a part of the pth row of A, but no row is a part of the q th row, then
| B0| = | A0| + k |C0| , where C0 is an (r + 1) -rowed square matrix which can be obtained
from A0 by replacing the elements of A0 in the row which corresponds to the pth row of
A by the corresponding elements in the q th row of A. Obviously all the r + 1rows of the
matrix C0 are exactly the same as the rows of some (r + 1)-rowed square sub-matrix of A,
though arranged in some different order. Therefore|C0| is ± 1times some (r + 1)-rowed
minor of A. Since the rank of A is r, therefore, every (r + 1)-rowed minor of A is zero, so
that | A0| = 0 ,|C0| = 0 , and consequently | B0| = 0 .
Thus we see that every (r + 1)-rowed minor of B also vanishes. Hence, s (the rank of B)
cannot exceed r ( the rank of A).
∴ s ≤ r.
[Link]
M-55
Also, since A can be obtained from B by an E-transformation of the same type i. e.,
Rp → Rp − kRq , therefore, interchanging the roles of A and B, we have r ≤ s.
Thus r = s.
We have thus shown that rank of a matrix remains unaltered by any E-row
transformation. Therefore we can also say that the rank of a matrix remains unaltered
by a series of elementary row transformations.
Similarly we can show that the rank of a matrix remains unaltered by a series of
elementary column transformations.
Finally, we conclude that the rank of a matrix remains unaltered by a finite chain of elementary
operations.
Corollary. We have already proved that every elementary row (column)
transformation of a matrix can be effected by pre-multiplication (post-multiplication)
with the corresponding elementary matrix. Combining this theorem with the theorem
just established, we conclude the following important results :
The pre-multiplication or post-multiplication by any elementary matrix, and as such by any
series of elementary matrices, does not alter the rank of a matrix.
2.9 Reduction to Normal Form
Ir O
Theorem: Every m × n matrix of rank r can be reduced to the form by a finite chain of
O O
E-operations where Ir , is the r-rowed unit matrix.
Proof: Let A = [aij] be an m × n matrix of rank r. If A is a zero matrix, then r is equal to
zero and we have nothing to prove. So let us take A as a non-zero matrix.
Since A is a non-zero matix, therefore A has at least one element different from zero,
say apq = k ≠ 0.
By interchanging the pth row with the first row and the q th column with the first
column respectively, we obtain a matrix B whose leading element is equal to k which is
not equal to zero.
Multiplying the elements of the first row of the matrix B by 1/k, we obtain a matrix C
whose leading element is equal to unity.
[Link]
M-56
1 c12 c13 … c1n
c21 c22 c23 … c2 n
Let C= .
… … … … …
c m1 c m2 c m3 … c mn
Subtracting suitable multiples of the first column of C from the remaining columns,
and suitable multiples of the first row from the remaining rows, we obtain a matrix D in
which all elements of the first row and first column except the leading element are
equal to zero.
1 0 0 0 … 0
0
0 A1
Let D= ,
0
…
0
where A1 is an (m − 1) × (n − 1) matrix.
If now, A1 be a non-zero matrix, we can deal with it as we did with A. If the elementary
operations applied to A1 for this purpose be applied to D, they will not affect the first
row and the first column of D . Continuing this process, we shall finally obtain a matrix
M, such that
Ik O
M= .
O O
The matrix M has the rank k. Since the matrix M has been obtained from the matrix A
by elementary transformations and elementary transformations do not alter the rank,
therefore we must have k = r.
Ir O
Hence every m × n matrix of rank r can be reduced to the form by a finite chain
O O
of elementary transformations.
Note. The above form is usually called the first canonical form or normal form of a
matrix.
Corollary 1: The rank of an m × n matrix A is r if and only if (iff) it can be reduced to the form
Ir O
by a finite chain of E-operations.
O O
[Link]
M-57
The condition is necessary. The proof has been given in the above theorem.
The condition is also sufficient. The matrix A has been transformed into the form
Ir O
by elementary transformations which do not alter the rank of the matrix.
O O
Ir O
Since the rank of the matrix is r, therefore the rank of the matrix A must also
O O
be r.
Corollary 2: If A be an m × n matrix of rank r, there exist non-singular matrices P and Q such
that
Ir O
PAQ = .
O O
Ir O
Proof: If A be an m × n matrix of rank r, it can be transformed into the form by
O O
elementary operations. Since E-row (column) operations are equivalent to
pre-(post)-multiplication by the corresponding elementary matrices, we have the
following result :
If A be an m × n matrix of rank r, there exist E-matrices P1, P2 ,… Ps , Q1, Q2 , …, Qt such
that
Ir O
Ps Ps − 1 … P1A Q1Q2 …Qt = .
O O
Now each elementary matrix is non-singular and the product of non-singular matrices
is also non-singular. Therefore if P = Ps Ps − 1 … P2 P1 and Q = Q1 Q2 , …Qt , then P and Q
are non-singular matrices. Hence
Ir O
PAQ = .
O O
2.10 Equivalence of Matrices
Definition. If B be an m × n matrix obtained from an m × n matrix A by finite number of
elementary transformations of A then A is called equivalent to B. Symbolically, we write
A ~ B, which is read as ‘A is equivalent to B.’
The following three properties of the relation ‘~’ in the set of all m × n matrices are
quite obvious :
[Link]
M-58
(i) Reflexivity: If A is any m × n matrix, then A ~ A. Obviously A can be obtained from
A by the elementary transformation Ri → kRi , where k = 1.
(ii) Symmetry: If A ~ B, then B ~ A. If B can be obtained from A by a finite number of
elementary transformations of A, then A can also be obtained from B by a finite
number of elementary transformations of B.
(iii) Transitivity: If A ~ B, B ~ C, then A ~ C.
If B can be obtained from A by a series of elementary transformations of A and C can be
obtained from B by a series of elementary transformations of B, then C can also be
obtained from A by a series of elementary transformations of A.
Therefore the relation ‘~’ in the set of all m × n matrices is an equivalence relation.
Example 7: If A and B be two equivalent matrices, then show that rank A = rank B.
Solution : If A ~ B, then B can be obtained from A by a finite number of elementary
transformations of A. Now the elementary transformations do not change the rank of
a matrix.
∴ If A ~ B, then rank A = rank B.
Example 8: Show that if two matrices A and B have the same size and the same rank, they are
equivalent.
Solution: Let A and B be two m × n matrices of the same rank r. Then by 2.9, we have
Ir O Ir O
A~ and also B ~ .
O O O O
By the symmetry of the equivalence relation,
Ir O Ir O
B~ implies ~ B.
O O O O
Now by the transitivity of the equivalence relation
Ir O Ir O
A~ and ~ B implies A ~ B.
O O O O
Example 9: (i) Use elementary transformations to reduce the following matrix A to triangular
form and hence find rank A :
[Link]
M-59
5 3 14 4
A = 0 1 2 1.
1 −1 2 0
(Bundelkhand 2010)
8 1 3 6
(ii) Find the rank of the matrix 0 3 2 2.
−8 −1 −3 4
Solution : (i) We have the matrix
1 −1 2 0
A ~ 0 1 2 1 by R1 ↔ R3
5 3 14 4
1 −1 2 0
~ 0 1 2 1 by R3 → R3 − 5 R1
0 8 4 4
1 −1 2 0
~ 0 1 2 1 by R3 → R3 − 8 R2 .
0 0 −12 −4
The last equivalent matrix is in Echelon form (or in triangular form). The number of
non-zero rows in this matrix is 3. Therefore its rank is 3. Hence rank A = 3.
(ii) Let us denote the given matrix by A. To find the rank of A, we shall reduce it to
1
Echelon form. Performing the column operation C1 → C1, we get
8
1 1 3 6 1 1 3 6
A~ 0 3 2 2 ~ 0 3 2 2 by R3 → R3 + R1.
−1 −1 −3 4 0 0 0 10
The last equivalent matrix is in Echelon form. The number of non-zero rows in this
matrix is 3. Therefore its rank is 3. Hence rank A = 3.
1 2 1
Example 10: Is the matrix −1 0 2 equivalent to I3 ?
2 1 −3
(Meerut 2008)
[Link]
M-60
1 2 1
Solution: Let A = −1 0 2.
2 1 −3
1 2 1 1 2 1
We have | A | = −1 0 2 = 0 2 3 R2 + R1, R3 − 2 R1
2 1 −3 0 −3 −5
= − 10 + 9 = − 1 i. e., ≠ 0.
Thus the matrix A is non-singular. Hence it is of rank 3. The rank of I3 is also 3. Since A
and I3 are matrices of the same size and the same rank, therefore A ~ I3 .
1 −1 2 −3
4 1 0 2 Ir O
Example 11: Reduce the matrix, A = to the nor mal form and
0 3 0 4 O O
0 1 0 2
hence determine its rank. (Meerut 2001, 09B, 10)
Solution: We have the matrix A
1 0 0 0
4 5 −8 14
~ by C2 → C2 + C1, C3 → C3 − 2 C1, C4 → C4 + 3 C1
0 3 0 4
0 1 0 2
1 0 0 0
0 5 −8 14
~ by R2 → R2 − 4 R1
0 3 0 4
0 1 0 2
1 0 0 0
0 1 0 2
~ by R2 ↔ R4
0 3 0 4
0 5 −8 14
1 0 0 0
0 1 0 0
~ by C4 → C4 − 2 C2
0 3 0 −2
0 5 −8 4
[Link]
M-61
1 0 0 0
0 1 0 0
~ by R3 → R3 − 3 R2 , R4 → R4 − 5 R2
0 0 0 −2
0 0 −8 4
1 0 0 0
0 1 0 0
~ by C3 ↔ C4
0 0 −2 0
0 0 4 −8
1 0 0 0
0 1 0 0 1 1
~ by C3 → − C3 , C4 → − C4
0 0 1 0 2 8
0 0 −2 1
1 0 0 0
0 1 0 0
~ by R4 → R4 + 2 R3
0 0 1 0
0 0 0 1
~ I4 . Hence the matrix A is of rank 4.
Example 12: Determine the rank of the following matrices :
2 −1 3 4
0 3 4 1
(i) 2 3 7 5
2 5 11 6 (Kanpur 2010)
−2 −1 −3 −1
1 2 3 −1
(ii) 1 0 1 1
0 1 1 −1 (Meerut 2003, 09 B, 10B)
Solution : (i) Let us denote the given matrix by A. Performing the elementary
operations R3 → R3 − R1, R4 → R4 − R1, we see that
[Link]
M-62
2 −1 3 4 2 −1 3 4
0 3 4 1 0 3 4 1
A~ ~ by R4 → R4 − 2 R2
0 4 4 1 0 4 4 1
0 6 8 2 0 0 0 0
2 −1 3 4
0 12 16 4
~ by R2 → 4 R2 , R3 → 3 R3
0 12 12 3
0 0 0 0
2 −1 3 4
0 12 16 4
~ by R3 → R3 − R2 .
0 0 −4 −1
0 0 0 0
The last equivalent matrix is in Echelon form. The number of non-zero rows in this
matrix is 3. Therefore its rank is 3. Hence rank A = 3.
(ii) Let us denote the given matrix by A. Performing the elementary operation
R1 ↔ R2 , we see that
1 2 3 −1
−2 −1 −3 −1
A~
1 0 1 1
0 1 1 −1
1 2 3 −1
0 3 3 −3
~ by R2 → R2 + 2 R1, R3 → R3 − R1
0 −2 −2 2
0 1 1 −1
1 2 3 −1
0 1 1 −1 1 1
~ by R2 → R2 , R3 → − R3
0 1 1 −1 3 2
0 1 1 −1
1 2 3 −1
0 1 1 −1
~ by R3 → R3 − R2 , R4 → R4 − R2 .
0 0 0 0
0 0 0 0
[Link]
M-63
The last equivalent matrix is in Echelon form. The number of non-zero rows in this
matrix is 2. Therefore rank A = 2.
2 −2 0 6
4 2 0 2
Example 13: Find the rank of the matrix A = by reducing it to normal form.
1 −1 0 3
1 −2 1 2
(Meerut 2008; Avadh 08; Rohilkhand 09)
1 1
Solution : Performing the operation R1 → R1, R2 → R2 , we see that
2 2
1 −1 0 3 1 −1 0 3
2 1 0 1 0 3 0 −5
A~ ~
1 −1 0 3 0 0 0 0
1 −2 1 2 0 −1 1 −1
by R2 → R2 − 2 R1, R3 → R3 − R1, R4 → R4 − R1
1 0 0 0
0 3 0 −5
~ by C2 → C2 + C1, C4 → C4 − 3 C1
0 0 0 0
0 −1 1 −1
1 0 0 0
0 −1 1 −1
~ by R2 ↔ R4
0 0 0 0
0 3 0 −5
1 0 0 0
0 1 −1 1
~ by R2 → (−1) R2
0 0 0 0
0 3 0 −5
1 0 0 0
0 1 −1 1
~ by R4 → R4 − 3 R2
0 0 0 0
0 0 3 −8
[Link]
M-64
1 0 0 0
0 1 0 0
~ by C3 → C3 + C2 , C4 → C4 − C2
0 0 0 0
0 0 3 −8
1 0 0 0
0 1 0 0
~ by R4 ↔ R3
0 0 3 −8
0 0 0 0
1 0 0 0
0 1 0 0 1 1
~ by C3 → C3 , C4 → − C4
0 0 1 1 3 8
0 0 0 0
1 0 0 0
0 1 0 0
~ by C4 → C4 − C3 ,
0 0 1 0
0 0 0 0
I3 O
which is the normal form . Hence rank A = 3.
O O
Example 14: Find two non-singular matrices P and Q such that PAQ is in the normal form where
1 1 1
A = 1 −1 −1 .
3 1 1
Also find the rank of the matrix A.
Solution: We write A = I3 AI3 i. e.,
1 1 1 1 0 0 1 0 0
1 −1 −1 = 0 1 0 A 0 1 0 ⋅
3 1 1 0 0 1 0 0 1
Now we go on applying E-operations on the matrix A (the left hand member of the
above equation) until it is reduced to the normal form. Every E-row operation will also
be applied to the pre-factor I3 (or its transform) of the product on the right hand
[Link]
M-65
member of the above equation and every E-column operation to the post-factor I3 (or
its transform).
Performing R2 → R2 − R1, R3 → R3 − 3 R1, we get
1 1 1 1 0 0 1 0 0
0 −2 −2 = −1 1 0 A 0 1 0 .
0 −2 −2 −3 0 1 0 0 1
Performing C2 → C2 − C1, C3 → C3 − C1, we get
1 0 0 1 0 0 1 −1 −1
0 −2 −2 = −1 1 0 A 0 1 0 .
0 −2 −2 −3 0 1 0 0 1
1
Performing R2 → − R2 , we get
2
1 0 0 1 0 0 1 −1 −1
1 1
0 1 1 = − 0 A 0 1 0 .
0 2 2
−2 −2 −3 0 1 0 0 1
Performing R3 → R3 + 2 R2 , we get
1 0 0 1 0 0 1 −1 −1
1 1
0 1 1 = − 0 A 0 1 0 .
0 2 2
0 0 −2 −1 1 0 0 1
Performing C3 → C3 − C2 , we get
1 0 0 1 0 0 1 −1 0
1 1
0 1 0 = − 0 A 0 1 −1 .
0 2 2
0 0 −2 −1 1 0 0 1
I2 O
∴ PAQ = ,
O O
1 0 0 1 −1 0
1 1
where P= − 0 , Q = 0 1 −1 .
−2 2
1 0 1
2 −1 0
I2 O
Since A~ , therefore rank A = 2.
O O
[Link]
M-66
Comprehensive Exercise 2
Determine the rank of the following matrices :
1 2 3 0 6 1 3 8
2 4 3 2 4 2 6 −1
1. 3 2.
2 1 3 10 3 9 7
6 8 7 5 16 4 12 15
(Bundelkhand 2008;
Rohilkhand 08; Avadh 05; Lucknow 11)
2 3 −1 −1
1 −1 −2 −4
3. 3 1 3 −2
6 3 0 −7 (Avadh 2006, 09; Purvanchal 06)
1 a b 0
1 3 4 3
0 c d 1
4. 1 (Lucknow 2007) 5. 3 9 12 9
a b 0 1 3
4 1
0 c d 1 (Rohilkhand 2007)
1 2 1 2
1 2 −1 4
1 3 2 2
6. 2 4 3 5 7.
−1 2 4 3 4
−2 6 −7
3 7 4 6 (Lucknow 08)
3 −2 0 −1
1 2 −4 5
0 2 2 1
8. 1 9. 2 −1 3 6
−2 −3 2 8
1 9 7
0 1 2 1
8 0 0 1
1 −1 3 6
1 0 8 1
10. 1 3 −3 −4 11.
5 0 0 1 8
3 3 11
0 1 1 8 (Kumaun 2008)
0 1 −3 −1
1 0 1 1
12. 3 1 0 2
1 1 −2 0 (Avadh 2010)
[Link]
M-67
1 1 1
1 −3 4 7
13. (Kumaun 2008) 14. 2 2 2
9 1 2 0 3
3 3
2 1 3 4 5 6
15. 4 7 13 16. 5 6 7
4 −3 −1 7 8 9
1 0 2 1
1 1 1
0 1 −2 1
17. 1 18. a b c
−1 4 0 a3
b3 c 3
−2 2 8 0 (Meerut 2011)
19. With the help of elementary transformations find the rank of the following
matrix:
1 1 2 3
1 3 0 3
1 −2 −3 −3
1 1 2 3 (Rohilkhand 2010)
20. Reduce the following matrix to its Echelon form and find its rank :
1 3 4 5
3 9 12 9
−1 −3 −4 −3
(Meerut 2004B; Rohilkhand 06, 10)
1 2 3
21. Find the rank of the matrix A = 2 3 4 after reducing it to normal form.
3 5 7
0 1 2 −1
1 0 1 1
22. Reduce the matrix to normal form and find its rank.
3 1 0 2
1 1 −2 0
(Meerut 2009; Bundelkhand 11; Avadh 11)
9 7 3 6
23. (i) Reduce the matrix 5 −1 4 1 to normal form and find its rank.
6 8 2 4
(Agra 2007)
(ii) Reduce the following matrix A into normal form and hence find its rank:
0 1 2 −2
A = 4 0 2 6 ⋅
2 1 3 1
(Kanpur 2011)
[Link]
M-68
24. Are the following pairs of matrices equivalent?
2 −1 3 4 1 0 −5 6
0 3 4 1 3 −2 1 2
(i) ,
2 3 7 5 5 −2 −9 14
2 5 11 5 4 −2 −4 8
4 0 2 3 9 0 2
(ii) 3 1 0 , 7 −2 0 1
5 0 0 8 1 1 5
25. Find the ranks of A, B, A + B, AB and BA where
1 1 −1 −1 −2 −1
A = 2 −3 4, B = 6 12 6 .
3 −2 3 5 10 5
26. Show that if A and B are equivalent matrices, then there exist non-singular
matrices P and Q such that B = PAQ.
27. Show that the rank of a matrix is not altered if a column of it is multiplied by a
non-zero scalar.
28. (i) What is the rank of a non-singular matrix of order n?
(ii) What is the rank of an elementary matrix?
A nswers 2
1. 3 2. 2 3. 3 4. 2 5. 2 6. 2 7. 3 8. 4 9. 3
10. 3 11. 4 12. 2 13. 2 14. 1 15. 2 16. 2 17. 3
18. rank (A ) = 3 if a ≠ b ≠ c and a + b + c ≠ 0; rank (A ) = 2 if a ≠ b ≠ c and
a + b + c = 0; Also rank (A ) = 2 if a = b ≠ c ; and rank (A ) = 1 if a = b = c
19. 3 20. 2 21. 2 22. 3 23. (i) 3 (ii) 2
24. (i) No, since the rank of the first matrix is 4 and that of the second matrix is 2.
(ii) No, since the marices are not of the same type
25. Rank A = 2; Rank B = 1; Rank A + B = 2;
Rank AB = 0; Rank BA = 1
28. (i) n. (ii) Equal to the order of the matrix
[Link]
M-69
2.11 Row and Column Equivalence of Matrices
Definition: A matrix A is said to be row equivalent to B if B is obtainable from A by a finite
R
number of E-row transformations of A. Symbolically, we then write A ~ B. Similarly a matrix A
is said to be column equivalent to B if B is obtainable from A by a finite number of E-column
C
transformations of A. Symbolically, we then write A ~ B.
2.12 Employment of Only Row Transformations
Theorem: If A be an m × n matrix of rank r, then there exists a non-singular matrix P such
G
that PA = , where G is an r × n matrix of rank r and O is (m − r) × n.
O
Proof. Since A is an m × n matrix of rank r, therefore there exist non-singular
matrices P and Q, such that
Ir O
PAQ = . …(1)
O O
Now every non-singular matrix can be expressed as the product of elementary
matrices. So let
Q = Q1 Q2 ...... Qt where Q1, Q2 , ........, Qt are elementary matrices.
Thus the relation (1) can be written as
Ir O
PAQ1 Q2 ..... Qt = ⋅ …(2)
O O
Now everyE-column transformation of a matrix, is equivalent to post-multiplication
with the corresponding elementary matrix. Since no column transformation can affect
the last (m − r) rows of the right hand side of (2), therefore post-multiplying the L. H.S.
of (2) by the elementary matrices Q−t 1, Q−t 1− 1, .... , Q2−1, Q1−1 successively and affecting
the corresponding column transformations in the right hand side of (2), we get a
relation of the form
G
PA = ⋅
O
[Link]
M-70
Since elementary transformations do not alter the rank, therefore the rank of the
G
matrix PA is the same as that of the matrix A which is r. Thus the rank of the matrix
O
is r and therefore the rank of the matrix G is also r as the matrix G has r rows and last
G
m − r rows of the matrix consist of zeros only.
O
2.13 Employment of Only Column Transformations
Theorem: If A be an m × n matrix of rank r, then there exists a non-singular matrix Q such
that AQ = [H O] , where H is an m × n matrix of rank r and O is m × (n − r).
Proof. Since A is an m × n matrix of rank r, therefore there exist non-singular
matrices P and Q such that
Ir O
PAQ = ⋅ …(1)
O O
Now every non-singular matrix can be expressed as the product of elementary
matrices. So let
P = P1 P2 ...... Ps where P1, P2 , ....., Ps are elementary matrices.
Thus the relation (1) can be written as
Ir O
P1 P2 .... Ps AQ = . …(2)
O O
Now every E-row transformation of a matrix is equivalent to pre-multiplication with
the corresponding elementary matrix. Again no row transformation can affect the last
Ir O
n − r columns of ⋅
O O
Therefore pre-multiplying the L.H.S. of (2) by the elementary matrices.
P1−1, P2−1, ..... Ps−1
successively and affecting the corresponding row transformations in the R.H.S. of (2),
we get a relation of the form AQ = [H O].
Now elementary transformations do not alter the rank. Therefore the rank of the
matrix AQ is the same as that of A which is r. Thus the rank of the matrix [H O] is r and
[Link]
M-71
therefore the rank of the matrix H is also r as the matrix H has r columns and the last
n − r columns of the matrix [H O] consists of zero only.
2.14 The Rank of a Product
Theorem: The rank of a product of two matrices cannot exceed the rank of either matrix.
(Lucknow 2009)
Let A and B be two m × n and n × p matrices respectively. Let r1, r2 be the ranks of A and
B respectively and let r be the rank of the product AB.
To prove r ≤ r1 and r ≤ r2 .
Since A is an m × n matrix of rank r1, therefore there exists a non-singular matrix P such
G
that PA = , where G is an r1 × n matrix of rank r1 and O is (m − r1) × n.
O
G
∴ PAB = B.
O
Since the rank of a matrix does not alter by multiplying it with a non-singular matrix,
therefore
Rank (PAB) = Rank (AB) = r.
G
∴ Rank of the matrix B = r.
O
G
Since the matrix G has only r1 non-zero rows, therefore the matrix B cannot have
O
more than r1 non-zero rows which arise by multiplying the r1 non-zero rows of G with
the columns of B.
G
∴ Rank of the matrix B is ≤ r1 i. e., t ≤ r1
O
i. e., Rank (AB) ≤ Rank of the prefactor A.
Again r = Rank (AB) = Rank (AB)′ = Rank (B ′ A ′ ) ≤ Rank of the prefactor B′
= Rank B [∵ Rank B = Rank B′]
= r2 .
∴ r ≤ r2 i. e., Rank (AB) ≤ Rank of the post-factor B.
[Link]
M-72
2.15 Theorem
Every non-singular matrix is row equivalent to a unit matrix.
Proof. We shall prove the theorem by induction on n, the order of the matrix. If the
matrix be of order 1 i. e., if A =| a11|, the theorem obviously holds.
Let us assume that the theorem holds for all non-singular matrices of order n − 1.
Let A = [aij] be an n × n non-singular matrix. The first column of the matrix A has at
least one element different from zero, for otherwise we shall have |A | = 0 and the
matrix A will not be non-singular.
Let ap1 = k ≠ 0 .
By interchanging the p th row with the first row (if necessary), we obtain a matrix B
whose leading element is equal to k which is not equal to zero.
Multiplying the elements of the first row of the matrix B by 1 / k, we obtain a matrix C
whose leading element is equal to unity.
1 c12 c13 .... c1n
c21 c22 c23 .... c2 n
Let C = c31 c32 c33 .... c3 n ⋅
.... .... .... .... ....
c n1 c n2 c n3 .... c nn
Subtracting suitable multiples of the first row of C from the remaining rows, we obtain
a matrix D in which all elements of the first column except the leading element are
equal to zero.
1 d12 d13 ... d1n
0
Let D = 0 A1 .
...
0
where A1 is an (n − 1) × (n − 1) matrix. The matrix A1 is non-singular, for otherwise
| A1| = 0 and so |D | is also equal to zero. Thus the matrix D will be non-singular, and
therefore A, which is row equivalent to D, will also not be non-singular.
[Link]
M-73
By the inductive hypothesis, A1 can be transformed to In −1 by E-row operations. If
these elementary row operations be applied to D, they will not affect the first row and
the first column of D and we shall obtain a matrix M such that
1 d12 d13 .... d1n
0 1 0 .... 0
M = 0 0 1 .... 0 ⋅
.... .... .... .... ....
0 0 0 .... 1
By adding suitable multiples of the second, third, …, n th rows to the first row of M, we
obtain the matrix In.
Thus the matrix A has been reduced to In by E-row operations only.
The proof is now complete by induction.
Corollary 1: If A be an n-rowed non-singular matrix, there exist E-matrices E1, E2 , ...., Et such
that
Et Et − 1.... E2 E1A = In.
If A be an n-rowed non-singular matrix, it can be reduced to In by E-row operations
only. Since every E-row operation is equivalent to pre-multiplication by the
corresponding E-matrix, therefore we can say that if A be an n -rowed non-singular
matrix, there exist E-matrices E1, E2 , ...., Et such that
Et Et − 1 ..... E2 E1A = In.
Corollary 2: Every non-singular matrix A is expressible as the product of elementary
matrices.
If A be an n-rowed non-singular matrix, there exist E-matrices E1, E2 , ...., Et such that
Et Et − 1 ..... E2 E1A = In. …(1)
Pre-multiplying both sides of the relation (1) by (Et Et − 1 ...... E2 E1)−1, we get
(Et Et − 1..... E2 E1)−1 (Et Et − 1...... E2 E1)A = (Et Et − 1.... .. E2 E1)−1 In
or InA = E1−1E2−1E3−1..... Et − 1−1Et−1In
−1
or A = E1−1E2−1 ..... Et − 1E−t 1.
[Link]
M-74
Since the inverse of an elementary matrix is also an elementary matrix of the same type
hence we get the result.
Corollary 3: The rank of a matrix does not alter by pre-multiplication or post-multiplication
with a non-singular matrix.
Every non-singular matrix can be expressed as the product of elementary matrices.
Also E-row (column) transformations are equivalent to pre-(post)-multiplication with
the corresponding elementary matrices and elementary transformations do not alter
the rank of a matrix. Hence we get the result.
2.16 Use of Elementary Transformations to Find the Inverse of
a Non-Singular Matrix
Let A be a non-singular matrix of order n. It can be easily shown that A can be reduced
to the unit matrix In by a finite number of E-row transformations only. Now each
E-row transformation of a matrix is equivalent to pre-multiplication by the
corresponding E-matrix. Therefore there exist elementary matrices, say, E1, E2 , ....., Et
such that (Et Et − t ...... E2 E1) A = In.
Post-multiplying both sides by A −1, we get
(Et Et − 1...... E2 E1)AA −1 = InA −1
or (Et Et − 1...... E2 E1)In = A −1 [∵ AA −1 = In, InA −1 = A −1]
or A −1 = (Et Et − 1...... E2 E1)In.
Hence we get the following result :
If a non-singular matrix A of order n is reduced to the unit matrix In by a sequence of E-row
transformations only, then the same sequence of E-row transformations applied to the unit matrix
In gives the inverse of A (i. e., A −1).
2.17 Working Rule for Finding the Inverse of a Non-Singular
Matrix by E-Row Transformations
Suppose A is a non-singular matrix of order n. Then we write A = InA.
Now we go on applying E-row transformations only to the matrix A and the pre-factor
In of the product InA till we reach the result In = BA.
Then obviously B is the inverse of A.
[Link]
M-75
1 2 1
Example 15: Find the inverse of the matrix A = 3 2 3 by using E-transformations.
1 1 2
(Avadh 2006, 08; Purvanchal 09; Rohilkhand 09; Lucknow 09)
1 2 1 1 0 0
Solution : We write A = I3 A, i. e., 3 2 3 = 0 1 0 A.
1 1 2 0 0 1
Now we go on applying E-row transformations to the matrix A (the left hand member
of the above equation) until it is reduced to the form I3 . Every E-row transformation
will also be applied to the prefactor I3 (or its transform) of the product on the right
hand side of the above equation.
Applying R2 → R2 − 3 R1, R3 → R3 − R1, we get
1 2 1 1 0 0
0 −4 0 = −3 1 0 A.
0 −1 1 −1 0 1
Now we should try to make 1 in the place of the second element of the second row of
1
the matrix on the left hand side. So applying R2 → − R2 , we get
4
1 2 1 1 0 0
3 1
0 1 0 = − 0 A.
0 4 4
−1 1 −1 0 1
Now we shall make zeros in the place of the second elements of the first and third rows
with the help of the second row. So applying R1 → R1 − 2 R2 , R3 → R3 + R2 , we get
1 1
− 0
1 0 1 2 2
3 1
0 1 0 = − 0 A.
0 4 4
0 1 1 1
− − 1
4 4
[Link]
M-76
Now the third element of the third row is already 1. So to make the third element of
the first row zero, we apply R1 → R1 − R3 , and we get
1 3
− −1
1 0 1 3 4
3 1
0 1 0 = − 0 A.
0 4 4
0 1 1 1
− − 1
4 4
1 3
− 4 −1
4
3 1
Thus I3 = BA, where B = − 0 .
4 4
− 1 −
1
1
4 4
1 3
− 4 −1
4
3 1
∴ A −1 =B= − 0 ⋅
4 4
− 1 −
1
1
4 4
0 1 2 2
1 1 2 3
Example 16: Find the inverse of the matrix A = by using E-transformations.
2 2 2 3
2 3 3 3
0 1 2 2 1 0 0 0
1 1 2 3 0 1 0 0
Solution : We write = A.
2 2 2 3 0 0 1 0
2 3 3 3 0 0 0 1
Performing R1 ↔ R2 , we get
1 1 2 3 0 1 0 0
0 1 2 2 1 0 0 0
2 = A.
2 2 3 0 0 1 0
2 3 3 3 0 0 0 1
[Link]
M-77
Performing R3 → R3 − 2 R1, R4 → R4 − 2 R1, we get
1 1 2 3 0 1 0 0
0 1 2 2 1 0 0 0
0 = A.
0 −2 −3 0 −2 1 0
0 1 −1 −3 0 −2 0 1
Performing R4 → R4 − R2 , R1 → R1 − R2 , we get
1 0 0 1 − 1 1 0 0
0 1 2 2 1 0 0 0
0 = A.
0 −2 −3 0 −2 1 0
0 1 −3 −5 −1 −2 0 1
Performing R4 → R4 − R2 , R1 → R1 − R2 , we get
1 0 0 1 − 1 1 0 0
0 1 2 2 1 0 0 0
0 = A.
0 −2 −3 0 −2 1 0
0 0 −3 −5 −1 −2 0 1
1
Performing R3 → − R3 , we get
2
1 0 0 1 − 1 1 0 0
0 1 2 2 1 0 0 0
0 3 = 1 A.
0 1 0 1 − 0
2 2
0 0 −3 −5 −1 −2 0 1
Performing R2 → R2 − 2 R3 , R4 → R4 + 3 R3 , we get
1 0 0 1 −1 1 0 0
0 1 0 −1 1 −2 1 0
0 3 = 1
0 1 0 1 − 0 A.
2 2
1 3
0 0 0 − −1
2
1 −
2
1
[Link]
M-78
Performing R4 → − 2 R4 , we get
1 0 0 1 −1 1 0 0
0 1 0 −1 1 −2 1 0
0 3 = 1 A.
0 1 0 1 − 0
2 2
0 0 0 1 2 −2 3 −2
3
Performing R3 → R3 − R4 , R2 → R2 + R4 , R1 → R1 − R4 we get
2
1 0 0 0 −3 3 −3 2
0 1 0 0 3 −4 4 −2
0 = A.
0 1 0 −3 4 −5 3
0 0 0 1 2 −2 3 −2
−3 3 −3 2
3 −4 4 −2
∴ A −1 = ⋅
−3 4 −5 3
2 −2 3 −2
Comprehensive Exercise 3
1 2 1
1. Reduce the matrix A = −1 0 2 to I3 by E-row transformations only.
2 1 3
2. Compute the inverse of the following matrices by using elementary
row-transformations :
1 2 3 i −1 2 i
(i) 2 4 5 (ii) 2 0 2
3 5 6 −1 0 1
−1 −3 3 −1
1 3 3
1 1 −1 0
(iii) (iv) 1 4 3
2 −5 2 −3 1
3 4
−1 1 0 1
(Avadh 2007)
[Link]
M-79
0 1 2
(v) 1 2 3
3 1 1
(Rohilkhand 2011; Bundelkhand 08)
A nswers 3
1 1
1 −3 2 0 4
−
2
1 3i i
2. (i) −3 3 −1 (ii) −1
157 4 2
−2 −1 0 0 1 1
4 2
0 2 1 3
7 −3 −3
1 1 −1 −2
(iii) (iv) −1 1 0
1 2 0 1 −1
0 1
−1 1 2 6
1 1 1
2 −
2 2
(v) −4 3 −1
5 3 1
−
2 2 2
Objective Type Questions
Multiple Choice Questions
Indicate the correct answer for each question by writing the corresponding letter from (a),
(b) , (c) and (d).
1. If A is a matrix such that there exists a square submatrix of order r which is
non-singular and every square submatrix of order r + 1or more is singular, then
(a) rank A = r + 1
(b) rank A = r
(c) rank A > r
(d) rank A ≥ r + 1
[Link]
M-80
2. Let A = [aij]m × n be a matrix such that aij = 1 for all i, j. Then
(a) rank A > 1
(b) rankA = 1
(c) rank A = m
(d) rank A = n
1 2 3 0
2 4 3 2
3. The rank of the matrix is
3 2 1 3
6 8 7 5
(a) 1 (b) 2
(c) 3 (d) 4
4. If A and B are two matrices such that rank of A = m and rank of B = n, then
(a) rank (A B) = m n
(b) rank (A B) ≥ rank A
(c) rank (A B) ≥ rank B
(d) rank (A B) ≤ min (rank A, rank B) (Agra 2007)
5. If A is an invertible matrix and B is a matrix, then
(a) rank (A B) = rank A
(b) rank (A B) = rank B
(c) rank (A B) > rank A
(d) rank (A B) > rank B
1 2 0 0
6. The rank of the matrix A = 2 3 0 0 is
0 0 0 0
(a) 0 (b) 1
(c) 2 (d) 3 (Bundelkhand 2001; Meerut 03)
[Link]
M-81
Fill in the Blank(s)
Fill in the blanks “……”, so that the following statements are complete and correct.
1. The rank of a matrix is the …… of any highest order non-vanishing minor of the
matr ix.
2. The rank of every non-singular matrix of order n is ……………
(Bundelkhand 2005)
3. The rank of a matrix in Echelon form is equal to the number of non-zero ……… of
the matrix.
(Bundelkhand 2008)
4. A is a non-zero column matrix and B is a non-zero row matrix, then rank
(AB) = ..... .
5. If a matrix B is obtained from A by an elementary transformation, A can also be
obtained from B by an elementary transformation of the …… type.
6. A matrix obtained from a …… by a single elementary transformation is called an
elementary matrix.
7. If a matrix A of order m × n can be expressed as :
Ir O
A~ , then rank of A is ………
O O (Meerut 2001)
8. If A be an m × n matrix of rank r, there exist non-singular matrices P and Q such
that PAQ = ..... .
9. If B be an m × n matrix obtained from an m × n matrix A by finite number of
elementary transformations of A, then A is called ……… to be B.
10. If two matrices A and B have the same size and the same rank, they are ……… .
1 0
11. The rank of a matrix is ………
0 1 (Agra 2008)
True or False
Write ‘T’ for true and ‘F’ for false statement.
1. The rank of the transpose of a matrix is the same as that of the original matrix.
2. Every elementary row transformation of a matrix can be obtained by
post-multiplication with the corresponding elementary matrix.
3. Elementary transformations change the rank of a matrix.
4. If A and B are two equivalent matrices, then rank A = rank B.
[Link]
M-82
5. The rank of a matrix does not alter by pre-multiplication or post-multiplication
with a non-singular matrix.
A nswers
Multiple Choice Questions
1. (b) 2. (b) 3. (c)
4. (d) 5. (b) 6. (c)
Fill in the Blank(s)
1. order 2. n 3. rows
4. 1 5. same 6. unit matrix
Ir O
7. r 8. 9. equivalent
O O
10. equivalent 11. 2
True or False
1. T 2. F 3. F
4. T 5. T
o
[Link]
M-83
3
Linear Equations
e shall devote this chapter to the study of the nature of solutions of a system of
W linear equations with the help of the theory developed in the preceding
chapters. Before discussing the solutions of linear equations, we shall discuss concepts
of linearly dependent and linearly independent sets of vectors.
3.1 Vectors
Definition: Any ordered n-tuple of numbers is called an n-vector. By an ordered n-tuple we
mean a set consisting of n numbers in which the place of each number is fixed. If
x1, x2 , ...., xn be any n numbers, then the ordered n -tuple X = ( x1, x2 , ..., xn) is called an
n-vector. The ordered triad ( x1, x2 , x3 ) is called a 3-vector. Similarly (1, 0 , 1, − 1) and
(1, 8, − 5, 7) are 4-vectors. The n numbers x1, x2 , ..., xn are called components of the
n-vector X = ( x1, x2 , ..., xn). A vector may be written either as a row vector or as a column
vector. If A be a matrix of the type m × n, then each row of A will be an n-vector and each
column of A will be an m-vector. A vector whose components are all zero is called a zero
vector and will be denoted by O.
If k be any number and X be any vector, then relative to the vector X, k is called a scalar.
Algebra of vectors. Since an n-vector is nothing but a row matrix or a column
matrix, therefore we can develop an algebra of vectors in the same manner as the
algebra of matrices.
[Link]
M-84
Equality of two vectors. Two n-vectors X and Y where X = ( x1, x2 ,..., xn) and
Y = ( y1, y2 , ..., yn) are said to be equal if and only if their corresponding components
are equal i. e., if xi = yi, for all i = 1, 2,.., n.
For example if
X = (1, 4, 7) and Y = (1, 4, 7),
then X = Y.
But if X = (1, 4, 7) and Y = (4, 1, 7), then X ≠ Y.
Addition of two vectors. If X = ( x1, x2 , …, xn) and Y = ( y1, y2 , …, yn) then by
definition X + Y = ( x1 + y1, x2 + y2 , …, xn + yn).
Thus X + Y is an n-vector whose components are the sums of corresponding
components of X and Y.
If X = (2, 4, − 7) and Y = (1, − 3, 5),
then X + Y = (2 + 1, 4 − 3, − 7 + 5) = (3, 1, − 2).
Multiplication of a vector by a scalar (number).
If k be any number and
X = ( x1, x2 , …, xn),
then by definition
kX = (kx1, kx2 , ..., kxn).
The vector k X is called the scalar multiple of the vector X by the scalar k.
If X = (1, 3, 8), then 4 X = (4, 12, 32)
and 0 X = (0 , 0 , 0 ).
Properties of addition and scalar multiplication of vectors. If X, Y, Z be any
three n-vectors and p, q be any two numbers, then obviously
(i) X + Y = Y + X. (ii) X + (Y + Z) = (X + Y) + Z.
(iii) p (X + Y) = pX + pY (iv) ( p + q)X = pX + qX.
(v) p (qX) = ( pq) X.
[Link]
M-85
3.2 Linear Dependence and Linear Independence of Vectors
(Agra 2005)
Linearly dependent set of vectors. Definition:
A set of r n-vectors X1, X2 , …, X r is said to be linearly dependent if there exist r scalars (numbers)
k1, k2 , ..., k r , not all zero, such that
k1X1 + k2 X2 + .. + k r X r = O,
where, O denotes the n-vector whose components are all zero.
Linearly independent set of vectors. Definition:
A set of r n-vectors X1, X2 , ..., X r is said to be linearly independent if every relation of the type
k1X1 + k2 X2 + … + k r X r = O
implies k1 = k2 = k3 … = k r = 0 .
Example 1: Show that the vectors X1 = (1, 2, 4), X2 = (3, 6, 12) are linearly dependent.
Solution: By a little inspection, we see that
3 X1 + (− 1) X 2 = (3, 6, 12) + (− 3, − 6, − 12) = (0 , 0 , 0 ) = O.
Thus there exist numbers k1 = 3, k2 = − 1 which are not all zero such that
k1X1 + k2 X2 = O.
Hence the vectors X1 and X2 are linearly dependent.
Example 2: Show that the set consisting only of the zero vector, O, is linearly dependent.
Solution : Let X = (0 , 0 , 0 , ...., 0 ) be an n-vector whose components are all zero. Then
the relation kX = O is true for some non-zero value of the number k. For example,1X = O
and 1 ≠ 0.
Hence the vector O is linearly dependent.
Example 3: Show that the vectors X1 = (1, 2, 3) and X 2 = (4, − 2, 7) are linearly independent.
Solution: Let k1 and k2 be two numbers such that k1X1 + k2 X2 = O,
i. e., k1 (1, 2, 3) + k2 (4, − 2, 7) = (0 , 0 , 0 )
[Link]
M-86
i.e., (k1 + 4 k2 , 2 k1 − 2 k2 , 3 k1 + 7 k2 ) = (0 , 0 , 0 ).
Equating the corresponding components, we get
k1 + 4 k2 = 0 , 2 k1 − 2 k2 = 0 , 3 k1 + 7 k2 = 0 .
The only common values of k1 and k2 which satisfy these equations are k1 = 0 , k2 = 0 .
Thus k1X1 + k2 X2 = O, iff k1 = k2 = 0.
Hence the vectors X1 and X2 are linearly independent.
Example 4: Show that the set of three 3-vectors
X1 = (1, 0 , 0 ), X2 = (0 , 1, 0 ), X3 = (0 , 0 , 1)
is linearly independent.
Solution : Let k1, k2 , k3 be three numbers such that
k1X1 + k2 X2 + k3 X3 = O,
i. e., k1(1, 0 , 0 ) + k2 (0 , 1, 0 ) + k3 (0 , 0 , 1) = (0 , 0 , 0 ),
i. e., (k1, 0 , 0 ) + (0 , k2 , 0 ) + (0 , 0 , k3 ) = (0 , 0 , 0 ),
i. e., (k1, k2 , k3 ) = (0 , 0 , 0 ).
Obviously this relation is true if and only if k1 = 0 , k2 = 0 , k3 = 0.
Hence the vectors X1, X2 , X3 are linearly independent.
3.3 A Vector as a Linear Combination of Vectors
Definition: A vector X which can be expressed in the form
X = k1X1 + k2 X2 + … + k r X r ,
is said to be a linear combination of the set of vectors X1, X2 , ..., X r .
Here k1, k2 , …, k r are any numbers.
The following two results are quite obvious :
(i) If a set of vectors is linearly dependent, then at least one member of the set can be expressed as
a linear combination of the remaining members.
(ii) If a set of vectors is linearly independent then no member of the set can be expressed as a linear
combination of the remaining members.
[Link]
M-87
3.4 The n-Vector Space
The set of all n vectors of a field F is called the n-vector space over F. It is usually
denoted by Vn( F ) or simply by Vn if the field is understood. Similarly the set of all
3-vectors is a vector space which is usually denoted by V3 . The elements of the field F
are known as scalars relatively to the vectors.
3.5 Sub-space of an n-Vector SpaceVn
Definition. A non-empty set, S, of vectors of Vn is called a vector sub-space of Vn, if
a + b belongs to S whenever a, b belong to S and k a belongs to S,where k is any scalar.
It is important to note that every sub-space of Vn contains the zero vector, being the
scalar product of any vector with the scalar zero.
Example: If a = (a1, a2 , a3 ) is any non-zero vector of V3 , then the set S of vectors ka is a
subspace of V3 , where k is a variable scalar which can take any value. The sum of any
two members k1a, k2 a of S, is the vector k1a + k2 a
i. e., (k1a1, k1a2 , k1a3 ) + (k2 a1, k2 a2 , k2 a3 )
i. e., (k1 + k2 ) a
which is also a member of S. Also the scalar multiple by any scalar x of any vector k1a of
S is the vector ( xk1)a which is again a member of S.
Hence the set S of vectors ka is a subspace of V3 .
3.6 Vector Subspace Spanned by a given System of Vectors
Let a, b, c be any three vectors of V3 . The set S of all vectors of the form xa + yb + zc,
where x, y, z are any scalars, is a subspace of V3 . For, if x1a + y1 b + z1c, x2 a + y2 b + z2 c
be any two members of S, then
( x1a + y1 b + z1c) + ( x2 a + y2 b + z2 c) = ( x1 + x2 )a + ( y1 + y2 ) b + (z1 + z2 )c,
which is also a member of S. Also if k be any scalar, then
k ( x1a + y1 b + z1c) = (kx1)a + (ky1) b + (kz1)c,
which is again a member of S. Thus S is a vector subspace and we say that S is a spanned
by the vectors a, b and c. More generally, if a1, a2 , ..., a r be a set of r fixed vectors of Vn,
then the set S of all n-vectors of the form p1a1 + p2 a2 + … pr a r where p1, p2 , ...., pr are
any scalars is a vector subspace of Vn.
[Link]
M-88
This vector space is said to be spanned by the vectors a 1, a 2, . . . . , a r . Thus a vector
space which arises as a set of all linear combinations of any given set of vectors,
is said to be spanned by the given set of vectors.
3.7 Basis and Dimension of a Subspace
A set of vectors a1, a2 , a3 ,..., a k belonging to the subspace S is said to be a basis of S,if
(i) the subspace S is spanned by the set a1, a2 , ..., a k and
(ii) the vectors a1, a2 , ..., a k are linearly independent.
It can be easily shown that every subspace, S, of Vn possesses a basis.
It can be easily shown that the vectors
e1 = (1, 0 , 0 , ..., 0 ), e2 = (0 , 1, 0 , ..., 0 ), e3 = (0 , 0 , 1, 0 , ..., 0), ...,
en = (0 , 0 , 0 , ..., 1)
constitute a basis of Vn.
We have already shown that these vectors are linearly independent. Moreover any
vector a = (a1, a2 , ..., an) of Vn is expressible as a = a1e1 + a2 e2 + a3 e3 + … + anen. Hence
the vectors e1, e2 , e3 , ..., en constitute a basis of Vn.
Theorem: A basis of a subspace, S, can always be selected out of a set of vectors which span S.
Let a1, a2 , ..., a r be a set of vectors which spans a subspace S. If these vectors are linearly
independent, they already constitute a basis of S as they span S. In case they are
linearly dependent, some member of the set is a linear combination of the members.
Deleting this member we obtain another set which also spans.
Continuing in this manner we shall ultimately, in a finite number to steps arrive at a
basis of S.
A vector subspace may (and in fact does) possess several bases
For examlple, if
a1 = (1, 0 , 0 ), a2 = (0 , 1, 0 ), a3 = (0 , 0 , 1), b1 = (1, 1, 1), b2 = (1, 1, 0 ), b3 = (1, 0 , 0 ),
then a1 , a2 , a3 constitute a basis of V3 . Also b1, b2 , b3 constitute a basis of V3 .
But it is important to note that the number of members in any one basis of a subspace is the
same as in any other basis. This number is called the dimension of the subspace.
[Link]
M-89
We have already shown that one basis of Vn possesses n members. Therefore every
basis of Vn must possess n members. Thus Vn is of dimension n. In a particular the
dimension of V3 is 3.
Also it can be easily shown that if r be the dimension of a subspace S and if a1, a2 ,… a k
be a linearly independent set of vectors belonging to S, then we can always find vectors
a k + 1, a k + 2 , .... a r such that the vectors a1, a2 , . . . , a k , a k + 1,..., a r constitute a basis of S.
In other words we can say that every linearly independent set of vectors belonging to a
subspace S can always be extended so as to constitute a basis of S.
Moreover if r be the dimension of a subspace S, then every set of more than r members
of S will be linearly dependent.
Intersection of subspaces. If S and T be two subspaces of Vn, then the vectors
common to both S and T also constitute a subspace. This subspace is called the
intersection of the subspaces S and T .
3.8 Row Rank and Column Rank of a Matrix
Row rank of a matrix. Let A = [aij] be any m × n matrix. Each of the m rows of A
consists of n elements. Therefore the row vectors of A are n-vectors. These row vectors
of A will span a subspace R of Vn. This subspace R is called the row space of the matrix A.
The dimension r of R is called the row rank of A. In other words the row rank of a matrix
A is equal to the maximum number of linearly independent rows of A.
Row rank of a matrix. Definition: The maximum number of linearly independent rows
of a matrix A is said to be the row rank of the matrix A.
Left nullity of a matrix. Suppose X is an m-vector written in the form of a row
[Link] the matrix product XA is defined. The subspace S of Vm generated by the
row vectors X belonging to Vm such that XA = O is called the row null space of the matrix
A. The dimension s of Sis called the left nullity or row nullity of the matrix A.
We shall now prove that the sum of the row rank and the row nullity of a matrix is equal to the
number of rows, i. e.,
r + s = m.
Proof: Since the row space R of A is spanned by the row vectors of A, therefore it will
be a set of all vectors of the form
x1(a11, a12 , a13 , ..., a1n) + x2 (a21, a22 , a23 , ...., a2 n) + …
… + xm (am1, am2 , am3 , ..., amn)
[Link]
M-90
i. e., of the form ( x1a11 + x2 a21 + … + xm am1, x1a12 + x2 a22 + .... + xm am2 ,
x1a13 + x2 a23 + … + xm am3 , ..., x1a1n + x2 a2 n + … + xm amn)
i. e., of the form XA, where X = ( x1, x2 , ..., xm ) is an m-vector.
Let u1, u2 , ..., u s be a basis of the subspace S of Vm generated by all vectors X such that
XA = O. Then, we have
u1A = u2 A = … = u s A = O.
Since the vectors u1, u2 , ..., u s belong to Vm and form a linearly independent set,
therefore we can find vectors u s + 1, u s + 2 , …, u m in Vm such that the vectors
u1, u2 , ..., u s , u s + 1, ..., u m constitute a basis of Vm . Then every vector X belonging to Vm
can be expressed in the form
X = h1u1 + h2 u2 + .... + hm u m .
Now every member of the subspace R is expressible as XA
i. e., as (h1u1 + h2 u2 + … + hm u m ) A
i. e., as h1u1A + h2 u2 A + … + hs u s A + hs + 1u s + 1A + hs + 2 u s + 2 A + … + hm u m A
i. e., as hs + 1u s + 1A + hs + 2 u s + 2 A + … + hm u m A.
Therefore the m − s n-vectors u s + 1A, u s + 2 A, ..., u m A span R. In fact, these vectors
form a basis of R. For any relation of the form
k s + 1u s + 1A + k s + 2 u s + 2 A + … + k m u m A = O
implies (k s + 1u s + 1 + k s + 2 u s + 2 + … + k m u m )A = O,
which shows that k s + 1u s + 1 + k s + 2 u s + 2 + … + k m u m is a member of the subspace S and
as such it can be linearly expressed in terms of the basis u1, u2 , ..., u s of S.
But the vectors u1, u2 , ..., u m are linearly independent. Therefore a relation of the form
k s + 1u s + 1 + k s + 2 u s + 2 + … + k m u m = p1u1 + p2 u2 + .....+ ps u s will exist if and only if
k s + 1 = k s + 2 = … = k m = 0 . Hence the vectors u s + 1A, u s + 2 A , ..., u m A are linearly
independent and form a basis of R. Thus the dimension of R is m − s.
Hence r=m−s or r + s = m.
[Link]
M-91
3.9 Column Rank of a Matrix
Let A = [aij] be any m × n matrix. Each of the n columns of A consists of m elements.
Therefore the column vectors of A are m-vectors. These column vectors of A will span a
subspace C of Vm . This subspace C is called the column space of the matrix A. The
dimension c of C is called the column rank of A. In other words the column rank of a
matrix A is equal to the maximum number of linearly independent columns of A.
Column rank of a matrix. Definition. The maximum number of linearly independent
columns of a matrix A is said to be the column rank of the matrix A.
Right nullity of a matrix. Suppose Y is an n-vector written in the form of a column
vector. Then the matrix product AY is defined. The subspace T of Vn generated by the
column vectors Y belonging to Vn such that AY = O is called the column null space of the
matrix [Link] dimension t of T is called the right nullity or column nullity of the matrix A.
As in 3.8, we can show that c + t = n.
3.10 Invariance of Row Rank under E-row Operations
Theorem. Row equivalent matrices have the same row rank.
Proof. Let A be any given m × n matrix. Let B be a matrix row equivalent to A. Since B
is obtainable from A by a finite chain of E-row operations and every E-row operation is
equivalent to pre-multiplication by the corresponding E-matrix, there exist E-matrices
E1, E2 , ..., Ek each of the type m × m such that
B = Ek Ek − 1 … E2 E1A, i. e., B = PA ,
where P = Ek Ek − 1 … E2 E1 is a non-singular matrix of the type m × m.
Let us write
p11 p12 ... p1m R1
p21 p22 ... p2 m R2
B = PA = ... ... ... ... ... …(1)
... ... ... ... ...
pm1 pm2 ... pmm R m
where the matrix A has been expressed as a matrix of its row sub-matrices
R1, R2 , ..., R m .
[Link]
M-92
From the product of the matrices on the R.H.S. of (1), we observe that the rows of the
matrix B are
p11R1 + p12 R2 + … + p1m R m ,
p21R1 + p22 R2 + … + p2 m R m ,
… … … … …
… … … … …
pm1R1 + pm2 R2 + … + pmm R m .
Thus we see that the rows of B are all linear combinations of the rows R1, R2 , ...., R m of
A. Therefore every member of the row space of B is also a member of the row space of A.
Similarly by writing A = P−1B and giving the same reasoning we can prove that every
member of the row space of A is also a member of the row space of B. Therefore the row
spaces of A and B are identical.
Thus we see that elementary row operations do not alter the row space of a matrix.
Hence the row rank of a matrix remains invariant under E-row transformations.
Note. From the above theorem we also conclude that pre-multiplication by a
non-singular matrix does not alter the row rank of a matrix.
3.11 Invariance of Column Rank under E-column Operations
Theorem. Column equivalent matrices have the same column rank.
Or
Post-multiplication by a non-singular matrix does not alter the column rank of a matrix.
Proof: Proceeding in the same way as in 3.10, we can show that post-multiplication
with a non-singular matrix does not alter the column space and therefore the column
rank of a matrix.
Note. Since every n-rowed E-matrix is obtained from In by single E-operation (row or
column operation as may be desired), therefore the row rank and column rank of an
E-matrix are equal to n.
[Link]
M-93
3.12 Invariance of Column Rank under E-row Operations
Theorem: Row equivalent matrices have the same column rank.
Proof: Let A be any given m × n matrix and let B be a matrix row equivalent to A. Then
there exists a non-singular matrix P such that B = PA.
For every column vector X such that AX = O, we have
BX = (PA) X = P (AX) = PO = O.
Since B = PA, therefore A = P−1B.
Therefore for every vector X such that BX = O, we have
AX = (P−1B) X = P−1 (BX) = P−1O = O.
Thus we see that the matrices A and B have the same right nullities and consequently
their column ranks are equal.
Similarly we can prove that column equivalent matrices have the same row rank.
3.13 Theorem
If r be the row rank of an m × n matrix A then there exists a non-singular matrix, P such that
K
PA = ,
O
where K is an r × n matrix consisting of a set of r linearly independent rows of A.
Proof. If the row rank r of A is zero, we have nothing to prove. Therefore let us
assume that r > 0. The matrix A has then r linearly independent rows. By elementary
row operations on A we can bring these linearly independent rows in the first r places.
Since the last m − r rows are now linear combinations of the first r rows, they can be
made zero by E-row operations without altering the first r rows.
Thus we see that the matrix A is row equivalent to a matrix B such that
K
B= ,
O
where K is an r × n matrix consisting of a set of r linearly independent rows of A.
[Link]
M-94
Since every elementary row operation is equivalent to pre-multiplication by the
corresponding E-matrix and the product of E-matrices is a non-singular matrix,
therefore there exists a non-singular matrix P such that
K
PA = ⋅
O
Similarly considering column transformations instead of row transformations, we can
show that if c be the column rank of a matrix A, then there exists a non-singular matrix R such
that
AR = [L O],
where L is an m × c matrix consisting of, c, linearly independent columns of A.
3.14 Equality of Row Rank, Column Rank and Rank
Theorem 1. The row rank of a matrix is the same as its rank.
Proof: Let, s be the row rank and r, the rank of an m × n matrix A. Since the matrix A is
of row rank s, therefore by 3.13, there exists a non-singular matrix P such that
K
PA = , where K is an s × n matrix.
O
Now we know that pre-multiplication by a non-singular matrix does not alter the rank
of a matrix.
∴ Rank (PA) = Rank A = r.
But each minor of order (s + 1) of the matrix PA involves atleast one row of zeros.
∴ Rank (PA) ≤ s.
∴ r≤s
Again, since the rank of the matrix A is r, therefore by 2.12 of chapter 2 there exists a
non-singular matrix R such that
G
RA = ,
O
where G is an r × n matrix.
Now we know that pre-multiplication by a non-singular matrix does not alter the row
rank of a matrix.
[Link]
M-95
∴ Row rank (RA) = Row rank A = s.
But the matrix RA has only r non-zero rows. Therefore the row rank of RA can, at the
most, be equal to r.
∴ s≤ r
Hence r = s.
Theorem 2. The column rank of a matrix is the same as its rank.
Proof: Let the matrix A′ be the transpose of the matrix A. Then the columns of A are
the rows of A′ .
∴ the column rank of A = the row rank of A′ = the rank of A′ = the rank of A.
Thus from theorems 1 and 2, we conclude that the rank, row rank and column rank of a
matrix are all equal. In other words we can say that the maximum number of linearly
independent rows of a matrix is equal to the maximum number of its linearly
independent columns and is equal to the rank of the matrix.
Thus we have proved that the row rank, the column rank and the rank of a matrix are
all equal. Therefore sometimes the rank of a matrix is also defined as the maximum number of
linearly independent row vectors or column vectors.
We shall now first consider systems of linear homogeneous equations and then
proceed to discuss systems of non-homogeneous linear equations.
3.15 Homogeneous Linear Equations
a11 x1 + a12 x2 + … + a1n xn = 0 ,
a21 x1 + a22 x2 + … + a2 n xn = 0 ,
Suppose … … … … … … … …(1)
… … … … … … …
am1 x1 + am2 x2 + … + amn xn = 0
is a system of m homogeneous equations in n unknowns x1, x2 , … , xn. Let
a11 a12 ... a1n x1 0
a21 a22 ... a2 n x2 0
a a32 ... a3 n x 0
31 3
A= X= O=
... ... ... ... ... ...
... ... ... ... ... ...
am1 am2 ... amn m × n, x n n × 1, 0 m × 1,
[Link]
M-96
where A, X, O are m × n, n × 1, m × 1matrices respectively. Then obviously we can write
the system of equations (1) in the form of a single matrix equation
AX = O. …(2)
The matrix A is called the coefficient matrix of the system of equations (1).
Obviously x1 = 0 , x2 = 0 , …, xn = 0 i. e., X = O is a solution of (1). It is a trivial
(self-obvious) solution of (1).
Again suppose X1 and X2 are two solutions of (2). Then their linear combination k1X1 + k2 X2 ,
where k1 and k2 are any arbitrary numbers, is also a solution of (2).
We have A(k1X1 + k2 X2 ) = k1(AX1) + k2 (AX2 )
= k1O + k2O = O. [∵ AX1 = O and AX2 = O]
Hence k1X1 + k2 X2 is also a solution of (2).
Therefore the collection of all the solutions of the system of equations AX = O forms a sub-space of
the n-vector space Vn.
Theorem: The number of linearly independent solutions of the system of m homogeneous
linear equations in n variables, AX = O, is (n − r), where r is the rank of the matrix A.
Proof:
a11 a12 ... a1n x1
a21 a22 ... a2 n x2
a a32 ... a3 n x
31 3
Let A= and X =
... ... ... ... ...
... ... ... ... ...
am1 am2 ... amn m × n , xn n × 1.
Since the rank of the coefficient matrix A is r, therefore it has r linearly independent
columns. Without loss of generality we can suppose that the first r columns from the
left of the matrix A are linearly independent, because it amounts only to renaming the
components of X.
The matrix A can be written as A = [C1, C2 , …, Cr , … Cn] 1 × n , where C1, C2 , ..., Cn are the
column vectors of the matrix A each of them being an m-vector.
The equation AX = O can now be written as the vector equation
x1C1 + x2C2 + … + xrCr + xr + 1Cr + 1 + … + xnCn = O . …(1)
[Link]
M-97
Since each of the vectors C r + 1 , C r + 2 , … , C n is a linear combination of the vectors
C1, C2 , ..., Cr , therefore we have relations of the type
Cr + 1 = p11C1 + p12C2 + … + p1rCr ,
Cr + 2 = p21C1 + p22C2 + … + p2 rCr ,
...(2)
… … … … … …
Cn = pk1C1 + pk2C2 + … + pkrCr , where k = n − r
The relations (2) can be written in the form
p11C1 + p12C2 + … + p1rCr − 1. Cr + 1 + 0 . Cr + 2 + … + 0 . Cn = O,
p21C1 + p22C2 + … + p2 rCr + 0 . Cr + 1 − 1. Cr + 2 + … + 0 . Cn = O,
...(3)
… … … … … … … … … … … …
pk1C1 + pk2C2 + … + pkrCr + 0 . Cr + 1 + 0 . Cr + 2 + ..... − 1. Cn = O
Comparing (1) and (3), we find that the vectors
p11 p11 p k1
p12 p22 pk2
... ... ...
... ... ...
p1r p2 r pkr
X1 = −1 , X 2 = 0 ,..., X n − r = 0
0 −1 0
0 0 0
... ... ...
... ... ...
0 0 0
are (n − r) solutions of the equation AX = O.
The vectors X1, X2 , ..., X n − r form a linearly independent set. For, if we have a relation
of type
l1X1 + l2 X2 + … + ln − r X n − r = O, …(4)
then comparing the (r + 1)th, (r + 2)th, … , nth components on both sides of (4), we get
− l1 = 0 , − l2 = 0 , ..., − ln − r = 0 ,
i. e., the vectors X1, X2 , ..., X n − r are linearly independent.
[Link]
M-98
It can now be easily seen that every solution of the equation AX = O is some suitable
linear combination of these n − r solutions X1, X2 , ...., X n − r .
Suppose the vector X, with components x1, x2 , ..., xn is any solution of the equation
AX = O. The the vector
X + xr + 1X1 + xr + 2 X2 + … + xnX n − r , …(5)
which, being a linear combination of solutions, is also a solution. It is quite obvious
that the last n − r components of the vector (5) are all equal to zero. Let z1, z2 , ..., z r be
the first r components of the vector (5). Then the vector whose components are
(z1, z2 , ..., z r , 0 , 0 , ..., 0 ) is a solution of the equation
AX = O.
Therefore from (1), we have
z1C1 + z2C2 + … + z rCr = O.
But the vectors C1, C2 , ..., Cr
are linearly independent. Therefore, we have z1 = 0 , z2 = 0 , ..., z r = 0 . Hence (5) is a
zero vector.
Therefore
X = − xr + 1X1 − xr + 2 X2 − … − xnX n − r .
Thus every solution X is a linear combination of the n − r linearly independent
solutions X1, X2 , …, X n − r .
Therefore the set of solutions
{ X1, X2 , …, X n − r }
forms a basis of the vector space of all the solutions of the system of equations AX = O.
3.16 Some Important Conclusions About the Nature of
Solutions of the Equation AX = O
Suppose we have m equations in n unknowns. Then the coefficient matrix A will be of
the type m × n. Let r be the rank of the matrix A. Obviously r cannot be greater than n
(the number of columns of the matrix A). Therefore we have either r = n or r < n.
[Link]
M-99
Case I. If r = n, the equation AX = O will have n − n i. e., no linearly independent
solutions. In this case the zero solution will be the only solution. We know that zero
vector forms a linearly dependent set.
Case II. If r < n, we shall have n − r linearly independent solutions. Any linear
combination of these n − r solutions will also be a solution of AX = O. Thus in this case
the equation AX = O will have an infinite number of solutions.
Case III. Suppose m < n i. e., the number of equations is less than the number of
unknowns. Since r ≤ m, therefore r is definitely less than n. Hence in this case the given
system of equations must possess a non-zero solution. The number of solutions of the
equation AX = O will be infinite.
3.17 Fundamental Set of Solutions of the Equation AX = O
Suppose the rank r of the coefficient matrix A is less than the number of the unknowns
n. In this case the given equations have a set of n − r linearly independent solutions and
every possible solution is a linear combination of these n − r solutions. This set of n − r
solutions is called a fundamental set of solutions of the equation AX = O.
Definition: A set of linearly independent solutions X1, X2 , ..., X k of the system of
homogeneous equations AX = O is called the fundamental system of solutions of AX = O, if every
solution X of AX = O can be written as a linear combination of these vectors i. e., in the form
X = c1X1 + c2 X2 + … + c k X k ,
where c1, c2 , ..., c k are suitable numbers.
3.18 Working Rule for Finding the Solutions of the Equation
AX = O
Reduce the coefficient matrix A to Echelon form by applying elementary row
transformations only. This Echelon form will help us to know the rank of the matrix
A. Suppose the matrix A is of the type m × n and its rank comes out to be r. If r < m, then
in the process of reducing the matrix A to Echelon form, (m − r) equations will be
eliminated. The given system of m equations will thus be replaced by an equivalent
system of r equations. Solving these r equations (by Cramer’s rule or otherwise), we
can express the values of some r unknowns in terms of the remaining n − r unknowns.
These n − r unknowns can be given any arbitrarily chosen values.
If r = n, the zero solution (trivial solution) will be the only solution. If r < n, there will
be an infinity of solutions.
[Link]
M-100
Example 5: Does the following system of equations possess a common non-zero solution?
x + 2 y + 3 z = 0, 3 x + 4 y + 4 z = 0 , 7 x + 10 y + 12 z = 0 . (Lucknow 2005)
Solution : The given system of equations can be written in the form of the single matrix
equation
1 2 3 x 0
AX = 3 4 4 y = 0 = O.
7 10 12 z 0
We shall start reducing the coefficient matrix A to triangular form by applying only
E-row transformations on it. Applying R2 → R2 − 3 R1, R3 → R3 − 7 R1, the given
system of equations is equivalent to
1 2 3 x
0 −2 −5 y = O.
0 −4 −9 z
Here we find that the determinant of the matrix on the left hand side of this equation is
not equal to zero. Therefore the rank of this matrix is 3. So there is no need of further
applying E-row transformations on the coefficient matrix. The rank of the coefficient
matrix A is 3, i. e., equal to the number of unknowns. Therefore the given system of
equations does not possess any linearly independent solution. The zero solution, i. e.,
x = y = z = 0 is the only solution of the given system of equations.
Example 6: Solve completely the system of equations
x + 3 y − 2 z = 0 , 2 x − y + 4 z = 0 , x − 11 y + 14 z = 0 . (Meerut 2005B)
Solution : The given system of equations is equivalent to the single matrix equation
1 3 −2 x
AX = 2 −1 4 y = O.
1 −11 14 z
We shall reduce the coefficient matrix A to Echelon form by applying only E-row
operations on it. Performing R2 → R2 − 2 R1, R3 → R3 − R1, we have
[Link]
M-101
1 3 −2 x
0 −7 8 y = O.
0 −14 1 z
Performing R3 → R3 − 2 R2 , we have
1 3 −2 x
0 −7 8 y = O.
0 0 0 z
The coefficient matrix is now triangular. The coefficient matrix being of rank 2, the
given system of equations possess 3 − 2 = 1 linearly independent solution. We shall
assign arbitrary values to n − r = 3 − 2 = 1 variable and the remaining r = 2 variables
shall be found in terms of these. The given system of equations is equivalent to
x + 3 y − 2 z = 0, − 7 y + 8 z = 0 .
8 10
Thus y= z, x = − z.
7 7
Choose z = c.
8 10
Then y= c, x = − c.
7 7
10 8
Hence x=− c, y = c, z = c
7 7
constitute the general solution of the given system, where c is an arbitrary parameter.
Remark. In matrix form, the general solution can be expressed as
10 10
− c − 7
x 7
8 8
y = c =c , where c is an arbitrary number.
z 7 7
c 1
Example 7: Does the following system of equations possess a common non-zero solution?
x + y + z = 0 , 2 x − y − 3 z = 0 , 3 x − 5 y + 4 z = 0 , x + 17 y + 4 z = 0 .
Solution: The given system of equations is equivalent to the single matrix equation
[Link]
M-102
1 1 1
x
2 −1 −3
AX = y = O.
3 −5 4 z
1 17 4
We shall first find the rank of the coefficient matrix A by reducing it to Echelon form
by applying elementary row transformations only.
Applying R2 → R2 − 2 R1, R3 → R3 − 3 R1, R4 → R4 − R1, we get
1 1 1 1 1 1
0 −3 −5 0 −3 −5
A~ ~ by R3 → 3 R3 , R4 → 3 R4
0 −8 1 0 −24 3
0 16 3 0 48 9
1 1 1
0 −3 −5
~ by R3 → R3 − 8 R2 , R4 → R4 + 16 R2
0 0 43
0 0 −71
1 1 1
0 −3 −5 71
~ by R4 → R4 + R3 .
0 0 43 43
0 0 0
Above is the Echelon form of the coefficient matrix A. We have rank A = the number
of non-zero rows in this Echelon form = 3. The number of unknowns is also 3. Since
rank A is equal to the number of unknowns, therefore the given system of equations
does not possess any linearly independent solution. Thus the given system of
equations possesses no non-zero solution. Hence the zero solution i. e., x = y = z = 0 is
the only solution of the given system of equations.
Example 8: Find all the solutions of the following system of equations :
3 x + 4 y − z − 6w = 0, 2 x + 3 y + 2z − 3w = 0,
2 x + y − 14 z − 9 w = 0 , x + 3 y + 13 z + 3 w = 0 .
Solution: The given system of equations is equivalent to the single matrix equation
[Link]
M-103
3 4 −1 −6 x
2 3 2 −3 y
AX = z = O.
2 1 −14 −9
1 3 13 3 w
We shall first find the rank of the coefficient matrix A by reducing it to Echelon form
by applying E-row transformations only.
Applying R4 ↔ R1, we get
1 3 13 3
2 3 2 −3
A~
2 1 −14 −9
3 4 −1 −6
1 3 13 3
0 −3 −20 −9
~
0 −5 −40 −15
0 −5 −40 −15
By R2 → R2 − 2 R1, R3 → R3 − 2 R1, R4 → R4 − 3 R1
1 3 13 3
0 1 8 3
~
0 1 8 3
0 1 8 3
1 1 1
by R2 → − R2 , R3 → − R3 , R4 → − R4
3 5 5
1 3 13 3
0 1 8 3
~ by R3 → R3 − R2 , R4 → R4 − R2 .
0 0 0 0
0 0 0 0
The rank of A is obviously 2 which is less than the number of unknowns 4. Therefore
the given system of equations possesses 4 − 2, i. e., 2 linearly independent solutions.
The given system of equations is equivalent to the equation
[Link]
M-104
1 3 13 3 x 0
0 1 8 3 y 0
0 0 0 0 z = 0 ⋅
0 0 0 0 w 0
Thus the given system of four equations is equivalent to the system of two equations,
i. e.,
x + 3 y + 13 z + 4 w = 0
⋅
y + 8z + 3w = 0
From these equations, we get
y = − 8 z − 3 w, x = − 3 (− 8 z − 3 w) − 13 z − 3 w
i. e., y = − 8 z − 3 w, x = 11z + 6 w.
Hence x = 11c1 + 6 c2 , y = − 8 c1 − 3 c2 , z = c1, w = c2 constitute the general solution of
the given system of equations, where c1 and c2 are arbitrary numbers. Since we can give
any arbitary values to c1 and c2 , therefore the given system of equations have an infinite
number of solutions.
Remark: In matrix form, the general solution of the given system of equations can
be expressed as
x 11c1 + 6 c2 11 6
y − 8 c1 − 3 c2 − 8 − 3
z = 1c + 0 c = c1 1 + c2 0 .
1 2
w 0 c1 + 1c2 0 1
x 11 x −6
y − 8 y −3
Here =
z 1 and z = 0
w 0 w 1
are two linearly independent solutions of the given system of equations and all their
linear combinations will also be the solutions of the given system of equations.
Example 9: Show that the only real value of λ for which the following equations have non-zero
solutions is 6 :
x + 2 y + 3 z = λx, 3 x + y + 2 z = λy, 2 x + 3 y + z = λz .
(Kanpur 2011; Bundelkhand 10; Lucknow 10)
[Link]
M-105
Solution: The given system of equations is equivalent to the single matrix equation
1 − λ 2 3 x
AX = 3 1− λ 2 y = O.
2 3 1 − λ z
If the given system of equations is to possess a non-zero solution, the coefficient
matrix A must be of rank less than 3. If the matrix A is to be of rank less than 3 its
determinant must be equal to zero. Thus we have
1 − λ 2 3
3 1− λ 2 =0
2 3 1 − λ
6 − λ 6−λ 6 − λ
or 3 1− λ 2 = 0,
2 3 1 − λ
adding the second and the third rows to the first
1 1 1
or (6 − λ ) 3 1− λ 2 =0
2 3 1 − λ
1 1 0
or (6 − λ ) 3 −λ − 2 −1 = 0 , C2 − C1, C3 − C1
1 1 − λ − 1
or (6 − λ ) [(λ + 2) (λ + 1) + 1] = 0
or (6 − λ ) [λ 2 + 3 λ + 3] = 0 .
− 3 ± (9 − 12)
The roots of the equation λ 2 + 3 λ + 3 = 0 are λ = i. e.,are imaginary.
2
Hence the only real value of λ for which the system of equations is to have a non-zero
solution is 6.
[Link]
M-106
Comprehensive E xercise 1
Find all the solutions of the following system of linear homogeneous equations :
1. 2 x − 3 y + z = 0 , x + 2 y − 3 z = 0 , 4 x − y − 2 z = 0 .
2. x + y − 3 z + 2 w = 0 , 2 x − y + 2 z − 3 w = 0 , 3 x − 2 y + z − 4 w = 0 ,
− 4 x + y − 3z + w = 0.
3. x + y + z = 0 , 2 x + 5 y + 7 z = 0 , 2 x − 5 y + 3 z = 0 .
4. x + 2 y + 3 z = 0 , 2 x + 3 y + 4 z = 0 , 7 x + 13 y + 19 z = 0 .
5. 4 x + 2 y + z + 3 u = 0 , 6 x + 3 y + 4 z + 7 u = 0 , 2 x + y + u = 0 .
6. 2 x − 2 y + 5 z + 3 w = 0 , 4 x − y + z + w = 0 , 3 x − 2 y + 3 z + 4 w = 0 ,
x − 3 y + 7 z + 6 w = 0.
7. x − 2 y + z − w = 0 , x + y − 2 z + 3 w = 0 , 4 x + y − 5 z + 8 w = 0 ,
5 x − 7 y + 2z − w = 0
A nswers 1
1. x = 0 , y = 0 , z = 0 2. x = 0 , y = 0 , z = 0 , w = 0
3. x = 0 , y = 0 , z = 0 4. x = 2 c , y = − 2 c , z = c
5. x = c1, u = c2 , y = − 2 c1 − c2 , z = − c2
5 7
6. x = c , y = 4 c , z = c , w = c
9 9
5 4
7. x = c1 − c2 , y = c1 − c2 , z = c1, w = c2
3 3
3.19 System of Linear Non-homogeneous Equations
Sometimes we think that we can solve every two simultaneous equations of the type
a1 x + b1 y = c1
.
a2 x + b2 y = c2
But it is not so. For example, consider the simultaneous equations
3x + 4 y = 5
.
6 x + 8 y = 13
There is no set of values of x and y which satisfies both these equations. Such
equations are said to be inconsistent.
[Link]
M-107
Let us take another example. Consider the simultaneous equations
3x + 4 y = 5
.
6 x + 8 y = 10
These equations are consistent since there exist values of x and y which satisfy both of
4 5
these equations. We see that x = − c + , y = c constitute a solution of these
3 3
equations, where c is arbitrary. Thus these equations possess an infinite number of
solutions.
Now we shall discuss the nature of solutions of a system of non-homogeneous linear
equations :
a11 x1 + a12 x2 + … + a1n xn = b1
a21 x1 + a22 x2 + … + a2 n xn = b2
Let … … … … … … … … …(1)
… … … … … … … …
am1 x1 + am2 x2 + .. + amn xn = bm
be a system of m non-homogeneous equations in n unknown x1, x2 , ..., xn.
If we write
a11 a12 … a1n x1 b1
a21 a22 … a2 n x2 b2
A= , X = ,B =
… … … … … …
am1 am2 … amn m × n xn bm
n ×1 m ×1
where A, X, B are m × n, n × 1and m × 1 matrices respectively, the above equations can
be written in the form of a single matrix equation AX = B.
Any set of values of x1, x2 , ..., xn which simultaneously satisfy all these equations is called a
solution of the system (1). When the system of equations has one or more solutions, the equations
are said to be consistent otherwise they are said to be inconsistent.
The matrix
a11 a12 … a1n b1
a21 a22 … a2 n b2
[A B] =
… … … … …
am1 am2 … amn bm
is called the augmented matrix of the given system of equations.
[Link]
M-108
3.20 Condition for Consistency
Theorem. The system of equations AX = B is consistent i. e., possesses a solution, if any only if
the coefficient matrix A and the augmented matrix [A B] are of the same rank.
[Meerut 2007B; Kanpur 11; Lucknow 05]
Proof. Let C1, C2 , ..., C n denote the column vectors of the matrix A. The equation
AX = B is then equivalent to
x1
x2
[C1, C2 , ...., Cn] … = B i. e., x1C1 + x2C2 + … + xnCn = B. …(1)
…
xn
Let now r be the rank of the matrix A. The matrix A has then r linearly independent
columns and without loss of generality, we can suppose that the first r columns
C1, C2 , ..., Cr form a linearly independent set so that each of the remaining n − r
columns is a linear combination of these r columns.
The condition is necessary. If the given system of equations is consistent, there
must exist n scalars (numbers) k1, k2 , ..., k n such that
k1C1 + k2C2 + …… + k nCn = B. …(2)
Since each of the n − r columns Cr + 1, Cr + 2 , ...., Cn is a linear combination of first r
columns C1, C2 , ..., Cr it is obvious from (2) that B is also a linear combination of
C1, C2 , ..., Cr . Thus the maximum number of linearly independent columns of the
matrix [A B] is also r. Therefore the matrix [A B] is also of rank r. Hence the matrices A
and [A B] are of the same rank.
The condition is sufficient. Now suppose that the matrices A and [A B] are of the
same rank r. The maximum number of linearly independent columns of the matrix
[A B] is then r. But the first r columns C1, C2 , ..., Cr of the matrix [A B] already form a
linearly independent set. Therefore the column B should be expressed as a linear
combination of the columns C1, C2 , ...., Cr .
Thus there exist r scalars k1, k2 , ..., k r such that
k1C1 + k2C2 + … + k rCr = B. …(3)
Now (3) may be written as
k1C1 + k2C2 + … + k rCr + 0 . Cr + 1 + 0 . Cr + 2 + … + 0 . Cn = B. …(4)
[Link]
M-109
Comparing (1) and (4), we see that
x1 = k1, x2 = k2 , ...., xr = k r , xr + 1 = 0 , xr + 2 = 0 , ..., xn = 0
consitute a solution of the equation AX = B.
Therefore the given system of equations is consistent.
3.21 Condition for a System of n Equations in n Unknowns to
have a Unique Solution
Theorem: If A be an n-rowed non-singular matrix, X be an n × 1matrix,B be an n × 1matrix,
the system of equations AX = B has a unique solution.
Proof: If A be an n-rowed non-singular matrix, the ranks of the matrices A and [A B]
are both n. Therefore the system of equations AX = B is consistent i. e., possesses a
solution.
Pre-multiplying both sides of AX = B by A −1, we have
A −1AX = A −1B i. e., IX = A −1B i. e., X = A −1B
is a solution of the equation AX = B.
To show that the solution is unique, let us suppose that X1 and X2 be two solutions of
AX = B.
Then AX1 = B, AX 2 = B ⇒ AX 1 = AX 2 ⇒ A −1AX1 = A −1AX2
⇒ IX 1 = IX 2 ⇒ X 1 = X 2.
Hence the solution is unique.
3.22 Working Rule for Finding the Solution of the Equation
AX = B
Suppose the coefficient matrix A is of the type m × n, i. e., we have m equations in n
unknowns. Write the augmented matrix [A B] and reduce it to Echelon form by
applying only E-row transformations on it. This Echelon form will enable us to know
the ranks of the augmented matrix [A B] and the coefficient matrix A. Then the
following different cases arise :
Case I. Rank A < Rank [A B] .
[Link]
M-110
In this case the equations AX = B are inconsistent i. e., they have no solution.
Case II. Rank A = Rank [A B] = r (say).
In this case the equations AX = B are consistent i. e., they possess a solution. If r < m,
then in the process of reducing the matrix [A B] to Echelon form, (m − r) equations
will be eliminated. The given system of m equations will then be replaced by an
equivalent system of r equations. From these r equations we shall be able to express
the values of some r unknowns in terms of the remaining n − r unknowns which can be
given any arbitrarily chosen values.
If r = n, then n − r = 0, so that no variable is to be assigned arbitrary values and
therefore in this case there will be a unique solution.
If r < n, then n − r variables can be assigned arbitrary values. So in this case there will be
an infinite number of solutions. Only n − r + 1 solutions will be linearly independent
and the rest of the solutions will be linear combinations of them.
If m < n, then r ≤ m < n. Thus in this case n − r > 0. Therefore when the number of
equations is less than the number of unknowns, the equations will always have an
infinite number of solutions, provided they are consistent.
Example 10: Show that the equations
x + y + z = − 3, 3 x + y − 2 z = − 2, 2 x + 4 y + 7 z = 7
are not consistent. (Avadh 2008; Purvanchal 07)
Solution : The given system of equations is equivalent to the single matrix equation
1 1 1 x −3
AX = 3 1 −2 y = −2 = B.
2 4 7 z 7
1 1 1 : −3
The augmented matrix [A B] = 3 1 −2 : −2.
2 4 7 : 7
We shall reduce the augmented matrix [A B] to Echelon form by applying E-row
transformations only. Applying
R2 → R2 − 3 R1, R3 → R3 − 2 R1, we get
[Link]
M-111
1 1 1: −3
[A B] ~ 0 −2 −5 : 7
0 2 5 : 13
1 1 1: −3
~ 0 −2 −5 : 7, aplying R3 → R3 + R2 .
0 0 0 : 20
Above is the Echelon form of the matrix [A B]. We have rank [A B] = the number of
non-zero rows in this Echelon form = 3
Also by the same E-row transformations, we get
1 1 1
A ~ 0 −2 −5.
0 0 0
Obviously rank A = 2.
Since Rank A ≠ Rank [A B], therefore the given equations are inconsistent i. e., they
have no solution.
Remark: The inconsistency of the given equations can also be shown as below :
The given system of equations is equivalent to the matrix equation
1 1 1 x −3
0 −2 −5 y = 7.
0 0 0 z 20
This matrix equation is equivalent to the system of equations
x + y + z = −3
0 x − 2 y − 5z = 7 .
0 x + 0 y + 0 z = 20
The last equation shows that 0 = 20 , which is not possible. Hence the given equations
are inconsistent.
Example 11: Show that the equations
x + y + z = 6, x + 2 y + 3 z = 14, x + 4 y + 7 z = 30
are consistent and solve them. (Meerut 2010 B; Avadh 09)
[Link]
M-112
Solution : The given system of equations is equivalent to the single matrix equation
1 1 1 x 6
AX = 1 2 3 y = 14 = B.
1 4 7 z 30
1 1 1 M 6
The augmented matrix [A B] = 1 2 3 M 14 .
1 4 7 M 30
We shall reduce the augmented matrix [A B] to Echelon form by applying elementary
row transformations only. Applying R2 → R2 − R1, R3 → R3 − R1, we get
1 1 1M 6 1 1 1M 6
[A B] ~ 0 1 2M 8 ~ 0 1 2M 8, by R3 → R3 − 3 R2 .
0 3 6M 24 0 0 0 M 0
Above is the Echelon form of the matrix [A B]. We have rank [A B] = the number of
non-zero rows in this Echelon form = 2 .
1 1 1
By the same elementary transformations, we get A ~ 0 1 2.
0 0 0
Obviously rank A = 2. Since rank A = rank [A B], therefore the given equations are
consistent. Here the number of unknowns is 3. Since rank A is less than the number of
unknowns therefore the given system will have an infinite number of solutions. We
see that the given system of equations is equivalent to the matrix equation
1 1 1 x 6
0 1 2 y = 8.
0 0 0 z 0
This matrix equation is equivalent to the system of equations
x + y + z = 6
.
y + 2z = 8
∴ y = 8 − 2 z , x = 6 − y − z = 6 − (8 − 2 z ) − z = z − 2.
Taking z = c , we see that x = c − 2, y = 8 − 2 c , z = c constitute the general solution of
the given system, where c is an arbitrary constant.
[Link]
M-113
Example 12: Apply the test of rank to examine if the following equations are consistent :
2 x − y + 3 z = 8, − x + 2 y + z = 4, 3 x + y − 4 z = 0
and if consistent, find the complete solution. (Meerut 2007)
Solution : The given system of equations is equivalent to the single matrix equation
2 −1 3 x 8
AX = −1 2 1 y = 4 = B.
3 1 −4 z 0
The augmented matrix
2 −1 3 M 8
[A B] = −1 2 1M 4.
3 1 −4 M 0
We shall reduce the augmented matrix to Echelon form by applying elementary row
transformations only. Applying R1 ↔ R2 , we get
−1 2 1M 4
[A B] ~ 2 −1 3 M 8
3 1 −4 M 0
−1 2 1M 4
~ 0 3 5 M 16, by R2 → R2 + 2 R1, R3 → R3 + 3 R1
0 7 −1 M 12
−1 2 1 M 4
0 3 5 M 16, by R3 → 3 R3
0 21 −3 M 36
−1 2 1 M 4
~ 0 3 5 M 16, by R3 → R3 − 7 R2
0 0 −38 M −76
−1 2 1 M 4
1
~ 0 3 5 M 16, by R3 → − R3 .
0 38
0 1 M 2
[Link]
M-114
Above is the Echelon form of the matrix [A B]. We have the rank [A B] = the number
of non-zero rows in this Echelon form = 3.
−1 2 1
By the same transformations, we get A ~ 0 3 5 ⋅
0 0 1
Obviously rank A = 3. Since rank A = rank [A B], therefore the given equations are
[Link] the number of unknowns is 3. Since rank A is equal to the number of
unknowns, therefore the given equations have a unique solution. We see that the
given equations are equivalent to the matrix equation
−1 2 1 x 4
0 3 5 y = 16.
0 0 1 z 2
This matrix equation is equivalent to the equations
− x + 2 y + z = 4, 3 y + 5 z = 16, z = 2.
These give z = 2, y = 2, x = 2.
Example 13: Show that the equations
x + 2 y − z = 3, 3 x − y + 2 z = 1, 2 x − 2 y + 3 z = 2, x − y + z = − 1
are consistent and solve them. (Meerut 2006B, 09; Rohilkhand 06)
Solution: The given system of equations is equivalent to the single matrix equation
1 2 −1 3
x
3 −1 2 1
AX = y = = B.
2 −2 3 2
z
1 −1 1 −1
The augmented matrix
1 2 −1 &: 3
3 −1 2 &: 1
[A B] = .
2 −2 3 &: 2
1 −1 1 &: −1
Performing R2 → R2 − 3 R1, R3 → R3 − 2 R1, R4 → R4 − R1, we get
[Link]
M-115
1 2 −1 M 3
0 −7 5 M −8
[A B] ~
0 −6 5 M −4
0 −3 2 M −4
1 2 −1 M 3
0 −1 0 M −4
~ , by R2 → R2 − R3
0 −6 5 M −4
0 −3 2 M −4
1 2 −1 M 3
0 −1 0 M −4
~ , by R3 → R3 − 6 R2 , R4 → R4 − 3 R2
0 0 5 M 20
0 0 2 M 8
1 2 −1 M 3
0 −1 0 M −4 1 1
~ , by R3 → R3 , R4 → R4
0 0 1 M 4 5 2
0 0 1 M 4
1 2 −1 M 3
0 −1 0 M −4
~ , by R4 → R4 − R3 .
0 0 1 M 4
0 0 0 M 0
Thus the matrix [A B] has been reduced to Echelon form. We have rank [A B] = the
number of non-zero rows in this Echelon form = 3. Also
1 2 −1
0 −1 0
A ~ .
0 0 1
0 0 0
We have rank A = 3. Since rank [A B] = rank A, therefore the given equations are
consistent. Since rank A = 3 = the number of unknowns, therefore the given equations
have unique solution. The given equations are equivalent to the equations
x + 2 y − z = 3, − y = − 4, z = 4.
These give z = 4, y = 4, x = − 1.
[Link]
M-116
Example 14: State the conditions under which a system of non-homogeneous equations will have
(i) no solution (ii) a unique solution (iii) infinity of solutions. (Lucknow 2008)
Solution : Let AX = B be a system of linear non-homogeneous equations, where A, X, B
are m × n, n × 1, m × 1 matrices respectively.
(i) These equations will have no solution if the coefficient matrix A and the
augmented matrix [A B] are not of the same rank.
(ii) These equations will possess a unique solution if the matrices A and [A B] are of
the same rank and the rank is equal to the number of variables. In particular if A is
a square matrix, these equations will possess a unique solution if and only if the
matrix A is non-singular.
(iii) These equations will have infinity of solutions if the matrices A and [A B] are of
the same rank and the rank is less than the number of variables.
Example 15: By the use of matrices, solve the equations :
x + y + z = 9, 2 x + 5 y + 7 z = 52, 2 x + y − z = 0 .
(Bundelkhand 2011; Avadh 10; Rohilkhand 10)
Solution : The given system of equations is equivalent to the single matrix equation
1 1 1 x 9
AX = 2 5 7 y = 52 = B.
2 1 −1 z 0
The augmented matrix
1 1 1 M 9
[A B] = 2 5 7 M 52
2 1 −1 M 0
1 1 1 M 9
~ 0 3 5 M 34, by R2 → R2 − 2 R1, R3 → R3 − 2 R1
0 −1 −3 M −18
1 1 1 M 9
~ 0 −1 −3 M −18 , by R2 ←→ R3
0 3 5 M 34
[Link]
M-117
1 1 1 M 9
~ 0 −1 −3 M −18 , by R3 → R3 + 3 R2 .
0 0 −4 M −20
Above is the Echelon form of the matrix [A B]. We have rank [A B] = the number of
non-zero rows in this Echelon form = 3.
Also by the same E-row transformations, we get
1 1 1
A ~ 0 −1 −3 .
0 0 −4
∴ rank A = 3.
Since rank A = rank [A B], therefore the given equations are consistent. Also rank
A = 3 and the number of unknowns is also 3. Hence the given equations will have a
unique solution. To find the solution we see that the given system of equations is
equivalent to the matric equation
1 1 1 x 9
0 −1 −3 y = −18.
0 0 −4 z −20
This matrix equation is equivalent to the system of equations
x + y + z = 9, − y − 3 z = − 18, − 4 z = − 20 .
Solving these, we get z = 5, y = 3, x = 1.
Example 16: Investigate for what values of λ , µ the simultaneous equations
x + y + z = 6, x + 2 y + 3 z = 10 , x + 2 y + λz = µ
have (i) no solution, (ii) a unique solution, (iii) an infinite number of solutions.
(Meerut 2006, 09B; Bundelkhand 09; Rohilkhand 07; Kanpur 09)
Solution : The matrix form of the given system of equations is
1 1 1 x 6
AX = 1 2 3 y = 10 = B .
1 2 λ z µ
[Link]
M-118
1 1 1 M 6
The augmented matrix [A B] = 1 2 3 M 10
1 2 λ M µ
1 1 1 M 6
~ 0 1 2 M 4 by R2 → R2 − R1, R3 → R3 − R1
0 1 λ −1 M µ − 6
1 1 1 M 6
~ 0 1 2 M 4 by R3 → R3 − R2 .
0 0 λ −3 M µ − 10
If λ ≠ 3, we have rank [A B] = 3 = rank A.
So in this case the given system of equations is consistent. Since rank A = the number
of unknowns, therefore the given system of equations possesses a unique solution.
Thus if λ ≠ 3,the given system of equations possesses a unique solution for any value of
µ.
If λ = 3 and µ ≠ 10, we have rank [A B] = 3 and rank A = 2. Thus in this case rank
[A B] ≠ rank A and so the given system of equations is inconsistent i. e., possesses no
solution.
If λ = 3 and µ = 10, we have rank [A B] = 2 = rank A.
So in this case the given system of equations is again consistent. Since rank A < the
number of unknowns, therefore in this case the given system of equations possesses an
infinite number of solutions.
Example 17: For what values of η the equations
x + y + z = 1, x + 2 y + 4 z = η, x + 4 y + 10 z = η2 ,
have a solution and solve them completely in each case.
(Meerut 2011; Agra 07; Kanpur 06, 08, 10)
Solution : The matrix form of the given system is
1 1 1 x 1
1 2 4 y = η.
1 4 10 z η2
[Link]
M-119
Performing R2 → R2 − R1, R3 → R3 − R1, we get
1 1 1 x 1
0 1 3 y = η − 1 .
0 3 9 z η2 − 1
Performing R3 → R3 − 3 R2 , we have
1 1 1 x 1
0 1 3 y = η−1 …(1)
0 0 0 z η2 − 3 η + 2
Now the given equations will be consistent if and only if
η2 − 3 η + 2 = 0 , i. e., iff (η − 2) (η − 1) = 0 , i. e., iff η = 2 or η = 1.
Case I. If η = 2, the equation (1) becomes
1 1 1 x 1
0 1 3 y = 1 .
0 0 z
0 0
The above system of equations is equivalent to
y + 3 z = 1, x + y + z = 1.
∴ y = 1 − 3z , x = 2z .
Thus, x = 2 k , y = 1 − 3 k , z = k constitute the general solution where k is an arbitrary
constant.
Case II. If η = 1, the equation (1) becomes
1 1 1 x 1
0 1 3 y = 0 .
0 0 0 z 0
The above system of equations is equivalent to
y + 3 z = 0 , x + y + z = 1.
∴ y = − 3z , x = 1 + 2z .
Thus x = 1 + 2 c , y = − 3 c , z = c constitute the general solution, where c is an arbitrary
constant.
[Link]
M-120
Comprehensive Exercise 2
1. Use the test of rank to show that the following equations are not consistent :
2 x − y + z = 4, 3 x − y + z = 6, 4 x − y + 2 z = 7, − x + y − z = 9.
(Rohilkhand 2009)
2. Show that the equations −2 x + y + z = a, x − 2 y + z = b, x + y − 2 z = c have no
solution unless a + b + c ≠ 0 (Kanpur 2009)
3. Apply the test of rank to examine if the following system of equations is
consistent and if consistent, find the complete solution :
x + y + z = 6, x + 2 y + 3 z = 10 , x + 2 y + 4 z = 1.
4. Use matrix method to solve the equations :
2 x − y + 3 z = 9, x + y + z = 6, x − y + z = 2. (Meerut 2005; Avadh 06,11)
5. Solve completely the equations
x + 2 y + 3 z = 6, 2 x + 4 y + z = 7, 3 x + 2 y + 9 z = 4.
6. Show that the equations x + 2 y − 5 z = − 9, 3 x − y + 2 z = 5, 2 x + 3 y − z = 3,
4 x − 5 y + z = − 3 are consistent and solve the same.
7. Show that the equations
x − 3 y − 8 z + 10 = 0 , 3 x + y − 4 z = 0 , 2 x + 5 y + 6 z − 13 = 0
are consistent and solve the same. (Bundelkhand 2007)
8. Solve completely the equations 2 x + 3 y + z = 9, x + 2 y + 3 z = 6, 3 x + y + 2 z = 8.
9. Solve completely the equations
x + y + z = 1, x + 2 y + 3 z = 4, x + 3 y + 5 z = 7, x + 4 y + 7 z = 10 . (Lucknow 2009)
10. Express the following system of equations into the matrix equation form AX = B:
x + 2 y + z = − 1, 6 x + y + z = − 4, 2 x − 3 y − z = 0 , − x − 7 y − 2 z = 7, x − y = 1.
Determine if this system of equations is consistent and if so find its solution.
11. Examine if the system of equations :
x + y + 4 z = 6, 3 x + 2 y − 2 z = 9, 5 x + y + 2 z = 13
is consistent. Find also the solution if it is consistent.
12. Prove, without actually solving, that the following system of equations has a
unique solution :
5 x + 3 y + 14 z = 4, y + 2 z = 1, x − y + 2 z = 0 .
13. For what values of the parameter λ will the following equations fail to have a
unique solution
3 x − y + λz = 1, 2 x + y + z = 2, x + 2 y − λz = − 1?
Will the equations have any solutions for these values of λ?
[Link]
M-121
14. Solve the equations
λx + 2 y − 2 z − 1 = 0 , 4 x + 2 λy − z − 2 = 0 , 6 x + 6 y + λz − 3 = 0 ,
considering specially the case when λ = 2.
15. Discuss for all values of λ, the system of equations
x + y + 4 z = 6, x + 2 y − 2 z = 6, λx + y + z = 6,
as regards existence and nature of solutions.
16. Solve the following equations by matrix method :
5 x + 3 y + 7 z = 4, 3 x + 26 y + 2 z = 9, 7 x + 2 y + 10 z = 5.
17. Solve the following system of linear equations by matrix method :
5 x − 6 y + 4 z = 15, 7 x + 4 y − 3 z = 19, 2 x + y + 6 z = 46.
18. Solve the following equations by matrix method :
x − y + 2 z = 4, 3 x + y + 4 z = 6, x + y + z = 1.
19. Solve the following equations by matrix method :
x − 2 y + 3 z = 6, 3 x + y − 4 z = − 7, 5 x − 3 y + 2 z = 5. (Meerut 2010)
A nswers 2
3. x = − 7, y = 22, z = − 9
4. x = 1, y = 2, z = 3
5. x = 1, y = 1, z = 1
1 3 5
6. x= , y = ,z =
2 2 2
7. x = − 1 + 2 c , y = 3 − 2 c , z = c , where c is arbitrary
35 29 5
8. x= , y= ,z =
18 18 18
9. x = c − 2, y = 3 − 2 c , z = c
10. Consistent; x = − 1, y = − 2, z = 4
1
11. Consistent ; x = 2, y = 2, z =
2
7 7
13. λ ≠ , the solution is unique; λ = − , no solution
2 2
14. In case λ = 2, the general solution of the given system of equations is given by
1
x = − c, y = c, z = 0
2
[Link]
M-122
7 7
15. λ ≠ , the solution is unique; λ = − , no solution
10 10
7 16 3 1
16. x = − c, y = + c, z = c
11 11 11 11
17. x = 3, y = 4, z = 6
5 3 3 1
18. x = − c, y = − + c, z = c
2 2 2 2
−8 5 25 13
19. x = + c, y = − + c, z = c
7 7 7 7
Objective Type Questions
Multiple Choice Questions
Indicate the correct answer for each question by writing the corresponding letter from (a),
(b), (c) and (d).
1. A set of r n-vectors X1, X2 , ........., X r is said to be linearly independent if every
relation of the type k1 X1 + k2 X2 + ......... + k r X r = O implies
(a) k1 + k2 + ......... + k r = 0
(b) k1 = k2 = ......... = k r = 0
(c) k1 + k2 + ......... + k r = 1
(d) none of these
2. Consider the system of equations
a1 x + b1 y + c1z = 0 , a2 x + b2 y + c2 z = 0 , a3 x + b3 y + c3 z = 0 .
a1 b1 c1
If a b2 c2 = 0 , then the system has
2
a3 b3 c3
(a) more than two solutions
(b) only trivial solution
(c) no solution
(d) none of these
3. The system of linear equations x + y + z = 2, 2 x + y − z = 3, 3 x + 2 y + kz = 4 has
a unique solution if
(a) k ≠ 0
[Link]
M-123
(b) −1 < k < 1
(c) −2 < k < 2
(d) k = 0
4. The system of equations x + 2 y + 3 z = 1, 2 x + y + 3 z = 2, 5 x + 5 y + 9 z = 4 has
(a) only one solution
(b) infinitely many solutions
(c) no solution
(d) none of these
(Meerut 2003)
Fill in the Blank(s)
Fill in the blanks ‘’.........,’’ so that the following statements are complete and correct.
1. The set consisting only of the zero vector, O, is linearly ......... .
2. The set of three vectors X1 = (1, 0 , 0 ), X2 = (0 , 1, 0 ), X 3 = (0 , 0 , 1) is linearly ........
3. The number of linearly independent solutions of the system of m homogeneous
linear equations in n variables, AX = O, is ......., where r is the rank of the matrix A.
4. When the system of equations has one or more solutions, the equations are said
to be ....., otherwise they are said to be .......
5. The sytem of m non-homogeneous linear equations in n unknowns, AX = B, is
....... if and only if the coefficient matrix A and the ungmented matrix [A B] are
of the same rank. (Meerut 2001)
6. If A be an n-rowed non-singular matrix, X be an n × 1matrix, B be an n × 1matrix,
the system of equations A X = B, has a ...... solution.
7. If the number of equations is less than the number of unknowns, the equations
will aways have an ....... number of solutions, provided they are consistent.
True or False
Write ‘T’ for true and ‘F’ for false statement.
1. A set of r n-vectors X1, X2 , ......., X r , is said to be linearly dependent if there exist r
scalars k1, k2 , ........., k r , not all zero, such that k1X1 + k2 X2 + ......... + k r X r = O,
where O denotes the n-vector whose components are all zero.
2. If a set of vectors is linearly independent, then at least one member of the set can
be expressed as a linear combination of the remaining members.
[Link]
M-124
3. The zero solution i. e., x = y = z = 0 is the only solution of the system of equations
x + 2 y + 3 z = 0, 3 x + 4 y + 4 z = 0 , 7 x + 10 y + 12 z = 0 .
4. Let AX = B be a system of m non-homogeneous linear equations in n unknowns.
These equations will have no solution if the coeffcient matrix A and the
augmented matrix [A B] are not of the same rank.
5. The system of m non-homogeneous linear equations in n unknowns, AX = B will
possess a unique solution if the matrices A and [A B] are of the same rank and the
rank is less than the number of variables.
A nswers
Multiple Choice Questions
1. (b) 2. (a) 3. (a) 4. (a)
Fill in the Blank(s)
1. dependent 2. independent
3. (n − r) 4. consistent; inconsistent
5. consistent 6. unique
7. infinite
True or False
1. T 2. F 3. T
4. T 5. F
o
[Link]
M- 125
4
Eigenvalues and
Eigenvectors
4.1 Matric Polynomials
Definition: An expression of the form
F (λ ) = A0 + A1λ + A2 λ2 + … + A m − 1λ m − 1 + A m λ m ,
where A0 , A1, A2 , ..., A m are all square matrices of the same order, is called a Matric
polynomial of degree m provided A m is not a null matrix. The symbol λ is called
indeterminate. If the order of each of the matric coefficients A0 , A1, …, A m is n, then we
say that the matric polynomial is n-rowed. According to this definition of a matric
polynomial, each square matrix can be expressed as a matric polynomial with zero
degree. For example, if A be any square matrix, we can write A = λ0 A.
Equality of Polynomials : Two matric polynomials are equal iff (if and only if), the
coefficients of the like powers of λ are the same.
Theorem: Every square matrix whose elements are ordinary polynomials in λ, can essentially be
expressed as a matric polynomial in λ of degree m, where m is the highest power of λ occurring in
any element of the matrix. We shall illustrate this theorem by the following example.
Consider the matrix
[Link]
M-126
1 + 2 λ + 3 λ2 λ2 4 − 6λ
A = 1 + λ3 3 + 4λ 2
1 − 2λ + 4λ 3
3
2 − 3 λ + 2 λ 5 6
in which the highest power of λ occurring in any element is 3. Rewriting each element
as a cubic in λ, supplying missing coefficients with zeros, we get
1 + 2 . λ + 3 . λ2 + 0 . λ3 0 + 0 . λ + 1 . λ2 + 0 . λ3 4 − 6 . λ + 0 . λ2 + 0 . λ3
A = 1 + 0 . λ + 0 . λ2 + 1 . λ3 3 + 0 . λ + 4. λ2 + 0 . λ3 1 − 2 . λ + 0 . λ2 + 4 . λ3
2 3
2 − 3 . λ + 0 . λ + 2 . λ 5 + 0 . λ + 0 . λ2 + 0 . λ3 6 + 0 . λ + 0 . λ2 + 0 . λ3
Obviously A can be written as the matric polynomial
1 0 4 2 0 −6 3 1 0 0 0 0
2 3
A = 1 3
1 + λ 0 0 −2 + λ 0 4 0 + λ 1 0 4.
2 5 6 −3 0 0 0 0 0 2 0 0
4.2 Characteristic Values and Characteristic
Vectors of a Matrix
Let A = [aij]n × n be a given n-rowed square matrix. Let
x1
x
2
A = …
…
xn
be a column vector. Consider the vector equation
AX = λX. ..(1)
where λ is a scalar (i.e., number).
It is obvious that the zero vector X = O is a solution of (1) for any value of λ. Now let us
see whether there exist scalars λ and non-zero vectors X which satisfy (1).
If I denotes the unit matrix of order n, then the equation (1) may be written as
AX = λIX
or (A − λI) X = O. ...(2)
[Link]
M- 127
The matrix equation (2) represents the following system of n homogeneous equations
in n unknowns :
(a11 − λ ) x1 + a12 x2 + … + a1n xn = 0
a21 x1 + (a22 − λ ) x2 + … + a2 n xn = 0
...(3)
… … … … … … … …
an1 x1 + an2 x2 + … + (ann − λ ) xn = 0
The coefficient matrix of the equations (3) is A − λI. The necessary and sufficient
condition for equations (3) to possess a non-zero solution (X ≠ O) is that the
coefficient matrix A − λI should be of rank less than the number of unknowns n. But
this will be so if and only if the matrix A − λI is singular i. e., if and only if| A − λI| = 0.
Thus the scalars λ for which
| A − λI| = 0
are of special importance.
Definitions: (Lucknow 2009)
Let A = [aij]n × n be any n-rowed square matrix and λ an indeterminate. The matrix
A − λI is called the characteristic matrix of A where I is the unit matrix of order n.
Also the determinant
a11 − λ a12 ... a1n
a21 a22 − λ ... a2 n
| A − λI| = ,
... ... ... ...
an1 an2 ... ann − λ
which is an ordinary polynomial in λ of degree n, is called the characteristic
polynomial of A. The equation| A − λI| = 0 is called the characteristic equation of A
and the roots of this equation are called the characteristic roots or characteristic
values or eigenvalues or latent roots or proper values of the matrix A. The set of the
eigenvalues of A is called the spectrum of A.
If λ is a characteristic root of the matrix A, then
| A − λI| = 0
and the matrix A − λI is singular. Therefore there exists a non-zero vector X such that
(A − λI) X = O or AX = λX.
[Link]
M-128
Characteristic vectors. Definition: If λ is a characteristic root of an n × n matrix A, then a
non-zero vector X such that AX = λX is called a characteristic vector or eigenvector of A
corresponding to the characteristic root λ. (Lucknow 2009)
4.3 Certain Relations between Characteristic Roots and
Characteristic Vectors
Theorem 1. λ is a characteristic root of a matrix A if and only if there exists a non-zero vector X
such that AX = λX.
Proof. Suppose λ is a characteristic root of the matrix A. Then | A − λI| = 0 and the
matrix A − λI is singular. Therefore, the matrix equation (A − λI) X = O possesses a
non-zero solution i. e., there exists a non-zero vector X such that (A − λI)X = O or
AX = λX.
Conversely suppose there exists a non-zero vector X such that AX = λX
i. e., (A − λI) X = O. Since the matrix equation (A − λI) X = O possesses a non-zero
solution, therefore the coefficient matrix A − λI must be singular i. e.,| A − λI | = 0.
Hence λ is a characteristic root of the matrix A.
Theorem 2. If X is a characteristic vector of a matrix A corresponding to the characteristic value
λ, then kX is also a characteristic vector of A corresponding to the same characteristic value λ.
Here k is any non-zero scalar.
Proof. Suppose X is a characteristic vector of A corresponding to the characteristic
value λ. Then X ≠ O and AX = λX.
If k is any non-zero scalar, then k x ≠ O. Also
A ( kX) = k (AX) = k (λX) = λ (kX).
Now kX is a non-zero vector such that A(kX) = λ (kX). Hence kX is a characteristic
vector of A corresponding to the characteristic value λ. Thus corresponding to a
characteristic value λ, there corresponds more than one characteristic vectors.
Theorem 3: If X is a characteristic vector of a matrix A, then X cannot correspond to more than
one characteristic values of A.
Proof: Let X be a characteristic vector of a matrix A corresponding to two
characteristic values λ1 and λ 2 . Then AX = λ1X and AX = λ 2 X. Therefore
λ1 X = λ 2 X
⇒ (λ1 − λ 2 ) X = O
[Link]
M- 129
⇒ λ1 − λ 2 = 0 [∵ X ≠ O]
⇒ λ1 = λ 2 .
Hence the result.
4.4 Nature of the Characteristic Roots of Special Types of
Matrices
Theorem 1: The characteristic roots of a Hermitian matrix are real.
(Meerut 2011; Lucknow 05, 11)
Proof. Suppose A is a Hermitian matrix, λ a characteristic root of A and X a
corresponding eigenvector. Then
AX = λX. ...(1)
Premultiplying both sides of (1) by X θ , we get
X θ AX = λX θ X. ...(2)
Taking conjugate transpose of both sides of (2), we get
(X θ AX)θ = (λX θ X)θ
or X θ A θ (X θ )θ = λX θ (X θ )θ
or X θ AX = λX θ X ...(3)
[∵ (X θ )θ = X and A θ = A, A being Hermitian]
From (2) and (3), we have
λX θ X = λX θ X
or (λ − λ )X θ X = O.
But X is not a zero vector, therefore X θ X ≠ O.
Hence λ − λ = 0, so that λ = λ and consequently λ is real.
Corollary 1. The characteristic roots of a real symmetric matrix are all real.
[Link]
M-130
If the elements of a Hermitian matrix A are all real, then A is a real symmetric matrix.
Thus a real symmetric matrix is Hermitian and therefore the result follows.
Corollary 2. The characteristic roots of a skew-Hermitian matrix are either pure imaginary or
zero. (Lucknow 2010)
Suppose A is a skew-Hermitian matrix. Then iA is Hermitian. Let λ be a characteristic
root of A. Then
AX = λX
or (iA)X = (iλ )X.
From this it follows that iλ is a characteristic root of iA which is Hermitian. Hence iλ is
real. Therefore either λ must be zero or pure imaginary.
Corollary 3. The characteristic roots of a real skew-symmetric matrix are either pure imaginary
or zero, for every such matrix is skew-Hermitian.
Theorem 2. The characteristic roots of a unitary matrix are of unit modulus.
Proof. Suppose A is a unitary matrix. Then A θ A = I.
Let λ be a characteristic root of A and X a corresponding eigenvector. Then
AX = λX. ...(1)
Taking conjugate transpose of both sides of (1), we get
(AX) θ = (λX)θ
or X θ A θ = λX θ . ...(2)
From (1) and (2), we have
( X θ A θ )(AX) = λλX θ X
or X θ (A θ A) X = λλX θ X [∵ A θ A = I]
or X θIX = λλX θ X
or X θ X = λλX θ X
or X θ X(λλ − 1) = O. ...(3)
Since X θ X ≠ O, therefore, (3) gives
λλ − 1 = 0
[Link]
M- 131
or λ λ =1
or | λ |2 = 1.
Corollary. The characteristic roots of an orthogonal matrix are of unit modulus.
We know that if the elements of a unitary matrix A are all real, then A is said to be an
orthogonal matrix. Hence the result follows.
4.5 The Process of Finding the Eigenvalues and Eigenvectors
of a Matrix
Let A = [aij]n × n be a square matrix of order n. First we should write the characteristic
equation of the matrix A i. e., the equation| A − λI| = 0. This equation will be of degree n
in λ. So it will have n roots. These n roots will give us the eigenvalues of the matrix A. If
λ1 is an eigenvalue of A, then the corresponding eigenvectors of A will be given by the
non-zero vectors
X = [ x1, x2 , ..., xn]′
satisfying the equation AX = λ1X
or (A − λ1I) X = O.
Example 1. Determine the characteristic roots of the matrix
0 1 2
A = 1 0 −1.
2 −1 0 (Meerut 2010)
Solution : The characteristic matrix of A
= A − λI
0 1 2 1 0 0
= 1 0 −1 − λ 0 1 0
2 −1 0 0 0 1
[Link]
M-132
0 − λ 1 2
= 1 0−λ −1 .
2 −1 0 − λ
It should be noted that in order to obtain the characteristic matrix of a matrix A, we
should simply subtract λ from each of its principal diagonal elements.
The characteristic polynomial of A
= | A − λI|
− λ 1 2
= 1 −λ −1
2 −1 −λ
= − λ (λ2 − 1) − 1(− λ + 2) + 2(−1 + 2 λ )
= − λ3 + λ + λ − 2 − 2 + 4 λ
= − λ3 + 6 λ − 4.
∴ the characteristic equation of A is
| A − λI| = 0
i. e., λ3 − 6 λ + 4 = 0
i. e., (λ − 2)(λ2 + 2 λ − 2) = 0 .
The roots of this equation are λ = 2, − 1 ± 3 .
Hence the characteristic roots of the matrix A are 2, − 1 ± 3 .
Example 2: Determine the eigenvalues of the matrix
a h g
A = 0 b 0 .
0 c c (Meerut 2010B; Kanpur 09)
a − λ h g
Solution: Here | A − λI| = 0 b−λ 0
0 c c − λ
= (a − λ )(b − λ ) (c − λ ) .
[Link]
M- 133
The characteristic equation of A is
| A − λI| = 0
i. e., (a − λ )(b − λ )(c − λ ) = 0.
The roots of this equation are λ = a, b, c . Hence the eigenvalues of A are a, b, c .
Example 3: Determine the eigenvalues and eigenvectors of the matrix
5 4
A= .
1 2 (Bundelkhand 2006)
Solution: The characteristic equation of A is
| A − λI| = 0
5 − λ 4
i. e., = 0
1 2 − λ
i. e., (5 − λ ) (2 − λ ) − 4 = 0
i. e., λ2 − 7 λ + 6 = 0 .
The roots of this equation are λ1 = 6, λ 2 = 1. Therefore the eigenvalues of A are 6, 1.
x1
The eigenvectors X = of A corresponding to the eigenvalue 6 are given by the
x2
non-zero solutions of the equation
(A − 6 I) X = O
5 − 6 4 x1 0
or 1 =
2 − 6 x2 0
−1 4 x1 0
or =
1 −4 x2 0
−1 4 x1 0
or 0 0 x = 0 , applying R2 → R2 + R1.
2
The coefficient matrix of these equations is of rank 1. Therefore these equations have
2 − 1 i. e., 1 linearly independent solution. These equations reduce to the single
equation − x1 + 4 x2 = 0. Obviously x1 = 4, x2 = 1 is a solution of this equation.
4
Therefore X1 = is an eigenvector of A corresponding to the eigenvalue 6. The set of
1
[Link]
M-134
all eigenvectors of A corresponding to the eigenvalue 6 is given by c1X1 where c1 is any
non-zero scalar.
The eigenvectors X of A corresponding to the eigenvalue 1 are given by the non-zero
solutions of the equation
(A − 1I) X = O
4 4 x1 0
or 1 1 x = 0
2
or 4 x1 + 4 x2 = 0 , x1 + x2 = 0 .
1
From these x1 = − x2 . Let us take x1 = 1, x2 = − 1. Then X2 = is an eigenvector of A
−1
corresponding to the eigenvalue 1. Every non-zero multiple of the vector X2 is an
eigenvector of A corresponding to the eigenvalue 1.
Example 4: Determine the characteristic roots and the corresponding characteristic vectors of the
matrix
8 −6 2
A = −6 7 −4.
2 −4 3
(Purvanchal 2010; Bundelkhand 08; Rohilkhand 05; Agra 07; Kanpur 09; Avadh 05)
Solution : The characteristic equation of the matrix A is
| A − λI| = 0
8 − λ −6 2
i. e., −6 7−λ −4 = 0
2 −4 3 − λ
or (8 − λ ) {(7 − λ )(3 − λ ) − 16} + 6 { −6 (3 − λ ) + 8} + 2 {24 − 2(7 − λ ) } = 0
or λ3 − 18 λ2 + 45 λ = 0
or λ (λ − 3) (λ − 15) = 0.
Hence the characteristic roots of A are 0, 3, 15.
The eigenvectors X = [ x1, x2 , x3 ]′ of A corresponding to the eigenvalue 0 are given by
the non-zero solutions of the equation
(A − 0I)X = O
[Link]
M- 135
8 −6 2 x1 0
or −6 7 −4 x = 0
2
2 −4 3 x3 0
2 −4 3 x1 0
or −6 7 −4 x = 0 , by R ←→ R
2 1 3
8 −6 2 x3 0
2 −4 3 x1 0
or 0 −5 5 x = 0 , by R → R + 3 R , R → R − 4 R
2 2 2 1 3 3 1
0 10 −10 x3 0
2 −4 3 x1 0
or 0 −5 5 x = 0 , by R → R + 2 R .
2 3 3 2
0 0 0 x3 0
The coefficient matrix of these equations is of rank 2. Therefore these equations have
3 − 2 = 1 linearly independent solution. Thus there is only one linearly independent
eigenvector corresponding to the eigenvalue 0. These equations can be written as
2 x1 − 4 x2 + 3 x3 = 0 , − 5 x2 + 5 x3 = 0 .
From the last equation, we get x2 = x3 . Let us take x2 = 1, x3 = 1. Then the first equation
1 1
gives x1 = . Therefore X1 = 1 1 ′ is an eigenvector of A corresponding to the
2 2
eigenvalue 0. If c1 is any non-zero scalar, then c1X1 is also an eigenvector of A
corresponding to the eigenvalue 0.
The eigenvectors of A corresponding to the eigenvalue 3 are given by the non-zero
solutions of the equation
(A − 3 I)X = O
5 −6 2 x1 0
or −6 4 −4 x = 0
2
2 −4 0 x3 0
−1 −2 −2 x1 0
or −6 4 −4 x = 0 , by R → R + R
2 1 1 2
2 −4 0 x3 0
[Link]
M-136
−1 −2 −2 x1 0
or 0 16 8 x2 = 0 , by R2 → R2 − 6 R1, R3 → R3 + 2 R1
0 −8 −4 x3 0
−1 −2 −2 x1 0
1
or 0 16 8 x2 = 0 , by R3 → R3 + R2 .
0 2
0 0 x3 0
The coefficient matrix of these equations is of rank 2. Therefore these equations have
3 − 2 = 1 linearly independent solution. These equations can be written as
− x1 − 2 x2 − 2 x3 = 0 , 16 x2 + 8 x3 = 0 .
1
From the second equation we get x2 = − x3 . Let us take x3 = 4, x2 = − 2. Then the first
2
equation gives x1 = − 4. Therefore X2 = [−4 − 2 4]′
is an eigenvector of A corresponding to the eigenvalue 3. Every non-zero multiple of
the X2 is an eigenvector of A corresponding to the eigenvalue 3.
The eigenvectors of A corresponding to the eigenvalue 15 are given by the non-zero
solutions of the equation
(A − 15 I) X = O
8 − 15 −6 2 x1 0
or −6 7 − 15 −4 x = 0
2
2 −4 3 − 15 x3 0
−7 −6 2 x1 0
or −6 −8 −4 x = 0
2
2 −4 −12 x3 0
−1 2 6 x1 0
or −6 −8 −4 x = 0 , by R → R − R
2 1 1 2
2 −4 −12 x3 0
−1 2 6 x1 0
or 0 −20 −40 x = 0 , by R → R − 6 R , R → R + 2 R .
2 2 2 1 3 3 1
0 0 0 x3 0
The coefficient matrix of these equations is of rank 2. Therefore these equations have
3 − 2 = 1 linearly independent solution. These equations can be written as
[Link]
M- 137
− x1 + 2 x2 + 6 x3 = 0 , − 20 x2 − 40 x3 = 0 .
The last equation gives x2 = − 2 x3 . Let us take x3 = 1, x2 = − 2. Then the first equation
gives x1 = 2. Therefore
X3 = [2 − 2 1]′
is an eigenvector of A corresponding to the eigenvalue 15. If k is any non-zero scalar,
then kX3 is also an eigenvector of A corresponding to the eigenvalue 15.
Example 5: Determine the characteristic roots and the corresponding characteristic vectors of the
matrix
6 −2 2
A = −2 3 −1.
2 −1 3 (Meerut 2006B, 09; Purvanchal 07)
Solution : The characteristic equation of A is
|A − λI| = 0
6 − λ −2 2
or −2 3−λ −1 = 0
2 −1 3 − λ
6 − λ −2 0
or −2 3−λ 2 − λ = 0 , by C3 → C3 + C2
2 −1 2 − λ
6 − λ −2 0
or (2 − λ ) −2 3−λ 1 = 0
2 −1 1
6 − λ −2 0
or (2 − λ ) −4 4−λ 0 = 0 , by R2 → R2 − R3
2 −1 1
or (2 − λ ) [(6 − λ ) (4 − λ ) − 8] = 0
or (2 − λ ) (λ2 − 10 λ + 16) = 0
or (2 − λ ) (λ − 2) (λ − 8) = 0 .
Therefore the characteristic roots of A are given by λ = 2, 2, 8.
[Link]
M-138
The characteristic vectors of A corresponding to the characteristic root 8 are given by
the non-zero solutions of the equation
(A − 8 I) X = O
6 − 8 −2 2 x1 0
or −2 3 − 8 −1 x = 0
2
2 −1 3 − 8 x3 0
−2 −2 2 x1 0
or −2 −5 −1 x = 0
2
2 −1 −5 x3 0
−2 −2 2 x1 0
or 0 −3 −3 x = 0 , by R → R − R , R → R + R
2 2 2 1 3 3 1
0 −3 −3 x3 0
−2 −2 2 x1 0
or 0 −3 −3 x = 0 , by R → R − R .
2 3 3 2
0 0 0 x3 0
The coefficient matrix of these equations is of rank 2. Therefore these equations
possess 3 − 2 = 1 linearly independent solution. These equations can be written as
−2 x1 − 2 x2 + 2 x3 = 0 ,
−3 x2 − 3 x3 = 0 .
The last equation gives x2 = − x3 . Let us take x3 = 1, x2 = − 1. Then the first equation
2
gives x1 = 2. Therefore X1 = −1 is an eigenvector of A corresponding to the eigenvalue
1
8. Every non-zero multiple of X1 is an eigenvector of A corresponding to the
eigenvalue 8.
The eigenvectors of A corresponding to the eigenvalue 2 are given by the non-zero
solutions of the equation
(A − 2 I) X = O
4 −2 2 x1 0
or −2 1 −1 x = 0
2
2 −1 1 x3 0
[Link]
M- 139
−2 1 −1 x1 0
or 4 −2 2 x = 0 , by R ←→ R
2 1 2
2 −1 1 x3 0
−2 1 −1 x1 0
or 0 0 0 x = 0 , by R → R + 2 R , R → R + R .
2 2 2 1 3 3 1
0 0 0 x3 0
The coefficient matrix of these equations is of rank 1. Therefore these equations
possess 3 − 1 = 2 linearly independent solutions. We see that these equations reduce to
the single equation
−2 x1 + x2 − x3 = 0 .
−1 1
Obviously X2 = 0 , X3 = 2
2 0
are two linearly independent solutions of this equation. Therefore X2 and X3 are two
linearly independent eigenvectors of A corresponding to the eigenvalue 2. If c1, c2 are
scalars not both equal to zero, then c1X2 + c2 X3 gives all the eigenvectors of A
corresponding to the eigenvalue 2.
Example 6: Show that 0 is a characteristic root of a matrix if and only if the matrix is singular.
Solution: We have 0 is an eigenvalue of A ⇒ λ = 0 satisfies the equation
| A − λI| = 0 ⇒| A| = 0 ⇒ A is singular.
Conversely, A is singular ⇒| A| = 0
⇒ λ = 0 satisfies the equation | A − λI| = 0 ⇒ 0 is an eigenvalue of A.
Example 7: If λ1, ........., λ n are the eigenvalues of A, then show that kλ1, ..., kλ n are the
eigenvalues of kA.
Solution: If k = 0, then kA = O and each eigenvalue of O is 0. Thus 0 λ1, ...., 0 λ n are the
eigenvalues of kA if λ1, ..., λ n are the eigen-values of A.
So let us suppose that k ≠ 0.
We have | kA − λkI| = | k (A − λI)|
= k n | A − λI|. [∵ | kB| = k n | B|]
∴ if k ≠ 0, then | kA − λkI| = 0 if and only if | A − λI| = 0
[Link]
M-140
i. e., kλ is an eigenvalue of kA if and only if λ is an eigenvalue of A.
Thus kλ1, ...., kλ n are the eigenvalues of kA if λ1, ..., λ n are eigenvalues of A.
Example 8: If A is non-singular, prove that the eigenvalues of A −1 are the reciprocals of the
eigenvalues of A.
Solution: Let λ be an eigenvalue of A and X be a corresponding eigenvector. Then
AX = λX
⇒ X = A −1 (λX) = λ (A −1X)
1
⇒ X = A − 1X [∵ A is non-singular ⇒ λ ≠ 0]
λ
1
A − 1X = X
λ
1
⇒ is an eigenvalue of A −1 and X is a corresponding eigenvector.
λ
Conversely suppose that k is an eigenvalue of A −1. Since A is non-singular ⇒ A −1 is
non-singular and (A −1)−1 = A, therefore, it follows from the first part of this question
1
that is an eigenvalue of A. Thus each eigenvalue of A −1 is equal to the reciprocal of
k
some eigenvalue of A.
Hence the eigenvalues of A −1 are nothing but the reciprocals of the eigenvalues of A.
Example 9: If the characteristic roots of A are λ1, λ 2 ,… λ n , then the characteristic roots of
A2 are λ12, λ 22 , …, λ n2 . (Kumaun 2008)
Solution : Let λ be a characteristic root of the matrix A. Then there exists a non-zero
vector X such that
AX = λX
⇒ A (AX) = A (λX)
⇒ A2 X = λ (AX)
⇒ A 2 X = λ (λX) [∵ AX = λ X]
⇒ A2 X = λ2 X. …(1)
[Link]
M- 141
Since X is a non-zero vector, therefore from the relation (1) it is obvious that λ2 is a
characteristic root of the matrix A2 . Therefore if λ1, …, λ n are the characteristic roots
2 2
of A, then λ1 , … , λ n are the characteristic roots of A2 .
Example 10: The characteristic roots of an idempotent matrix are either zero or unity.
Solution: Let A be an idempotent matrix so that A2 = A. Let λ be a characteristic root
of the matrix A. Then there exists a non-zero vector X such that
AX = λX …(1)
⇒ A (AX) = A (λX)
⇒ A2 X = λ (AX)
⇒ AX = λ (λX) [∵ A2 = A and AX = λX]
⇒ AX = λ2 X. ...(2)
From (1) and (2), we get λ2 X = λX
or (λ2 − λ ) X = O
or λ2 − λ = 0 [∵ X ≠ O]
or λ (λ − 1) = 0 .
∴ λ = 0 or λ = 1.
Hence the characteristic roots of an idempotent matrix are either zero or unity.
Example 11: The product of the characteristic roots of a square matrix of order n is equal to the
determinant of the matrix.
Solution: Let A = [aij ]n × n be a square matrix of order n. Then the characteristic
polynomial f (λ ) of A is given by
a11 − λ a12 … a1n
a22 a22 − λ … a2 n
f (λ ) = | A − λI| =
… … … …
an1 an2 … ann − λ
= (− 1)n [λ n + p1 λ n − 1 + p2 λ n − 2 + …+ pn], say.
[Link]
M-142
If λ1, λ 2 , …, λ n are the characteristic roots of A, then λ1, λ 2 , …, λ n are the roots of the
equation | A − λI| = 0.
∴ | A − λI| = (− 1)n [λ n + p1λ n − 1 + p2 λ n− 2 + … + pn]
= (− 1)n (λ − λ1) (λ − λ 2 ) … (λ − λ n). …(1)
Putting λ = 0 on both sides of (1), we get
| A| = (− 1)n (− λ1) (− λ 2 ) …(− λ n)
= (− 1)n (− 1)n λ1 λ 2 … λ n = (− 1)2 n λ1 λ 2… λ n = λ1 λ 2 … λ n
Hence λ1 λ 2 … λ n = | A|.
Example 12: Any two characteristic vectors corresponding to two distinct characteristic roots of a:
(i) Hermitian, (ii) Real symmetric, (iii) Unitary matrix are orthogonal.
Solution : (i) A is Hermitian.
We have A θ = A.
Also, we have AX 1 = λ1 X 1 and AX2 = λ 2 X2
θ θ θ θ
or λ1 X1 (X2 )θ = λ 2 X1 X2 ; [λ1 is real ⇒ λ1 = λ1]
θ θ
or λ1X1 X2 = λ 2 X 1 X2
θ
or (λ1 − λ 2 ) X1 X2 = O.
θ
Since λ1 − λ 2 ≠ 0 , we have X1 X2 = O.
⇒ X1 and X2 are orthogonal with respect to each other.
(ii) A is real symmetric. The real symmetric matrix is always Hermitian, so the result
follows at once from (i).
(iii) A is unitary. We have A θ A = I.
Now, AX1 = λ1X1 and AX2 = λ 2 X2 ,
where λ1, λ 2 are characteristic roots of a unitary matrix which must be uni-modular
i. e., λ1λ1 = 1, λ 2 λ 2 = 1.
[Link]
M- 143
θ θ
Now, AX2 = λ 2 X2 ⇒ (AX2 )θ = (λ 2 X2 )θ ⇒ X2 A θ = λ 2 A2
θ θ θ
⇒ X2 A θ AX1 = λ 2 X2 λ1 X1 = λ 2 λ1 X2 X1
Again A θ A = I, hence we get
θ θ
X2 X1 = λ 2 λ1 X2 X1
θ
⇒ (1 − λ 2 λ1) X2 X1 = O.
But λ 2 λ1 ≠ 1 ; so 1 − λ 2 λ1 ≠ 0 .
θ
This implies that X2 X1 = O ⇒ X1 and X2 are orthogonal.
Example 13: Show that the two matrices A, C−1AC have the same characteristic roots.
(Lucknow 2005, 07)
Solution: Let B = C −1AC.
Then B − λI = C−1AC − λI
= C−1AC − C−1λIC [∵ C−1(λI) C = λC−1C = λI]
= C−1(A − λI) C.
∴ | B − λI| = |C−1|| A − λI||C|
= | A − λI|| C−1||C| = | A − λI||C−1C|
= | A − λI|| I| = | A − λI|.
Thus the two matrices A and B have the same characteristic determinants and hence
the same characteristic equations and the same characteristic roots.
4.6 The Cayley-Hamilton Theorem
Every square matrix satisfies its characteristic equation i.e., if for a square matrix A of order n,
| A − λI| = (−1)n [λ n + a1λ n − 1 + a2 λ n − 2 + … + an],
then the matrix equation X n + a1X n − 1 + a2 X n − 2 + a3 X n − 3 + … anI = O
[Link]
M-144
is satisfied by X = A i. e., A n + a1A n − 1 + a2 A n − 2 + … + anI = O.
(Meerut 2000, 03, 05B, 06B, 09B, 10B; Rohilkhand 05, 08; Agra 07; Avadh 05;
Purvanchal 08; Lucknow 08, 09)
Proof. Since the elements of A − λI are at most of the first degree in λ, the elements of
Adj (A − λI) are ordinary polynomials in λ of degree n − 1 or less. Therefore Adj
(A − λI) can be written as a matrix polynomial in λ, given by
Adj (A − λI) = B0 λ n − 1 + B1λ n − 2 + … + B n − 2 λ + B n − 1,
where B0 , B1, ..., B n − 1 are matrices of the type n × n whose elements are functions of
aij’s.
Now (A − λI) Adj. (A − λI) = | A − λI| I. [∵ A . Adj A = | A| In| ]
∴ (A − λI) (B0 λ n − 1 + B1λ n − 2 + … + B n − 2 λ + B n − 1)
= (−1)n [λ n + a1λ n − 1 + … + an] I.
Comparing coefficients of like powers of λ on both sides, we get
− IB0 = (−1)n I,
AB0 − IB1 = (−1)n a1I,
AB1 − IB2 = (−1)n a2 I,
...........................
...........................
AB n − 1 = (−1)n anI.
Pre-multiplying these successively by A n, A n −1, ..., I and adding, we get
O = (−1)n [A n + a1A n − 1 + a2 A n − 2 + … + anI].
Thus A n + a1A n − 1 + a2 A n − 2 + … + an − 1 A + anI = O. ...(1)
Corollary. 1. If A be a non-singular matrix, then | A| ≠ 0 . Also | A| = (−1)n an and
therefore an ≠ 0.
Pre-multiplying (1) by A −1, we get
[Link]
M- 145
A n − 1 + a1A n − 2 + a2 A n − 3 + … + an − 1 I + anA −1 = O
or A −1 = − (1 / an) [A n−1 + a1A n−2 + … + an−1 I].
Corollary. 2. If m be a positive integer such that m ≥ n, then multiplying the result (1)
by A m − n, we get
A m + a1A m − 1 + … + anA m − n = O,
showing that any positive integral power A m (m ≥ n) of A is linearly expressible in terms
of those of lower order.
2 −1 1
Example 14: Find the characteristic equation of the matrix A = −1 2 −1 and verify that
1 −1 2
it is satisfied by A and hence obtain A −1. (Meerut 2002, 05; Bundelkhand 05, 08, 10, 11;
Kanpur 10; Lucknow 05,11)
2 − λ −1 1
Solution : We have | A − λI| = −1 2 − λ −1
1 −1 2 − λ
= (2 − λ ){(2 − λ )2 − 1} + 1{ −1(2 − λ ) + 1} + 1{1 − (2 − λ )}
= (2 − λ ) (3 − 4 λ + λ2 ) + (λ − 1) + (λ − 1)
= − λ3 + 6 λ2 − 9 λ + 4.
∴ the characteristic equation of the matrix A is λ3 − 6 λ2 + 9 λ − 4 = 0 .
We are now to verify that A3 − 6 A2 + 9 A − 4 I = O. ...(1)
We have
1 0 0 2 −1 1 6 −5 5
2
I = 0 1 0 , A = −1 2 −1, A = A × A = −5 6 −5,
0 0 1 1 −1 2 5 −5 6
[Link]
M-146
22 −21 21
3 2
A = A A = −21 22 −21.
21 −21 22
Now we can verify that A3 − 6 A2 + 9A − 4I
22 −21 21 6 −5 5
= −21 22 −21 − 6 −5 6 −5
21 −21 22 6
5 −5
2 −1 1 1 0 0
+ 9 −1 2 −1 − 4 0 1 0
1 −1 2 0 0 1
0 0 0
= 0 0 0 .
0 0 0
Multiplying (1) by A −1, we get A2 − 6 A + 9 I − 4 A −1 = O.
1 2
∴ A −1 = (A − 6 A + 9 I)
4
6 −5 5 −12 6 −6 9 0 0
2
Now A − 6 A + 9 I = −5 6 −5 + 6 −12 6 + 0 9 0
5 −5 6 −6 6 −12 0 0 9
3 1 −1
= 1 3 1 .
−1 1 3
3 1 −1
−1 1
∴ A = 1 3 1.
4
−1 1 3
1 0 2
Example 15: Obtain the characteristic equation of the matrix A = 0 2 1 and verify that
2 0 3
it is satisfied by A and hence find its inverse. (Meerut 2001, 10B; Bundelkhand 09, 11;
Kanpur 07; Purvanchal 10; Lucknow 07, 10)
[Link]
M- 147
Solution : We have
1 − λ 0 2
| A − λI| = 0 2−λ 1
2 0 3 − λ
= (1 − λ )(2 − λ ) (3 − λ ) + 2 [0 − 2 (2 − λ )]
= (2 − λ ) [(1 − λ ) (3 − λ ) − 4]
= (2 − λ ) [λ2 − 4 λ − 1]
= − (λ3 − 6 λ2 + 7 λ + 2).
∴ the characteristic equation of A is
λ3 − 6 λ2 + 7 λ + 2 = 0 . …(1)
By the Cayley-Hamilton theorem
A3 − 6 A2 + 7 A + 2 I = O. ...(2)
Verification of (2). We have
1 0 2 1 0 2 5 0 8
2
A = 0 2 1 × 0 2 1 = 2 4 5 .
2 0 3 2 0 3 8 0 13
1 0 2 5 0 8 21 0 34
3 2
Also A = A . A = 0 2 1 × 2 4 5 = 12 8 23.
2 0 3 8 0 13 34 0 55
Now A 3 − 6 A 2 + 7 A + 2I
21 0 34 5 0 8 1 0 2 1 0 0
= 12 8 23 − 6 2 4 5 + 7 0 2 1 + 2 0 1 0
34 0 55 8 0 13 2 0 3 0 0 1
21 0 34 30 0 48 7 0 14 2 0 0
= 12 8 23 − 12 24 30 + 0 14 7 + 0 2 0
34 0 55 48 0 78 14 0 21 0 0 2
[Link]
M-148
30 0 48 30 0 48 0 0 0
= 12 24 30 − 12 24 30 = 0 0 0 = O.
48 0 78 48 0 78 0 0 0
Hence Cayley-Hamilton theorem is verified. Now we shall compute A −1.
Multiplying (2) by A −1, we set A 2 − 6 A + 7I + 2 A − 1 = O.
−3 0 2
−1 1 2 1 1
∴ A = − (A − 6 A + 2I ) = −1 .
2 2 2 2
0 −1
0 c − b
Example 16: Show that the matrix A = − c 0 a satisfies Cayley-Hamiltion theorem.
b −a 0
(Meerut 2006B)
0 − λ c −b
Solution : We have | A − λI | = − c 0 −λ a
b −a 0 − λ
= − λ (λ2 + a2 ) − c (cλ − ab) − b (ac + bλ ) = − λ3 − λ (a2 + b2 + c 2 ).
∴ the characteristic equation of the matrix A is λ3 + λ (a2 + b2 + c 2 ) = 0 .
We are to verify that A 3 + (a2 + b2 + c 2 ) A = O.
We have
0 c − b 0 c − b
2
A = − c 0 a × − c 0 a
b −a 0 b −a 0
− c 2 − b2 ab ac
= ab − c 2 − a2 bc .
ac bc − b2 − a2
− c 2 − b2 ab ac 0 c − b
2 2
∴ A 3 = A 2A = ab −c − a bc × − c 0 a
ac bc − b2 − a2 b −a 0
[Link]
M- 149
0 − c 3 − b2 c − a2 c bc 2 + b3 + a2 b
3
= c + a c + b2 c
2
0 − ab2 − ac 2 − a3
2 3 2
− bc − b − a b ac + ab2 + a3
2
0
0 c − b
2 2 2
= − (a + b + c )− c 0 a = − (a2 + b2 + c 2 )A.
b −a 0
∴ A3 + (a2 + b2 + c 2 ) A = − (a2 + b2 + c 2 )A + (a2 + b2 + c 2 )A = 0 A = O.
Hence A satisfies Cayley-Hamilton theorem.
1 2 0
Example 17: Find the characteristic equation of the matrix A = 2 −1 0 and hence find
0 0 −1
−1
A . (Rohilkhand 2009, 10)
Solution : The characteristic equation of the matrix A is
| A − λI| = 0
1 − λ 2 0
or 2 −1 − λ 0 =0
0 0 −1 − λ
or (1 − λ )(− 1 − λ ) (−1 − λ ) − 2 [2 (−1 − λ )] = 0
or (1 − λ ) (1 + λ )2 + 4 (1 + λ ) = 0
or (1 − λ )(1 + 2 λ + λ2 ) + 4 + 4 λ = 0
or 1 + 2 λ + λ2 − λ − 2 λ2 − λ3 + 4 + 4 λ = 0
or − λ3 − λ2 + 5 λ + 5 = 0
or λ3 + λ2 − 5 λ − 5 = 0 . ...(1)
Now by Cayley-Hamilton theorem the matrix A must satisfy its characteristic
equation (1). Therefore we have
A 3 + A2 − 5A − 5I = O
[Link]
M-150
or 5I = A3 + A2 − 5 A ...(2)
Pre-multiplying both sides of (2) by A −1 , we have
5 A −1I = A −1A3 + A −1A2 − 5 A −1A
or 5 A −1 = A2 + A − 5 I
1 2
or A −1 = (A + A − 5 I ) …(3)
5
1 2 0 1 2 0 5 0 0
2
Now A = 2 −1 0 2 −1 0 = 0 5 0 .
0 0 −1 0 0 −1 0 0 1
5 0 0 1 2 0 5 0 0
2
∴ A + A − 5 I = 0 5 0 + 2 −1 0 − 0 5 0
0 0 1 0 0 −1 0 0 5
1 2 0
= 2 −1 0 .
0 0 −5
1 2 0
−1 1
Hence from (3), A = 2 −1 0 .
5
0 0 −5
Example 18: State Cayley-Hamilton theorem. Use it to express 2 A5 − 3 A4 + A2 − 4 I as a
3 1
linear polynomial inA, when A = .
−1 2 (Lucknow 2005)
Solution : Statement of Cayley-Hamilton theorem. Every square matrix satisfies its
characteristic equation.
Now let us find the characteristic equation of the matrix A. We have
3−λ 1
| A − λI| =
−1 2−λ
= (3 − λ ) (2 − λ ) + 1 = λ2 − 5 λ + 7.
[Link]
M- 151
The characteristic equation of A is | A − λI| = 0 i.e., is
λ2 − 5 λ + 7 = 0 ...(1)
By Cayley-Hamilton theorem,the matrix A must satisfy (1). Therefore we have
A2 − 5 A + 7 I = O. …(2)
From (2), we get
A2 = 5 A − 7 I … (3)
Multiplying both sides of (3) by A, we get
A3 = 5 A2 − 7 A …(4)
∴ A4 = 5 A3 − 7 A2 …(5)
and A5 = 5 A4 − 7 A3 …(6)
5
Now 2A − 3 A4 + A 2 − 4 I = 2 (5 A4 − 7 A3 ) − 3 A4 + A2 − 4 I
[Substituting for A5 from (6)]
= 7 A4 − 14 A3 + A2 − 4 I = 7 (5 A3 − 7 A2 ) − 14 A3 + A2 − 4 I
[by (5)]
= 21A3 − 48 A2 − 4 I = 21 (5 A2 − 7 A) − 48 A2 − 4 I [by (4)]
= 57 A2 − 147 A − 4 I = 57 (5 A − 7 I) − 147 A − 4 I [by (3)]
= 138 A − 403 I, which is a linear polynomial in A.
4.7 Diagonalization of Square Matrices with Distinct Eigen
Values
Similarity of Matrices. Definition: Let A and B be square matrices of order n. Then B is
said to be similar to A if there exists a non-singular matrix P such that
B = P–1AP.
Theorem 1: Similarity of matrices is an equivalence relation.
[Link]
M-152
Theorem 2: Similar matrices have the same determinant.
Theorem 3: Similar matrices have the same characteristic polynomial and hence the same
eigenvalues. If X is an eigenvector of A corresponding to the eigenvalue λ , then P–1X is an
eigenvector of B corresponding to the eigenvalue λ, where
B = P–1AP.
Corollary. If A is similar to a diagonal matrix D , the diagonal elements of D are the eigenvalues
of A.
Diagonalizable matrix. Definition: A matrix A is said to be diagonalizable if it is similar
to a diagonal matrix.
Thus a matrix A is diagonalizable if there exists an invertible matrix P such that
P-1AP = D where D is a diagonal matrix. Also the matrix P is then said to diagonalize A
or transform A to diagonal form.
Theorem 1: An n × n matrix is diagonalizable if and only if it possesses n linearly independent
eigenvectors.
Proof: Suppose A is diagonalizable. Then A is similar to a diagonal matrix D = dia.
[λ1, λ 2 ,…, λ n].Therefore there exists an invertible matrix P = [X1, X 2, …, X n ] such that
P−1AP = D
i. e., AP = PD
i. e., A[X1, X 2, …, X n ] = [X1, X 2, …, X n ] dia.[λ1, λ 2 ,…, λ n]
i. e, [AX1, AX2 ,…, AX n] = [λ1X1, λ 2 X2 ,…, λ nX n]
i. e., AX1 = λ1X1, AX2 = λ 2 X2 ,…, AX n = λ nX n.
Therefore X1, X 2, …, X n are eigenvectors of A corresponding to the eigenvalues
λ1, λ 2 ,…, λ n respectively. Since the matrix P is non-singular, therefore its column
vectors X1, X 2,…, X n are linearly independent. Hence A possesses n linearly
independent eigenvectors.
Conversely, suppose that A possesses n linearly independent eigenvectors
X1, X 2, …, X n and let λ1, λ 2 ,…, λ n be the corresponding eigenvalues. Then
AX1 = λ1X1, AX2 = λ 2 X2 ,…, AX n = λ nX n.
Let P = [X1, X2 ,…, X n] and D = dia. [λ1, λ 2 ,..., λ n].
Then AP = [X1, X 2 ,…, X n] = [AX1, AX2 ,…, AX n]
[Link]
M- 153
= [λ1X1, λ 2 X2 ,…, λ nX n]
= [X1, X2 ,…, X n] dia. [λ1, λ 2 ,…, λ n] = PD.
Since the column vectors X1 , X2 ,…, X n of the matrix P are linearly independent,
therefore X is invertible and X −1 exists.
Therefore AP = PD ⇒ P−1AP = P−1PD
⇒ P−1AP = D
⇒ A is similar to a diagonal matrix D
⇒ A is diagonalizable.
Remark: In the proof of the above theorem we have shown that if A is diagonalizable
and P diagonalizes A, then
λ1 0 … 0
0 λ2 … 0
P−1AP = =D
… … … …
0 0 … λ n
if and only if the j th column of P is an eigenvector of A corresponding to the eigenvalue
λ j of A, ( j = 1, 2,…, n). The diagonal elements of D are the eigenvalues of A and they
occur in the same order as is the order of their corresponding eigenvectors in the
column vectors of P.
Theorem 2: If the eigenvalues of an n × n matrix are all distinct then it is always similar to a
diagonal matrix.
Proof: Let A be a square matrix of order n and suppose it has n distinct eigenvalues
λ1, λ 2 ,.…, λ n. We know that eigenvectors of a matrix corresponding to distinct
eigenvalues are linearly independent. Therefore A has n linearly independent
eigenvectors and so it is similar to a diagonal matrix
D = dia. [λ1, λ 2 ,…, λ n].
Corollary: Two n × n matrices with the same set of n distinct eigenvalues are similar.
Proof: Suppose A and B are two n × n matrices with the same set of n distinct
eigenvalues λ1, λ 2 ,…, λ n. Let D = dia. [λ1, λ 2 ,…, λ n]. Then both A and B are similar to
D. Now A is similar to D and D is similar to B implies that A is similar to B.
Note that the relation of similarity is transitive.
[Link]
M-154
Theorem 3: The necessary and sufficient condition for a square matrix to be similar to a
diagonal matrix is that the geometric multiplicity of each of its eigenvalues coincides with the
algebraic multiplicity.
Proof: The condition is necessary. Suppose A is similar to a diagonal matrix D = dia.
[λ1, λ 2 ,…, λ n]. Then λ1, λ 2 ,…, λ n are the eigenvalues of A and there exists a
non-singular matrix P such that
P−1AP = D.
Let α be an eigenvalue of A of algebraic multiplicity k. Then exactly k among
λ1, λ 2 ,…, λ n, are equal to α.
Let m = rank (A − αI). Then the system of equations
(A − αI) X = O
have n − m linearly independent solutions and so n − m will be the geometric
multiplicity of α . We are to prove that k = n − m. We know that the rank of a matrix
does not change on multiplication by a non-singular matrix. Therefore
rank (A − αI) = rank [P−1 (A − αI)P]
= rank [P−1AP − αI]
= rank [D − αI] = rank dia. [λ1 − α, λ 2 − α,…, λ n − α]
= n − k , since exactly k elements of dia. [λ1 − α, λ 2 − α,…, λ n − α]
are equal to zero.
Thus rank (A − αI) = m = n − k. Therefore k = n − m.
Thus there are exactly k linearly independent eigenvectors corresponding to the
eigenvalue α.
The condition is sufficient. Suppose that the geometric multiplicity of each
eigenvalue of A is equal to its algebraic multiplicity. Let λ1,…, λ p be the set of p distinct
eigenvalues of A with respective multiplicities r1,…, rp . We have
r1 + … + rp = n.
To prove that A is diagonalizable.
[Link]
M- 155
C11, C12 , …, C1r
1
… … … …
Let ...(1)
… … … …
Cp1, Cp2 , …, Cprp
be linearly independent sets of eigenvectors corresponding to the eigenvalues
λ1,…, λ p respectively. We claim that the n vectors given in (1) are linearly
independent. Let
(a11C11 + a12 C12 + … + a1r C1r ) + … + (ap1Cp1 + … + aprp Cprp ) = O …(2)
1 1
The relation (3) may be written as
X1 + X2 + … + X p = O, ...(3)
where X1, X2 ,…, X p denote the vectors written within brackets in (2) i. e., X1
= a11C11 + … + a1r C1r , and so on.
1 1
Now X1is a linear combination of eigenvectors of A corresponding to the eigenvalue λ1.
Therefore if X1 ≠ O, then X1 is also an eigenvector of A corresponding to the eigenvalue
λ1 .
Similarly we can speak for X2 ,…, X p.
In case some one of X1,…, X p is not zero, then the relation (3) implies that a system of
eigenvectors of A corresponding to distinct eigenvalues of A is linearly dependent. But
this is not possible. Hence each of the vectors X1, X2…, X p must be zero.
Since C11, C12 ,…, C1r is a set of linearly independent vectors, therefore
1
O = X1 = a11 C11 + … + a1r C1r implies that
1 1
a11 = 0 ,…, a1r = 0 .
1
Similarly we can show that each of the scalars in relation (2) is zero. Therefore the n
vectors give in (1) are linearly independent. Thus A has n linearly independent
eigenvectors. So it is similar to a diagonal matrix.
Example 19: Show that the rank of every matrix similar to A is the same as that of A.
Solution: Let B be a matrix similar to A. Then there exists a non-singular matrix P such
that B = P−1AP. We know that the rank of a matrix does not change on multiplication
by a non-singular matrix.
[Link]
M-156
Therefore
rank (P−1AP) = rank A ⇒ rank B = rank A.
Example 20: If U be a unitary matrix such that Uθ AU = diag [λ1,…, λ n], show that λ1,…, λ n
are the eigenvalues of A.
Solution: Let diag [λ1,…, λ n] = D. Since U is unitary, therefore Uθ = U−1. So
Uθ AU = D ⇒ U−1AU = D.
Thus A is similar to the diagonal matrix D. But similar matrices have the same
eigenvalues and eigenvalues of D are its diagonal elements. Therefore λ1,…, λ n are the
eigenvalues of A.
Example 21: Prove that if A is similar to a diagonal matrix, then AT is similar to A.
Solution: Suppose A is similar to a diagonal matrix D. Then there exists a non-singular
matrix P such that
P−1 AP = D
⇒ A = P D P−1
⇒ AT = (P D P−1)T = (P−1)T DT PT
⇒ AT = (PT )−1 D PT [∵ D is diagonal ⇒ DT = D]
⇒ AT is similar to D
⇒ D is similar to AT .
Finally A is similar to D and D is similar to AT implies that A is similar to AT .
Nilpotent Matrix. Definition.
A non-zero matrix A is said to be nilpotent, if for some positive integer r, A r = O.
Example 22: Show that a non-zero matrix is nilpotent if and only if all its eigenvalues are equal
to zero.
Solution: Suppose A ≠ O and A is nilpotent. Then
A r = O, for some positive integer r
⇒ the polynomial λ r annihilates A
[Link]
M- 157
⇒ the minimal polynomial m(λ ) of A divides λ r
⇒ m(λ ) is of the type λ s , where s is some positive integer
⇒ 0 is the only root of m(λ )
⇒ 0 is the only eigenvalue of A
⇒ all eigenvalues of A are zero.
Conversely, each eigenvalue of A = 0
⇒ characteristic equation of A is λ n = 0
⇒ A n = O, since A satisfies its characteristic equation
⇒ A is nilpotent.
Example 23: A square matrix A is defined by
1 2 −2
A= 1 2 1.
−1 −1 0
Find the modal matrix P and the resulting diagonal matrix D of A.
Solution: The characteristic equation of A is
1− λ 2 −2
1 2−λ 1 =0
−1 −1 0−λ
or (1 − λ )(−2 λ + λ2 + 1) − 2 (− λ + 1) − 2(− 1 + 2 − λ ) = 0
or (− λ + 1)(λ − 1)2 + (4 λ − 4) = 0
or (− λ + 1)(λ + 1)(λ − 3) = 0
The roots of this equation are 1, − 1, 3.
The eigenvectors X of A corresponding to the eigenvalue 1 are given by the equation
(A − 1 I) X = O or (A − I) X = O
[Link]
M-158
0 2 −2 x 0
i. e., 1 1 1 y = 0
−1 −1 −1 z 0
0 2 −2 x 0
or 1 1 1 y = 0 , by R3 → R3 + R1
0 0 0 z 0
The coefficient matrix of these equations is of rank 2. So these equations have 1
linearly independent solution. These equations can be written as
0 x + 2 y − 2z = 0 , x + y + z = 0
From these, we get y = z = 1, say. Then x = − 2.
−2
Therefore X1 = 1 is an eigenvector of A corresponding to the eigenvalue 1.
1
Now the eigenvectors of A corresponding to the eigenvalues −1 are given by
(A + I) X = O
2 2 −2 x 0
i. e., 1 3 1 y = 0
−1 −1 1 z 0
2 2 −2 x 0
or 1 3 1 y = 0 , by R3 → 2 R3 + R1
0 0 0 z 0
The coefficient matrix of these equations is of rank 2. So these equations have 1
linearly independent solution.
These equations can be written as
2 x + 2 y − 2z = 0 , x + 3 y + z = 0
−2
Let us take z = − 1, then y = 1and x = − 2. Therefore X2 = 1 is an eigenvector of A
−1
corresponding to the eigenvalue −1.
[Link]
M- 159
Now the eigenvectors of A corresponding to the eigenvalue 3 are given by
(A − 3 I) X = O
−2 2 −2 x 0
i. e., 1 −1 1 y = 0
−1 −1 −3 z 0
−2 2 −2 x 0
or −1 −1 −3 y = 0 , R2 ↔ R3
1 −1 1 z 0
Similarly, we have x = 2, y = 1, z = − 1.
2
Therefore X3 = 1 is an eigenvector of A
−1
corresponding to the eigenvalue 3.
−2 −2 2
Let modal matrix P = [X1, X2 , X3 ] = 1 1 1
1 −1 1
0 −4 −4
−1 Adj P 1
∴ P = =− 2 0 4 .
|P | 8
−2 −4 0
The matrix P will transform A to diagonal form D which is given by the relation
1 0 0
−1
P A P = 0 −1 0 = D
0 0 3
Example 24: Show that the matrix
8 −8 −2
A = 4 −3 −2
3 −4 1
is diagonalizable. Also find the transforming matrix and diagonal matrix.
[Link]
M-160
Solution: The characteristic equation of A is
8−λ −8 −2
4 −3 − λ −2 =0
3 −4 1− λ
1− λ −1 + λ −1 + λ
or 4 −3 − λ −2 = 0 , applying R1 − ( R2 + R3 )
3 −4 1− λ
1 −1 −1
or (1 − λ ) 4 −3 − λ −2 =0
3 −4 1− λ
1 0 0
or (1 − λ ) 4 1− λ 2 = 0 , applying C2 + C1, C3 + C1
3 −1 4−λ
or (1 − λ )[(1 − λ )(4 − λ ) + 2] = 0
or (1 − λ )(λ2 − 5 λ + 6) = 0
or (1 − λ )(λ − 2)(λ − 3) = 0
The roots of this equation are 1, 2, 3.
Since the eigenvalues of the matrix A are all distinct, therefore A is similar to a diagonal
matrix. Since the algebraic multiplicity of each eigenvalue of A is 1, therefore there will
be one and only one linearly independent eigenvector of A corresponding to each
eigenvalue of A.
The eigenvectors X of A corresponding to the eigenvalue 1 are given by the equation
(A − 1I) X = O or (A − I) X = O
7 −8 −2 x1 0
or 4 −4 −2 x2 = 0
3 −4 0 x3 0
7 −8 −2 x1 0
or −3 4 0 x2 = 0 , by R2 → R2 − R1
3 −4 0 x3 0
[Link]
M- 161
7 −8 −2 x1 0
or −3 4 0 x2 = 0 , by R3 → R3 + R2.
0 0 0 x3 0
The matrix of coefficients of these equations has rank 2. Therefore these equations
have only one linearly independent solution as it should have been because the
algebraic multiplicity of the eigenvalue 1 is 1. Note that the geometric multiplicity
cannot exceed the algebraic multiplicity. The above equations can be written as
7 x1 − 8 x2 − 2 x3 = 0 , − 3 x1 + 4 x2 = 0 . From the last equation, we get x1 = 4, x2 = 3.
4
Then the first gives x3 = 2. Therefore X1 = 3 is an eigenvector of A corresponding to
2
the eigenvalue 1.
The eigenvectors X of A corresponding to the eigenvalue 2 are given by the equation
(A − 2 I) X = O
6 −8 −2 x1 0
or 4 −5 −2 x2 = 0
3 −4 −1 x3 0
6 −8 −2 x1 0
or −2 3 0 x2 = 0 ,
0 0 0 x3 0
1
applying R2 → R3 − R1, R3 → R3 − R1.
2
These equations can be written as 6 x1 − 8 x2 − 2 x3 = 0 , − 2 x1 + 3 x2 = 0 . From these, we
get x1 = 3, x2 = 2, x3 = 1.
3
Therefore X2 = 2 is an eigenvector of A corresponding to the eigenvalue 2.
1
The eigenvectors X of A corresponding to the eigenvalue 3 are given by the equation
(A − 3 I) X = O
5 −8 −2 x1 0
or 4 −6 −2 x2 = 0
3 −4 −2 x3 0
[Link]
M-162
5 −8 −2 x1 0
or −1 2 0 x2 = 0 ,
0 0 0 x3 0
applying R2 → R2 − R1, R3 → R3 + R1 − 2 R2 .
These equations can be written as
5 x1 − 8 x2 − 2 x3 = 0 , − x1 + 2 x2 = 0 .
From these, we get x1 = 2, x2 = 1, x3 = 1.
2
∴ X3 = 1 is an eigenvector of A corresponding to the eigenvalue 3.
1
4 3 2
Let P = [X1 X2 X3 ] = 3 2 1.
2 1 1
The columns of P are linearly independent eigenvectors of A corresponding to the
eigenvalues 1, 2, 3 respectively. The matrix P will transform A to diagonal form D
which is given by the relation
1 0 0
−1
P A P = 0 2 0 = D.
0 0 3
Comprehensive Exercise 1
1. (i) State Cayley-Hamilton theorem.
(Bundelkhand 2005; Agra 05; Lucknow 07)
1 2 3
(ii) Find the characteristic roots of the matrix A = 0 −4 2.
0 0 7
(Meerut 2009 B)
[Link]
M- 163
a c b
2. (i) If a + b + c = 0, find the characteristic roots of the matrix A = b b a.
c a c
(Kumaun 2008; Kanpur 07, 08)
(ii) Prove that the matrices :
0 a b 0 b a
A = a 0 c and B = b 0 c have
b c 0 a c 0
same characteristic equation.
(Rohilkhand 2007; Purvanchal 06; Agra 05)
1 1
3. (i) Verify Cayley-Hamilton theorem for matrix A, where A = .
1 1
(Rohilkhand 2011)
(ii) Verify that the matrix A satisfies its characteristic equation, where
1 2
A= .
−1 3
Express A6 − 4 A5 + 8 A4 − 12 A3 + 14 A2 as a linear polynomial in A.
1 4
4. Find the characteristic roots o the matrix A = and verify Cayley-
2 3
Hamilton theorem for this matrhix. Find the inverse of the matrix A and also
express A 5 − 4 A 4 − 7 A 3 + 11A 2 − A − 10 I as a linear polynomial in A .
1 2 1
5. Verify that the matrix A = 0 1 −1 satisfies its characteristic equation and
3 −1 1
compute A −1. (Purvanchal 2009)
1 1 3
6. Find the characteristic roots of the matrix A = 5 2 6 and verify
−2 −1 −3
Cayley-Hamilton theorem. (Agra 2008; Bundelkhand 11; Meerut 09B)
7. Determine the eigenvalues and eigenvectors of the matrix
1 0 −2
A = 0 0 0 .
−2 0 4
[Link]
M-164
2 2 1
8. Show that the matrix A = 1 3 1 satisfies Calyley-Hamilton theorem.
1 2 2
Also determine the characteristic roots (i. e., latent roots) and the corresponding
characteristic vectors of the matrix A. (Meerut 2007)
0 0 1
9. Verify Cayley-Hamilton theorem for the matrix A = 3 1 0 .
−2 1 4
Hence or otherwise evaluate A −1. (Meerut 2003, 09; Avadh 05)
1 3 7
10. Find the characteristic equation of the matrix A = 4 2 3 and show that it
0 2 1
is satisfied by A. Hence obtain the inverse of the given matrix A.
1 2 1
11. Verify that the matrix A = −1 0 3 satisfies its own characteristic
2 −1 1
equation. Is it true of every square matrix? State the theorem that applies here.
12. Verify Cayley-Hamilton theorem for the following matrix :
1 0 2
A = 0 −1 1 .
0 1 0 (Avadh 2008)
13. Find the characteristic roots and the characteristic spaces of the matrix
1 2 3
0 2 3 .
0 0 2 (Meerut 2005; Rohilkhand 07)
14. Show that the characteristic roots of a triangular matrix are just the diagonal
elements of the matrix.
15. Let A and B be n-rowed square matrices and let A be non-singular. Show that
the matrices A −1B and BA −1 have the same eigenvalues.
16. If A and B are non-singular matrices of order n, show that the matrices AB and
BA are similar.
17. A and B are two n × n matrices with the same set of n distinct eigenvalues. Show
that there exist two matrices P and Q (one of them non-singular) such that
A = PQ, B = QP.
18. Prove that a non-zero nilpotent matrix cannot be similar to a diogonal matrix.
4 1
19. Find a matrix P which diagonalizes the matrix A = . Verify that
2 3
P−1 AP = D, where D is the diagonal matrix.
[Link]
M- 165
1 −1 2
20. Reduce the matrix A = 0 2 −1 to diagonal form.
0 0 3
A nswers 1
1. (ii) 1, –4, 7
1 /2
3
2. (i) λ = 0 , ± (a2 + b2 + c 2 )
2
3. (ii) −4 A + 5 I.
1 −3 4
4. 5, − 1;
5 2 −1
5. Characteristic equation is λ3 − 3 λ2 − λ + 9 = 0
0 3 3
1
A −1
= 3 2 −1
9
3 −7 −1
6. All the characteristic roots are zero.
7. Eigenvalues are 5, 0 , 0 . Corresponding to the eigenvalue 2 an eigenvector is
[1, 0 , − 2]′. Two linearly independent eigenvectors corresponding to the
eigenvalue 0 are [2, 0 , 1]′ and [0 , 1, 0 ]′
8. 5, 1, 1. Corresponding to the characteristic root 5 a characteristic vector is [1, 1, 1]′.
Two linearly independent characteristic vectors corresponding to the
characteristic root 1 are [1, 0 , − 1]′ and [1, 1, − 3]′
4 1 −1
1
9. A −1
= −12 2 3
5
5 0 0
10. λ3 − 4 λ2 − 13 λ − 40 = 0 ;
−4 11 −5
1
A −1 = −4 1 25
40
8 −2 −10
13. Characteristic roots are 1, 1, 2. Corresponding to the characteristic root 1 a
characteristic vector is [1, 0 , 0 ]′. Corresponding to the characteristic root 2 a
characteristic vector is [2, 1, 0 ]′ . The characteristic space corresponding to the
characteristic root 1 consists of the vector c [1, 0 , 0 ]′, where c is any scalar.
Similarly for the other characteristic space
[Link]
M-166
19. 2,5. Corresponding to the characteristic root 2 a characteristic vector is [1, − 2]′.
Corresponding to the characteristic root 5 a characteristic vector is [1 , 1]′ .
1 1
P=
−2 1
20. 1, 2, 3 corresponding to the characteristic root 1 a characteristic vector is [1, 0 , 0 ]′ .
Corresponding to the characteristic root 2 a characteristic vector is [1, − 1, 0 ]′.
Corresponding to the characteristic root 3 a characteristic vector is [3, − 2, 2]′.
Objective Type Questions
Multiple Choice Questions
Indicate the correct answer for each question by writing the corresponding letter from (a),
(b), (c) and (d).
1. If λ is a characteristic root of a matrix A, then a characteristic root of A −1 is
(a) 1 / λ (b) λ
(c) λ2 (d) 1 / λ2
(Bundelkhand 2001)
a h g
2. The eigenvalues of the matrix A = 0 b 0 are
0 c c
(a) a, h, g (b) a, h, c
(c) a, g, c (d) a, b, c
(Kanpur 2009, 11)
3. If λ is a characteristic root of the matrix A, then a characteristic root of the matrix
A + kI is
(a) λ (b) k + λ
(c) k − λ (d) none of these (Agra 2007)
3 1
4. The characteristic equation of A = is
−1 2
(a) λ2 − 5 λ + 7 = 0 (b) λ2 − 3 λ + 7 = 0
(c) λ2 − 2 λ + 7 = 0 (d) none of these (Bundelkhand 2007)
[Link]
M- 167
Fill in the Blank(s)
Fill in the blanks ‘‘.........’’, so that the following statements are complete and correct.
1. If A is any n-rowed square matrix and λ an indeterminate, then the equation
| A − λI| = 0 is called the ......... of A and the roots of this equation are called the
......... of the matrix A.
2. If λ is a characteristic root of an n × n matrix A, then a non-zero vector X such that
AX = λX is called a ......... of A corresponding to the characteristic root λ.
3. The characteristic roots of a Hermitain matrix are ......... (Avadh 2005)
4. The characteristic roots of a skew-Hermitian matrix are either ......... or .........
5. The matrices A and A ′ have the ......... eigenvalues. (Agra 2008)
6. If λ1, ........., λ n are the eigenvalues of A, then kλ1, ........., kλ n are the
eigenvalues of ......... (Lucknow 2011)
7. If the characteristic roots of A are λ1, λ 2 , ........., λ n, then the characteristic roots
2 2 2
of ......... are λ1 , λ 2 , ........., λ n .
8. Every square matrix ..... its characteristic equation. (Meerut 2001)
1 0 2
9. The characteristic equation of the matrix A = 0 −1 1 is ......... .
0 1 0
True or False
Write ‘T’ for true and ‘F’ for false statement.
1. The characteristic roots of a real symmetric matrix are all real.
2. The characteristic roots of a real skew-symmetric matrix are all pure imaginary.
3. The characteristic roots of a unitary matrix are of unit modulus.
4. The characteristic roots of a diagonal matrix are just the diagonal elements of the
matrix.
5. Two matrices A and C−1AC do not have the same characteristic roots.
6. Cayley-Hamilton theorem states that ‘Every square matrix satisfies its
characteristic equation’. (Rohilkhand 2006)
7. The characteristic equation of
3 1
A=
−1 2
is λ2 − 3 λ + 7 = 0 . (Meerut 2003; Kumaun 08)
[Link]
M-168
A nswers
Multiple Choice Questions
1. (a) 2. (d) 3. (b) 4. (a)
Fill in the Blank(s)
1. characteristic equations; characteristic roots
2. characteristic vector 3. real 3. pure imaginary; zero
4. same 5. kA 7. A2
8. satisfies 9. λ3 − 2 λ + 1 = 0
True or False
1. T 2. F 3. T
4. T 5. F 6. T
7. F
o
[Link]
M- 169
5
Quadratic Forms
5.1 Quadratic Forms
n n
Definition: An expression of the form ∑ ∑ aij xi x j , where aij's are elements of a field F, is
i =1 j =1
called a quadratic form in the n variables x1 , x2 , … ..., xn over a field F.
Real Quadratic Form. Definition : As expression of the form
n n
∑ ∑ aij xi x j,
i =1 j =1
where aij's are all real numbers, is called a real quadratic form in the n variables x1, x2 ,…, xn.
For example
(i) 2 x2 + 7 xy + 5 y2 is a real quadratic form in the two variables x and y.
(ii) 2 x2 − y2 + 2 z 2 − 2 yz − 4 zx + 6 xy is a real quadratic form in the three variables
x, y and z.
[Link]
M-170
2 2 2 2
(iii) x1 − 2 x2 + 4 x3 − 4 x4 − 2 x1 x2 + 3 x1 x4 + 4 x2 x3 − 5 x3 x4 is a real quadratic form
in the four variables x1, x2 , x3 and x4 .
Theorem : Every Quadratic form over a field F in n variables x1, x2 ,…, xn can be expressed in
the form X ′ BX where
X = [ x1, x2 ,……, xn]T
is a column vector and B is a symmetric matrix of order n over the field F.
n n
Proof : Let ∑ ∑ aij xi x j, …(1)
i =1 j =1
be a quadratic form over the field F in the n variables x1, x2 ,……, xn.
In (1) it is assumed that xi x j = x j xi. Then the total coefficient of xi x j in (1) is aij + a ji. Let
us assign half of this coefficient to xij and half to x ji. Thus we define another set of
1
scalars bij, such that bii = aii and bij = b ji = (aij + a ji), i ≠ [Link] we have
2
n n n n
∑ ∑ aij xi x j = ∑ ∑ bij xi x j.
i =1 j =1 i =1 j =1
Let B = [bij]n × [Link] B is a symmetric matrix of order n over the field F since bij = b ji.
x1
x2
Let X = …. Then XT or X′ = [ x1 x2 …… xn].
…
xn
Now XT BX is a matrix of the type 1 × 1. It can be easily seen that the single element of
n n
this matrix is ∑ ∑ bij xi x j. If we identify a1 × 1matrix with its single element i. e., if we
i =1 j =1
regard a 1 × 1matrix equal to its single element, then we have
n n n n
XT BX = ∑ ∑ bij xi x j = ∑ ∑ aij xi x j.
i =1 j =1 i =1 j =1
Hence the result.
[Link]
M- 171
4.2 Matrix of a Quadratic Form
n n
Definition : If φ = ∑ ∑ aij xi x j is a quadratic form in n variables x1, x2 ,……, then there
i =1 j =1
exists a unique symmetric matrix B of order n such that φ = XT BX , where X = [ x1 x2 …… xn]T .
n n
The symmetric matrix B is called the matrix of the quadratic form ∑ ∑ aij xi x j.
i =1 j =1
Since every quadratic form can always be so written that matrix of its coefficients is a
symmetric matrix, therefore we shall be considering quadratic forms which are so
adjusted that the coefficient matrix is symmetric.
4.2 Quadratic Form Corresponding to a Symmetric Matrix.
Let A = [aij]n × n be symmetric matrix over the field F and let X = [ x1 x2 …… xn]T be a
n n
column vector. Then XT AX determines a unique quadratic form ∑ ∑ aij xi x j in n
i =1 j =1
variables x1, x2 ,……, xn over the field F.
Thus we have seen that there exists a one-to-one correspondence between the set of all
quadratic forms in n variables over a field F and the set of all n-rowed symmetric
matrices over F.
Example 1: Write down the matrix of each of the following quadratic forms and verify that they
can be written as matrix products XT AX :
2 2
(i) x1 − 18 x1 x2 + 5 x2 .
2 2 2
(ii) x1 + 2 x2 − 5 x3 − x1 x2 + 4 x2 x3 − 3 x3 x1.
Solution: (i) The given quadratic form can be written as x1 x1 − 9 x1 x2 − 9 x2 x1 + 5 x2 x2 .
[Link]
M-172
1 −9
Let A be the matrix of this quadratic form. Then A = .
−9 5
x1
Let X= . Then X′ = [ x1 x2 ].
x2
1 −9
We have X ′ A = [ x1 x2 ] = [ x1 − 9 x2 − 9 x1 + 5 x2 ].
−9 5
x1
∴ X ′ AX = [ x1 − 9 x2 − 9 x1 + 5 x2 ]
x2
= x1( x1 − 9 x2 ) + x2 (− 9 x1 + 5 x2 )
2 2
= x1 − 9 x1 x2 − 9 x2 x1 + 5 x2
2 2
= x1 − 18 x1 x2 + 5 x2 .
(ii) The given quadratic form can be written as
1 3 1 3
x1 x1 − x1 x2 − x1 x3 − x2 x1 + 2 x2 x2 + 2 x2 x3 − x3 x1 + 2 x3 x2 − 5 x3 x3 .
2 2 2 2
Let A be the matrix of this quadratic form. Then
1 −1 / 2 −3 / 2
A = −1 / 2 2 2
−3 / 2 2 −5
Obviously A is a symmetric matrix.
x1
Let X = x2 . Then X′ = [ x1 x2 x3 ].
x
3
1 −1 / 2 −3 / 2
We have X ′ A = [ x1 x2 x3 ] −1 / 2 2 2
−3 / 2 2 −5
1 3 1 3
= [x1 − x2 − x3 − x1 + 2 x2 + 2 x3 − x1 + 2 x2 − 5 x3 ].
2 2 2 2
[Link]
M- 173
x1
1 3 1 3
∴ X ′ AX = [x1 − x2 − x3 − x1 + 2 x2 + 2 x3 − x1 + 2 x2 − 5 x3 ] x2
2 2 2 2 x
3
1 3 1 3
= x1 ( x1 − x2 − x3 ) + x2 ( − x1 + 2 x2 + 2 x3 ) + x3 ( − x1 + 2 x2 − 5 x3 )
2 2 2 2
2 1 3 1 2 3 2
= x1 − x1 x2 − x1 x3 − x2 x1 + 2 x2 + 2 x2 x3 − x3 x1 + 2 x3 x2 − 5 x3
2 2 2 2
2 2 2
= x1 + 2 x2 − 5 x3 − x1 x2 + 4 x2 x3 − 3 x3 x1.
Example 2: Obtain the matrices corresponding to the following quadratic forms
(i) ax2 + 2 hxy + by2 (ii) 2 x1 x2 + 6 x1 x3 − 4 x2 x3.
(iii) x12 + 5 x22 − 7 x32 (iv) 2 x12 − 7 x32 + 4 x1 x2 − 6 x2 x3 .
Solution: (i) The given quadratic form can be written as ax2 + hxy + hyx + by2 .
∴ if A is the matrix of this quadratic form, then
a h
A= , which is a symmetric matrix of order 2.
h b
(ii) The given quadratic form can be written as
2 2 2
0 x1 + x1 x2 + 3 x1 x3 + x2 x1 + 0 x2 − 2 x2 x3 +3 x3 x1 − 2 x3 x2 + 0 x3 .
∴ if A is the matrix of this quadratic form, then
0 1 3
A = 1 0 −2.
3 −2 0
(iii) The given quadratic form can be written as
2 2 2
x1 + 0 x1 x2 + 0 x1 x3 + 0 x2 x1 + 5 x2 + 0 x2 x3 + 0 x3 x1 + 0 x3 x2 − 7 x3 .
1 0 0
∴ if A is the matrix of this quadratic form, then A = 0 5 0 .
0 0 −7
[Link]
M-174
(iv) The given quadratic form can be written as
2 2
2 x1 + 2 x1 x2 + 0 x1 x3 + 2 x2 x1 + 0 x22 − 3 x2 x3 + 0 x3 x1 − 3 x3 x2 − 7 x3 .
2 2 0
∴ if A is the matrix of this quadratic form, then A = 2 0 −3.
0 −3 −7
Example 3: Find the matrix of the quadratic form
2 2
x12 − 2 x2 − 3 x3 + 4 x1 x2 + 6 x1 x3 − 8 x2 x3
and verify that it can be written as a matrix product X ′ AX.
Solution: The given quadratic form can be written as
2 2 2
x1 + 2 x1 x2 + 3 x1 x3 + 2 x2 x1 − 2 x2 − 4 x2 x3 + 3 x3 x1 − 4 x3 x2 − 3 x3 .
Let A be the matrix of this quadratic form. Then
1 2 3
A = 2 −2 −4.
3 −4 −3
x1
Let X = x2 . Then X′ = [ x1 x2 x3 ].
x
3
1 2 3
We have X ′ A = [ x1 x2 x3 ] 2 −2 −4
3 −4 −3
= [x1 + 2 x2 + 3 x3 2 x1 − 2 x2 − 4 x3 3 x1 − 4 x2 − 3 x3 ].
x1
∴ X ′ AX = [x1 + 2 x2 + 3 x3 2 x1 − 2 x2 − 4 x3 3 x1 − 4 x2 − 3 x3 ] x2
x
3
= ( x1 + 2 x2 + 3 x3 ) x1 + (2 x1 − 2 x2 − 4 x3 ) x2 + (3 x1 − 4 x2 − 3 x3 ) x3
2 2 2
= x1 − 2 x2 − 3 x3 + 4 x1 x2 − 8 x2 x3 + 6 x3 x1.
[Link]
M- 175
Example 4: Write down the quadratic forms corresponding to the following matrices :
0 1 2 3
0 5 −1
1 2 3 4
(i) 5 1 6 (ii) .
−1 2 3 4 5
6 2
3 4 5 6
Solution: (i) Let X = [ x1 x2 x2 ]T and A denote the given symmetric matrix. Then
XT AX is the quadratic form corresponding to this matrix. We have
0 5 −1
T
X A = [ x1 x2 x3 ] 5 1 6
−1 6 2
= [5 x2 − x3 5 x1 + x2 + 6 x3 − x1 + 6 x2 + 2 x3 ].
∴ XT AX = x1(5 x2 − x3 ) + x2 (5 x1 + x2 + 6 x3 ) + x3 ( − x1 + 6 x2 + 2 x3 )
2 2
= x2 + 2 x3 + 10 x1 x2 − 2 x1 x3 + 12 x2 x3 .
(ii) Let X = [ x1 x2 x3 x4 ]T and A denote the given symmetric matrix. Then
XT AX is the quadratic form corresponding to this matrix. We have
0 1 2 3
T 1 2 3 4
X A = [ x1 x2 x3 x4 ]
2 3 4 5
3 4 5 6
= [x2 + 2 x3 + 3 x4 x1 + 2 x2 + 3 x3 + 4 x4
2 x1 + 3 x2 + 4 x3 + 5 x4 3 x1 + 4 x2 + 5 x3 + 6 x4 ]
∴ XT AX = x1( x2 + 2 x3 + 3 x4 ) + x2 ( x1 + 2 x2 + 3 x3 + 4 x4 )
+ x3 (2 x1 + 3 x2 + 4 x3 + 5 x4 ) + x4 (3 x1 + 4 x2 + 5 x3 + 6 x4 )
2 2 2
= 2 x2 + 4 x3 + 6 x4 + 2 x1 x2 + 4 x1 x3 + 6 x1 x4 + 6 x2 x3 + 8 x2 x4 + 10 x3 x4 .
[Link]
M-176
Comprehensive Exercise 1
1. Obtain the matrices corresponding to the following quadratic forms :
(i) x2 + 2 y2 + 3 z 2 + 4 xy + 5 yz + 6 zx.
(ii) ax2 + by2 + cz 2 + 2 fyz + 2 gzx + 2 hxy.
2 2 2
(iii) a11 x1 + a22 x2 + a33 x3 + 2 a12 x1 x2 + 2 a23 x2 x3 + 2 a31 x3 x1.
2 2 2 2
(iv) x1 − 2 x2 + 4 x3 − 4 x4 − 2 x1 x2 + 3 x1 x4 + 4 x2 x3 − 5 x3 x4.
(v) x1 x2 + x2 x3 + x3 x1 + x1 x4 + x2 x4 + x3 x4.
2
(vi) x1 − 2 x2 x3 − x3 x4.
2 2 2 2 2
(vii) d1 x1 + d2 x2 + d3 x3 + d4 x4 + d5 x5 .
2. Write down the quadratic forms corresponding to the following symmetric matrices :
0 a b c
1 2 3
a 0 l m
(i) 2 0 3 (ii) .
3 b l 0 p
3 1
c m p 0
(iii) diag. [λ1, λ 2 ,……, λ n].
A nswers
1 2 3 a h g
1. (i) 2 2 5 / 2 (ii) h b f
3 5/2 3 g f c
3
1 −1 0
2
a11 a12 a31 −1
−2 2 0
(iii) a12 a22 a23 (iv) 5
0 2 4 −
a a23 a33 2
31
3 0 −
5
−4
2 2
[Link]
M- 177
1 1 1
0 2 2 2 1 0 0 0
1 1 1
0 0 0 −1 0
2 2 2 (vi) 0 1
(v) −1 0 −
1 1
0
1 2
2 2 2 1
1 1 1 0 0 −
2
0
2 0
2 2
(vii) diag. [d1, d2 , d3 , d4 , d5 ]
2 2
2. (i) x1 + x3 + 4 x1 x2 + 6 x1 x3 + 6 x2 x3
(ii) 2 ax1 x2 + 2 bx1 x3 + 2 cx1 x4 + 2 lx2 x3 + 2 mx2 x4 + 2 px3 x4
2 2 2
(iii) λ1 x1 + λ 2 x2 + ...... + λ n xn
Objective Type Questions
Fill in the Blank(s)
Fill in the blanks “……”, so that the following statements are complete and correct.
n n
1. An expression of the form ∑ ∑ aij xi x j, where aij's are elements of a field F, is
i =1 j =1
called a …… in the n variables x1, x2 ,……, xn over a field F.
2. There exists a one-to-one correspondence between the set of all quadratic forms
in n variables over a field F and the set of all n-rowed …… matrices over F.
3. The matrix A corresponding to the quadratic form
2 2 2 2 2
d1 x1 + d2 x2 + d3 x3 + d4 x4 + d5 x5 is A = .............
4. The quadratic form corresponding to the symmetric matrix diag. [λ1, λ 2 ,……, λ n]
is … .
2 1 5
5. The quadratic form corresponding to the matrix 1 3 −2 is ….
5 −2 4
[Link]
M-178
True or False
Write ‘T’ for true and ‘F’ for false statement.
1. The matrix corresponding to the quadratic form d1 x12 + d2 x22 + d3 x32 + d4 x42 is a
diagonal matrix.
2. Every quadratic form over a field F in n variables x1, x2 ,……, xn can be expressed in
the form X ′ BX where X = [ x1, x2 ,……, xn]T is a column vector and B is a skew
symmetric matrix of order n over the field F.
A nswers
Fill in the Blank(s)
1. quadratic form 2. symmetric 3. diag. [d1, d2 , d3 , d4 , d5 ]
2 2 2
4. λ1 x1 + λ 2 x2 + …… + λ n xn
2 2 2
5. 2 x1 + 3 x2 + 4 x3 + 2 x1 x2 + 10 x1 x3 − 4 x2 x3
True or False
1. T 2. F
Common questions
Powered by AIThe rank of a matrix determines the dimension of the solution space of a system of equations it represents. A rank equal to the number of unknowns suggests a unique solution, while a lower rank indicates multiple solutions. Matrix transformations preserve rank and hence the characteristics of the solution space, allowing for reduced forms without solution loss .
Every m×n matrix of rank r can be reduced to the form \( \begin{bmatrix} I_r & O \end{bmatrix} \) by a finite chain of elementary transformations, which do not alter the rank as they involve adding multiples of one row to another, swapping rows, or multiplying a row by a non-zero scalar .
The characteristic roots (eigenvalues) of a matrix A are found by solving the characteristic equation |A - λI| = 0. For each eigenvalue, the eigenvectors are determined by solving (A - λI)x = 0 for non-zero x. If a matrix's equation leads to solutions where x = 0 is the only solution, no non-zero eigenvectors exist for that λ .
Two matrices are row equivalent if one can be transformed into the other via a finite number of elementary row transformations, and similarly, they are column equivalent if one can be transformed into the other via elementary column transformations .
A matrix is said to be in Echelon form if all zero rows occur at the bottom, and the number of zeros before the first non-zero element in a row is less than in the next row. The rank of a matrix in Echelon form is determined by the number of non-zero rows .
A set of vectors is linearly independent if no vector in the set can be represented as a linear combination of the others. For example, the vectors \( \begin{bmatrix} 1 \end{bmatrix}, \begin{bmatrix} 0 \end{bmatrix}, \begin{bmatrix} 0 \end{bmatrix} \) are linearly independent because the only solution to \( k_1 \begin{bmatrix} 1 \end{bmatrix} + k_2 \begin{bmatrix} 0 \end{bmatrix} + k_3 \begin{bmatrix} 0 \end{bmatrix} = \begin{bmatrix} 0 \end{bmatrix} \) is \( k_1 = k_2 = k_3 = 0 \).
A quadratic form \( \phi \) in variables is represented by a symmetric matrix B such that \( \phi = X^T BX \), where X is a vector of the variables. The matrix is symmetric because all quadratic forms can be adjusted so that their coefficient matrix reflects this property, reflecting the requirement that mixed terms have a consistent coefficient .
To obtain the modal and diagonal matrices of a matrix A, solve for its eigenvalues and eigenvectors. The modal matrix P consists of eigenvectors as columns, and the diagonal matrix D has corresponding eigenvalues on its diagonal. A matrix is diagonalizable if P exists, meaning there are enough independent eigenvectors to span the space .
A system of linear equations has a unique solution if the rank of the coefficient matrix equals the rank of the augmented matrix and equals the number of variables. It has infinite solutions if the rank of both matrices is less than the number of variables but equal, and it has no solution if the rank of the coefficient matrix is less than the rank of the augmented matrix .
The unique correspondence between quadratic forms and symmetric matrices allows for consistent and efficient manipulation of multivariable functions, facilitating analysis in optimization, physics, and statistics. Symmetric matrices also simplify computations, such as finding eigenvalues and diagonalization, critical in these applications .
You might also like
- Discrete Mathematics Notes for BCA 3rd SemNo ratings yetDiscrete Mathematics Notes for BCA 3rd Sem281 pages
- Ordinary Differential Equations OverviewNo ratings yetOrdinary Differential Equations Overview5 pages
- Dynamics 21st Edition A.R. Vasishtha Ebook Unrestricted AccessNo ratings yetDynamics 21st Edition A.R. Vasishtha Ebook Unrestricted Access90 pages
- Quadratic Forms: Canonical Reduction GuideNo ratings yetQuadratic Forms: Canonical Reduction Guide5 pages
- Handwritten Notes on Differential GeometryNo ratings yetHandwritten Notes on Differential Geometry72 pages
- Krishna Publication Mathematics CatalogNo ratings yetKrishna Publication Mathematics Catalog25 pages
- Complex Variables for Post-Graduate Students100% (2)Complex Variables for Post-Graduate Students462 pages
- Limits and Continuity in Multivariable FunctionsNo ratings yetLimits and Continuity in Multivariable Functions80 pages
- NEP - Syllabus M.a.& M.sc. Mathematics (2022-2023) PDFNo ratings yetNEP - Syllabus M.a.& M.sc. Mathematics (2022-2023) PDF47 pages
- Group Theory (Isomorphism and Homomorphism)No ratings yetGroup Theory (Isomorphism and Homomorphism)6 pages
- Differential Geometry and Tensor Analysis CentumNo ratings yetDifferential Geometry and Tensor Analysis Centum33 pages
- Dips OrdinaryDifferentialEquation HandWrittenNotes 338pagesNo ratings yetDips OrdinaryDifferentialEquation HandWrittenNotes 338pages94 pages
- Chapter 1-Linear Algebra, 2Nd Edition - Kenneth Hoffmann and Ray Kunze PDFNo ratings yetChapter 1-Linear Algebra, 2Nd Edition - Kenneth Hoffmann and Ray Kunze PDF35 pages
- Introduction to Partial Differential EquationsNo ratings yetIntroduction to Partial Differential Equations33 pages
- First Order Differential Equations GuideNo ratings yetFirst Order Differential Equations Guide24 pages
- Matrix Project Overview and ApplicationsNo ratings yetMatrix Project Overview and Applications15 pages
- Introduction to Matrices and DefinitionsNo ratings yetIntroduction to Matrices and Definitions2 pages
- Alan Green : selected works from 1991 to 2003 : 22 May - 18 July 2008, Annely Juda Fine Art ... London by Green, Alan, 1932-2003, Green, June; Annely Juda Fine Art instant download ebook testbank solutions pdf available100% (1)Alan Green : selected works from 1991 to 2003 : 22 May - 18 July 2008, Annely Juda Fine Art ... London by Green, Alan, 1932-2003, Green, June; Annely Juda Fine Art instant download ebook testbank solutions pdf available51 pages
- Evaluating Landslide Early Warning ModelsNo ratings yetEvaluating Landslide Early Warning Models20 pages
- Parallel Algorithm for Recurrence EquationsNo ratings yetParallel Algorithm for Recurrence Equations8 pages
- An Introduction To Neural Networks For BeginnersNo ratings yetAn Introduction To Neural Networks For Beginners40 pages
- Quantum Explanation of Disjunction EffectsNo ratings yetQuantum Explanation of Disjunction Effects6 pages
- Banker's Algorithm for Deadlock AvoidanceNo ratings yetBanker's Algorithm for Deadlock Avoidance5 pages
- University of Toronto Math Competition 2002No ratings yetUniversity of Toronto Math Competition 20027 pages