Md Firoz Ali 1

1 Elementary algebraic structures

1.1 Binary operation
• Let S be a non-empty set. Then a mapping f : S × S → S is called a binary operation in
S.
• In general, binary operation is denoted by ◦ or ∗ and we write f (a, b) = a ◦ b or f (a, b) = a ∗ b.
• A non-empty set S along with one or more binary operations defined on it is called an alge-
braic structure.

Example 1.1 (i) ∗ : R × R → R defined by a ∗ b = a + b is a binary operation.

(ii) ∗ : R × R → R defined by a ∗ b = a − b is a binary operation.
(iii) ∗ : R × R → R defined by a ∗ b = a.b is a binary operation.
(iv) ∗ : R×R → R defined by a∗b = a/b is not a binary operation but ∗ : R\{0}×R\{0} → R\{0}
defined by a ∗ b = a/b is a binary operation.
(v) ∗ : R × R → R defined by a ∗ b = a + b − 1 is a binary operation.
(vi) ∗ : N × N → N defined by a ∗ b = a + b is a binary operation.
(vii) ∗ : N × N → N defined by a ∗ b = a − b is not a binary operation.

1.2 Group
Let G be a non empty set and ∗ be a binary operation on G. The algebraic structure G is called
a group if it satisfies the following properties
(i) a ∗ b ∈ G for all a, b ∈ G (closure property)
(ii) a ∗ (b ∗ c) = (a ∗ b) ∗ c for all a, b, c ∈ G (associativity property)
(iii) there exists an element e ∈ G such that a ∗ e = e ∗ a = a for all a ∈ G. (Existence of identity)
(iv) for each element a ∈ G, there exists an element a0 ∈ G such that a ∗ a0 = a0 ∗ a = e (Existence
of Inverse Element).
The group is denoted by (G, ∗). Here e is called the identity element of the group G. a0 is called
the inverse of a and is denoted by a−1 .

The group (G, ∗) is called commutative or abelian if a ∗ b = b ∗ a hold for all a, b ∈ G.

Example 1.2 (i) (R, +) is a commutative group.

(ii) (Z, +) is a commutative group.
(iii) (N, +) is not a group.
(iv) (R, .) is not a group but (R \ {0}, .) is a commutative group.
(v) (C, +) is a group.
(vi) (C, .) is not a group but (C \ {0}, .) is a commutative group.
(vii) Let Mm×n (R) be the set of all m × n matrices whose elements are real numbers. Then
(Mm×n (R), +) is a commutative group. But (Mn×n (R), .) is not a group.
(viii) Let GL(n, R) = {A ∈ Mn×n (R) : det A 6= 0}. Then (GL(n, R), .) is a non-commutative
group, known as general linear group.
(ix) Let Mm×n (C) be the set of all m × n matrices whose elements are complex numbers. Then
(Mm×n (C), +) is a commutative group. But (Mn×n (C), .) is not a group.
Md Firoz Ali 2

(x) Let X be a non empty set. Then (P(X), ∪) and (P(X), ∩) are not groups.

Example 1.3 Let n be a fixed integer. Define a relation ≡n on the set of integers Z by x ≡n y iff
x − y is divisible by n. Then ≡n is an equivalence relation. The equivalence classes associated with
this equivalence relation are known as classes of residues modulo n. The equivalence classes
are given by

0 = {m ∈ Z : m ≡n 0} = {m ∈ Z : m − 0 is divisible by n} = {0, ±n, ±2n, ±3n, . . .}

1 = {m ∈ Z : m ≡n 1} = {m ∈ Z : m − 1 is divisible by n} = {0, ±n + 1, ±2n + 1, ±3n + 1, . . .}
2 = {m ∈ Z : m ≡n 2} = {m ∈ Z : m − 2 is divisible by n} = {0, ±n + 2, ±2n + 2, ±3n + 2, . . .}
..
.
n − 1 = {m ∈ Z : m ≡n (n − 1)} = {m ∈ Z : m − (n − 1) is divisible by n}
= {0, ±n + (n − 1), ±2n + (n − 1), ±3n + (n − 1), . . .}

Then Z = 0 ∪ 1 ∪ 2 ∪ . . . ∪ n − 1. Let Zn = {0, 1, 2, . . . , n − 1}. Define +n and ·n on Zn by

x +n y = x + y & x ·n y = x · y.

Then (Zn , +n ) is a group. Here 0 is the identity element and the inverse of x ∈ Zn is n − x.
But (Zn , ·n ) is not a group. Here 1 is the identity element but the inverse of x ∈ Zn does not exist
always.

1.3 Subgroup
Let (G, ∗) be a group and H be a non-empty subset of G. If (H, ∗) is also a group then (H, ∗) is
called a subgroup of (G, ∗).

Theorem 1.1 Let (G, ∗) be a group. A non-empty subset H of G forms a subgroup of (G, ∗) if
and only if
(i) a, b ∈ H =⇒ a ∗ b ∈ H and
(ii) a ∈ H =⇒ a−1 ∈ H.
The above two conditions can be written as a, b ∈ H =⇒ a ∗ b−1 ∈ H.

Example 1.4 (i) (Z, +) is a subgroup of (Q, +).

(ii) (2Z, +) is a subgroup of (Z, +).
(iii) (R, +) is a subgroup of (C, +).
(iv) Let SL(n, R) = {A ∈ Mn×n (R) : det A = 1}. Then (SL(n, R), .) is a subgroup of (GL(n, R), .),
known as special linear group.

1.4 Ring
A non empty set R is said to form a ring with respect to two binary operation addition (+) and
multiplication (.) defined on it if the following condition satisfied
(i) (R, +) is a commutative group,
(ii) a.b ∈ R for all a, b ∈ R,
(iii) a.(b.c) = (a.b).c for all a, b, c ∈ R,
(iv) a.(b + c) = a.b + a.c and (b + c).a = b.a + c.a for all a, b, c ∈ R. (distributive law)
The ring is denoted by (R, +, .). The identity element of the group (R, +) is denoted by 0.
• The ring (R, +, .) is called commutative if a.b = b.a hold for all a, b ∈ R.
• If there is an element 1 ∈ R such that a.1 = 1.a = a for every a ∈ R, then R is called a ring
with unit element.
Md Firoz Ali 3

• R is said to be a ring without zero divisors if a.b = 0 =⇒ a = 0 or, b = 0.

• A commutative ring with unity is called an integral domain if it has no zero divisor.

Example 1.5 (i) (Z, +, .) is a commutative ring, with unit element.

(ii) (Q, +, .), (R, +, .), (C, +, .) are commutative rings.

(iii) (2Z, +, .) is a commutative ring without unit element.
(iv) (Mn×n (R), +, .) is a non-commutative ring with zero divisor.

1.5 Subring
Let (R, +, .) be a ring and S be a non-empty subset of R. If (S, +, .) is also a ring then (S, +, .) is
called a subring of (R, +, .).

Theorem 1.2 Let (R, +, .) be a ring. A non-empty subset S of R forms a subring of (R, +, .) if
and only if
(i) a, b ∈ S =⇒ a − b ∈ S and

(ii) a, b ∈ S =⇒ a.b ∈ S.

Example 1.6 (i) (Z, +) is a subring of (Q, +).

(ii) (2Z, +) is a subring of (Z, +).

1.6 Field
Let F be a non empty set and + and . be two binary operation on F . Then F is a field if F is
a commutative ring with unity and every non-zero element of F has a multiplicative inverse. In
other words, if it satisfies the following properties
(i) (F, +) is a commutative group, where the additive identity element is denoted by 0 and the
additive inverse of a is denoted by −a.

(ii) (F, .) is a commutative group with the exception that 0 has no multiplicative inverse, where
the multiplicative identity element is denoted by 1 and the multiplicative inverse of a is
denoted by a−1 .
(iii) Distributivity of multiplication over addition: a.(b + c) = (a.b) + (a.c).

The field is denoted by (F, +, .) or simply by F .

Example 1.7 (i) (Z, +, .) is not a field.

(ii) (Q, +, .), (R, +, .), (C, +, .) are fields.

√ √ √
(iii) Let Q[ 2] = {a + b 2 : a, b ∈ Q}. Then (Q[ 2], +, .) is a field.
(iv) (Mn×n (R), +, .) is not a field.

(v) (Z2 , +n , ·n ), (Z3 , +n , ·n ), (Z5 , +n , ·n ) are field. But (Z4 , +n , ·n ) is not a field.
Md Firoz Ali 4

2 Linear Algebra
2.1 Vector Space
A vector space over a field F is a non-empty set V together with two operations + : V × V → V
(called vector addition) and . : F ×V → V (called scalar multiplication) which satisfies the following
properties

(i) α + β ∈ V for all α, β ∈ V

(ii) α + (β + γ) = (α + β) + γ for all α, β, γ ∈ V
(iii) there exists an element θ ∈ V (called zero vector or null vector) such that α + θ = θ + α = α
for all α ∈ G.

(iv) for each element α ∈ G, there exists an element −α ∈ G such that α + (−α) = (−α) + α = θ.
(v) α + β = β + α for all α, β ∈ V
(vi) c.α ∈ V for all c ∈ F and α ∈ V

(vii) c.(α + β) = c.α + c.β for all c ∈ F and α, β ∈ V

(viii) (c + d).α = c.α + d.α for all c, d ∈ F and α ∈ V
(ix) c.(d.α) = (c.d).α for all c, d ∈ F and α ∈ V
(x) 1.α = α for all α ∈ V where 1 is the multiplicative identity element of the field F .

The vector space is denoted by (V (F ), +, .) or simply by V or V (F ). The elements of V are called

vectors whereas the elements of F are called scalars.

Example 2.1 (i) R2 = {(x1 , x2 ) : x1 , x2 ∈ R} is a vector space over the field R where the vector
addition + and scalar multiplication . is defined respectively as below

(x1 , x2 ) + (y1 , y2 ) = (x1 + y1 , x2 + y2 ) & c(x1 , x2 ) = (cx1 , cx2 ).

(ii) Rn = {(x1 , x2 , . . . , xn ) : x1 , x2 , . . . , xn ∈ R} is a vector space over the field R.

(iii) The set V = {xî + y ĵ + z k̂ : x, y, z ∈ R} is vector space over the field R.

(iv) The set C of complex numbers is a vector space over the field R.

(v) The set Mm×n (R) is a vector space over the field R.
(vi) If Pn is the set of all polynomials of degree less than or equal to n with coefficients in R
(respectively C) then Pn is a vector space over the field R (respectively C).

Theorem 2.1 Let V be a vector space over the field F . Then the following hold
(i) 0.α = θ for all α ∈ V

(ii) c.θ = θ for all c ∈ F

(iii) −1.α = −α for all α ∈ V
(iv) c.α = θ implies either c = 0 or α = θ.
Md Firoz Ali 5

2.2 Subspace
Let V be a vector space over the field F and W be a non-empty subset of V . If W is a vector
space under the operations of V , then it is called a vector subspace of V .
Theorem 2.2 Let V be a vector space over the field F and W be a non-empty subset of V . Then
W is a subspace of V if and only if the following two conditions hold
(i) α, β ∈ W =⇒ α + β ∈ W
(ii) α ∈ W, c ∈ F =⇒ c.α ∈ W .
The above two condition can be also written as
α, β ∈ W & c, d ∈ F =⇒ c.α + d.β ∈ W.
Example 2.2 Let V be a vector space over a field F . Then W = V is itself a subspace of V (called
the improper subspace of V ). Also, W = {θ} is a subspace of V (called the trivial subspace of V ).
Example 2.3 Let W1 = {(x, y, z) ∈ R3 : x = 0}. Then show that W1 is a subspace of R3 .
Solution: First we note that W1 is non-empty as θ = (0, 0, 0) ∈ W1 . Let α = (x1 , y1 , z1 ), β =
(x2 , y2 , z2 ) ∈ W1 and c ∈ R. Then
α + β = (x1 + x2 , y1 + y2 , z1 + z2 ) ∈ W1 as x1 + x2 = 0
c.α = (cx1 , cy1 , cz1 ) ∈ W1 as cx1 = 0
3
Thus W1 is a subspace of R .
Example 2.4 (i) Let W2 = {(x, y, z) ∈ R3 : y = 0}. Then show that W2 is a subspace of R3 .
(ii) Let W3 = {(x, y, z) ∈ R3 : z = 0}. Then show that W3 is a subspace of R3 .
(iii) Let W4 = {(x, y, z) ∈ R3 : x + y + z = 0}. Then show that W4 is a subspace of R3 .
(iv) Let W5 = {(x, y, z) ∈ R3 : x2 + y 2 + z 2 = 1}. Then show that W5 is not a subspace of R3 .
Hint: The null vector θ = (0, 0, 0) 6∈ W5 .
Theorem 2.3 Let W1 and W2 be two subspace of a vector space V over a field F . Then W1 ∩ W2
is a subspace of V . But W1 ∪ W2 need not be a subspace of V .
Example 2.5 Let
W1 = {(x, y, z) ∈ R3 : x = 0} & W2 = {(x, y, z) ∈ R3 : y = 0}.
Then W1 and W2 are subspaces of R3 . Let W = W1 ∪ W2 . Clearly, α = (0, 1, 1), β = (1, 0, 1) ∈ W
but α + β = (1, 1, 2) 6∈ W . Thus W is not a subspace of R3 .
Example 2.6 Let V be a vector space over a field F and α ∈ V . Then W = {c.α : c ∈ F } is a
subspace of V .
Solution: First we note that W is non-empty as θ = 0.α ∈ W .
Let c1 .α, c2 .α ∈ W and d ∈ F . Then
c1 .α + c2 .α = (c1 + c2 )α ∈ W as c1 + c2 ∈ F
d.c1 α = (d.c1 )α ∈ W
Thus W is a subspace of V .
Example 2.7 Let V be a vector space over a field F and α, β ∈ V . Then W = {c.α+d.β : c, d ∈ F }
is a subspace of V .
Example 2.8 Let V be a vector space over a field F and α1 , α2 , . . . , αk ∈ V . Then W = {c1 α1 +
c2 α2 + . . . + ck αk : c1 , c2 , . . . , ck ∈ F } is a subspace of V .
Solution: First we note that W is non-empty as θ = 0.α1 + 0.α2 + . . . + 0.αk ∈ W .
Let u = c1 α1 + c2 α2 + . . . + ck αk , v = d1 α1 + d2 α2 + . . . + dk αk ∈ W and a ∈ F . Then
u + v = (c1 + d1 )α1 + (c2 + d2 )α2 + . . . + (ck + dk )αk ∈ W as ci + di ∈ F
a.u = (a.c1 )α1 + (a.c2 )α2 + . . . + (a.ck )αk ∈ W
Thus W is a subspace of V .
Md Firoz Ali 6

2.3 Linear Combination and Linear Span

Linear Combination: Let V be a vector space over a field F and α1 , α2 , . . . , αr ∈ V (finite
number of vectors). A vector β ∈ V is said to be a linear combination of the vectors α1 , α2 , . . . , αr
if β can be expressed as
β = c1 α1 + c2 α2 + . . . + cr αr
for some scalars ci ∈ F .

Example 2.9 Let V be a vector space over the field R and α, β, γ ∈ V . Then u1 = α+2β +γ, u2 =
α + 2β, u3 = β + 5γ are linear combination of α, β, γ.

Linear span: Let V be a vector space over a field F and S be a non empty subset of V (maybe
finite or infinite). The linear span of S is

L(S) = {c1 α1 + c2 α2 + . . . + ck αk : ci ∈ F, αi ∈ S}.

If S is the finite set S = {α1 , α2 , . . . , αn } then L(S) is the set of all linear combinations of the
vectors α1 , α2 , . . . , αn .

Remark: In general, L(S) is a subspace of V and it is the smallest subspace containing S.

Example 2.10 In R3 , let α = (4, 3, 5), β = (0, 1, 3), γ = (2, 1, 1), δ = (4, 2, 2). Examine if
(i) α is a linear combination of β, γ.
(ii) β is a linear combination of γ, δ.

Solution: (i) Let α = cβ + dγ for some c, d ∈ R. Then

(4, 3, 5) = c(0, 1, 3) + d(2, 1, 1)

=⇒ 4 = 0 + 2d, 3 = c + d, 5 = 3c + d

=⇒ d = 2, c = 1

Thus α = β + 2γ and so α is a linear combination of β, γ.

(ii) Let β = cγ + dδ for some c, d ∈ R. Then

(0, 1, 3) = c(2, 1, 1) + d(4, 2, 2)

=⇒ 2c + 4d = 0, c + 2d = 1, c + 2d = 3

This system of equations are inconsistent. Thus β is not a linear combination of γ, δ.

Example 2.11 Determine the subspace of R3 spanned by α = (1, 2, 3), β = (3, 1, 0). Examine if
(i) γ = (2, 1, 3) is in the subspace.
(ii) δ = (−1, 3, 6) is in the subspace.

Solution: Let S = {α, β}. Then

L(S) = {cα + dβ : c, d ∈ R} = {c(1, 2, 3) + d(3, 1, 0) : c, d ∈ R}.

If γ = (2, 1, 3) ∈ L(S) then there exist c, d ∈ R such that

(2, 1, 3) = c(1, 2, 3) + d(3, 1, 0)

=⇒ 2 = c + 3d, 1 = 2c + d, 3 = 3c

This system of equation is inconsistent and so γ 6∈ L(S).

If δ = (−1, 3, 6) ∈ L(S) then there exist c, d ∈ R such that

(−1, 3, 6) = c(1, 2, 3) + d(3, 1, 0)

=⇒ − 1 = c + 3d, 3 = 2c + d, 6 = 3c =⇒ c = 2, d = −1

Thus δ = 2α − β and so δ ∈ L(S).

Md Firoz Ali 7

Theorem 2.4 Let S and T be two non-empty finite subset of a vector space V over a field F .
Then

(i) S ⊂ T =⇒ L(S) ⊂ L(T );

(ii) if each element of S is a linear combination of the vectors of T then L(S) ⊂ L(T ).

Example 2.12 Let S = {α, β, γ} and T = {α, β, α + β, β + γ. Show that L(S) = L(T ).

Solution: Clearly, each element of T is a linear combination of elements of S. Thus L(T ) ⊂

L(S). Also note that,

α = 1.α + 0.β + 0.(α + β) + 0.(β + γ)

β = 0.α + 1.β + 0.(α + β) + 0.(β + γ)
γ = 0.α + (−1).β + 0.(α + β) + 1.(β + γ).

Thus, each element of S is a linear combination of elements of T . Hence L(S) ⊂ L(T ). Therefore
L(S) = L(T ).

2.4 Linearly Dependent and Linearly Independent

A finite set of vectors {α1 , α2 , . . . , αn } of a vector space V over a field F is said to be Linearly
Dependent (L.D.) in V if there exist scalars c1 , c2 , . . . , cn not all zero in F such that

c1 α1 + c2 α2 + . . . + cn αn = θ. (2.1)

The set of vectors {α1 , α2 , . . . , αn } is said to be Linearly Independent (L.I.) in V if the equality
(2.1) is satisfied only when c1 = c2 = . . . = cn = 0.

An arbitrary set S of vectors of a vector space V over a field F is said to be L.D. if there exists
a finite subset of S which is L.D. in V. If S is not L.D. then it is called a L.I. set.

Remarks:
(i) A superset of a L.D. set is L.D.

(ii) A subset of L.I. set is L.I.

(iii) Any set which contains the null vector is linearly dependent set as cθ = θ for c 6= 0.

Example 2.13 Examine if the set of vectors S = {α = (2, 1, 1), β = (1, 2, 2), γ = (1, 1, 1)} is L.I.
in R3 . Do the same for T = {α = (1, 2, 2), β = (2, 1, 2), γ = (2, 2, 1)}.

Solution: Here S = {α = (2, 1, 1), β = (1, 2, 2), γ = (1, 1, 1)}. Let c1 α + c2 β + c3 γ = θ where
c1 , c2 , c3 ∈ R.

=⇒ c1 (2, 1, 1) + c2 (1, 2, 2) + c3 (1, 1, 1) = (0, 0, 0)

=⇒ 2c1 + c2 + c3 = 0, c1 + 2c2 + c3 = 0, c1 + 2c2 + c3 = 0.

This is a system of homogeneous equations and its coefficient determinant is

2 1 1
1 2 1 = 0.
1 2 1

Thus it has many non zero solutions. Moreover,

c1 c2 c3
= = = k.
−1 −1 3
Thus (c1 , c2 , c3 ) = k(−1, −1, 3). Therefore S is L.D.
Md Firoz Ali 8

Here T = {α = (1, 2, 2), β = (2, 1, 2), γ = (2, 2, 1)}. Let

c1 α + c2 β + c3 γ = θ, c1 , c2 , c3 ∈ R
=⇒ c1 + 2c2 + 2c3 = 0, 2c1 + c2 + 2c3 = 0, 2c1 + 2c2 + c3 = 0.

This is a system of homogeneous equation and coefficient determinant is

1 2 2
2 1 2 = 5 6= 0.
2 2 1

Thus it has an unique solution and the solution is c1 = c2 = c3 = 0. Hence T is L.I.

Remark: Two vectors α, β in a vector space V are L.D. if at least one of them is a scalar
multiple of the other.

2.5 Basis and Dimension

Basis: Let V be a vector space over the field F. A subset B of V is called a basis for V if
(i) B is L.I.
(ii) B generates V i.e L(B) = V.

Dimension: The vector space V is called finite dimensional if B contains finite number of
element. If V is not finite dimensional then it is called infinite dimensional. The dimension of V,
denoted by dim V is the number of elements in B. The null space is said to be of dimension 0.

Example 2.14 Show that the set E = {1 = (1, 0), 2 = (0, 1)} is a basis of R2 .

Solution: Let c1 1 + c2 2 = θ for some c1 , c2 ∈ R.

Then (c1 , c2 ) = (0, 0) =⇒ c1 = 0, c2 = 0 and so E is L.I.

Let v = (a, b) ∈ R2 . Then v = a(1, 0) + b(0, 1) = a1 + b2 and so R2 ⊆ L(E).

Also E ⊆ R2 and L(E) is the smallest subspace containing E. Then L(E) ⊆ R2 . Thus R2 =
L(E). Therefore E is a basis for R2 and dim R2 = 2.

Example 2.15 Show that the set set E = {1 = (1, 0, 0), 2 = (0, 1, 0), 3 = (0, 0, 1)} is a basis of
R3 .

Example 2.16 Show that the set E = {1 = (1, 0, 0, ..., 0), 2 = (0, 1, 0, ..., 0), ..., n = (0, 0, 0, ..., 1)}
is a basis of Rn . The basis E is called the standard basis of Rn .

Example 2.17 Let E = {E11 , E12 , . . . , E1m , E21 , E22 , . . . , E2m , . . . , Emn } where Eij is an m × n
matrix having 1 as the ij-th element and 0 elsewhere. Then E is a basis of the vector space
Mm×n (R) of all m × n. Also, dim Mm×n (R) = mn.

Example 2.18 Show that the set E = {1, x, x2 , ..., xn } is a basis of the vector space Pn .

Example 2.19 Show that the set E = {1, i} is a basis of the vector space C over the field R.

Theorem 2.5 Let V be a vector space of dimension n over a field F. Then

(i) Any set of L.I. vectors of V contains at most n vectors.
(ii) Any L.I. set of n vectors of V is a basis of V.
(iii) Any subset of n vectors of V that generates V is a basis of V.

Example 2.20 Prove that S = {α = (1, 0, 1), β = (0, 1, 1), γ = (1, 1, 0)} is a basis of R3 .
Md Firoz Ali 9

Solution: Here

c1 α + c2 β + c3 γ = θ, where c1 , c2 , c3 ∈ R3
=⇒ c1 + c3 = 0, c2 + c3 = 0, c1 + c2 = 0
=⇒ c1 = c2 = c3 = 0.

Thus S is L.I. Since R3 is a vector space of dimension 3 and S contains 3 L.I. vectors, S is a basis
of R3 .

Example 2.21 Let V be a real vector space with {α, β, γ} as basis. Prove that S = {α + β + γ, β +
γ, γ} is also a basis of V.

Solution: Let c1 (α + β + γ) + c2 (β + γ) + c3 γ = 0 where c1 , c2 , c3 ∈ R. Then

c1 α + (c1 + c2 )β + (c1 + c2 + c3 )γ = 0
=⇒ c1 = 0, c1 + c2 = 0, c1 + c2 + c3 = 0

as {α, β, γ} is L.I. So we get c1 = c2 = c3 = 0.

Thus S is L.I. Since V is a vector space of dimension 3 and S contain 3 L.I. vectors, S is a basis
of V.

Example 2.22 Find a basis and the dimension of the subspace W = {(x, y, z) ∈ R3 : x+y+z = 0}
of R3 .

Solution: Here we assume W is a subspace of R3 . Let v = (a, b, c) ∈ W. Then a + b + c = 0.

Therefore,

v = (a, b, −a − b) = a(1, 0, −1) + b(0, 1, −1)

Let α = (1, 0, −1), β = (0, 1, −1). Then v ∈ L{α, β} and so W ⊆ L{α, β}.

Since α, β ∈ W and L{α, β} is the smallest subspace containing α, β, then L{α, β} ⊆ W. Thus
W = L{α, β}.

Here α and β are L.I. as none of them is a scalar multiple of other. Thus {α, β} is a basis for
W and dim W = 2.

Example 2.23 Let S = {(x, y, z, w) ∈ R4 : 2x + y + 3z + w = 0} and T = {(x, y, z, w) ∈ R4 } :

x + 2y + z + 3w = 0. Find dim S ∩ T.

Solution: Clearly,

S ∩ T = {(x, y, z, w) ∈ R4 : 2x + y + 3z + w = 0, x + 2y + z + 3w = 0}.

is a subspace of R4 . Let v = (a, b, c, d) ∈ (S ∩ T ).Then

2a + b + 3c + d = 0, a + 2b + c + 3d = 0
=⇒ d = −2a − b − 3c, a + 2b + c − 6a − 3b − 9c = 0
=⇒ 5a + b + 8c = 0
1
=⇒ c = − (5a + b).
8
Therefore, d = −2a − b + 38 (5a + b) = − 18 a − 85 b. Thus

5 1 1 5 a b
v =(a, b, − a − b, − a, − ) = (8, 0, −5, −1) + (0, 8, −1, −5)
8 8 8 8 8 8
a b
= α + β ∈ L{α, β},
8 8
where α = (8, 0, −5, −1), β = (0, 8, −1, −5). Therefore, S ∩ T ⊆ L{α, β}.
Clearly α, β ∈ S ∩ T . Since S ∩ T is a subspace containing {α, β}, L{α, β} ⊆ S ∩ T. Therefore
L{α, β} = S ∩ T , Clearly α and β are L.I. Thus {α, β} is a basis for S ∩ T, and dim S ∩ T = 2.
Md Firoz Ali 10

2.6 Rank of a Matrix

A non-zero matrix A of order m × n is said to be of rank r when
(i) it has at least one non-zero minor of order r,
(ii) every minor of order higher than r is zero.
The rank of zero matrix is 0. Rank of a matrix is denoted by ρ(A) or r(A) or rank(A). For a
non-zero m × n matrix A, we have 0 < ρ(A) ≤ min{m, n} and ρ(A) = ρ(AT ).
   
1 0 3 2 3 −1 1
Example 2.24 Find the rank of the matrices A = 4 −1 5 and B = 3 0 4 2.
2 0 6 6 9 −3 3

Solution: We can easily verify that det(A) = 0. But

1 0
6= 0.
4 −1

Therefore ρ(A) = 2. Similarly we can show that ρ(B) = 2.

 
1 1 1 1
Example 2.25 Find all real λ for which rank of A is 2, where A = 1 2 −1 λ .
5 7 1 λ2

Solution:
1 1 1
1 −1 λ =1(−λ2 − λ) − 1(λ2 − 5λ) + 1(1 + 5)
5 1 λ2
= − 2λ2 + 4λ + 6.

Since ρ(A) = 2, we have

− 2λ2 + 4λ + 6 = 0
=⇒ (λ − 3)(λ + 1) = 0
=⇒ λ = −1, 3.

Elementary Operation: An elementary operation on a matrix A is an operation of the

following types
(i) Interchange of two rows or columns of A. Notation: Rij (Cij ).
(ii) Multiplication of a row or column by a non-zero scalar c. Notation: kRi (kCi ).
(iii) Addition of a scalar multiple of one row or column to another row or column. Notation:
Ri + kRj (Ci + kCj ).

Equivalent Matrices: Two matrices A and B are said to be row equivalent or column equiv-
alent if one can be obtained from the other by elementary row or column operations respectively.
Two equivalent matrices have the same order and same rank.

Row reduced Echelon form: An m × n matrix A is called row-reduced echelon matrix if

(i) the first non-zero element in each non-zero row is 1 (called the leading 1)
(ii) each column containing the leading 1 of some row has all other element zero.
(iii) there is an integer r such that first r rows of A are non-zero and the remaining are all zero.
(iv) if the leading element of i-th row occurs in column ki then k1 < k2 < k3 ... < ki .

Theorem 2.6 The number of non-zero rows in row-reduced echelon matrix is the rank of that
matrix.
Md Firoz Ali 11

 
0 1 −3 −1
1 0 1 1
Example 2.26 Find the rank of the matrix A =  3
.
1 0 2
1 1 −2 0
     
1 0 1 1 1 0 1 1 1 0 1 1
R12 0 1 −3 −1 R 3 −3R 1
0 1 −3 −1 R 3 −R2
0 1 −3 −1
Solution: A −−→   −−−−−→   −−−−−→  .
3 1 0 2  R4 −R1 0 1 −3 −1 R4 −R2 0 0 0 0
1 1 −2 0 0 1 −3 −1 0 0 0 0
Therefore ρ(A) = 2.
 
1 2 3 4
Example 2.27 Find the rank of the matrix A = 2 1 4 3 .
3 0 5 −10
Solution:
     
1 2 3 4 − 13 R2
1 2 3 4 1 0 5/3 2/3
R2 −2R1 R3 +6R2
A −−−−−→ 0 −3 −2 −5  −−−−→ 0 1 5/3  −−−−−→ 0 1 2/3
2/3 5/3 
R3 −3R1 R1 −2R2
0 −6 −4 −22 0 −6 −4−22 0 0 0 −12
   
− 1 R3
1 0 5/3 2/3 R − 2 R 1 0 5/3 0
1 3 3
−−12
−−→ 0 1 2/3 5/3 −−−−− −→ 0 1 2/3 0 .
R2 − 35 R3
0 0 0 1 0 0 0 1
Therefore, ρ(A) = 3.

 
2 0 4 2
3 2 6 5
Example 2.28 Find the rank of the matrix A = 
5 2 10 7


0 3 2 5
     
1 0 2 1 1 0 2 1 1 0 2 1
1/2R1 3 2 6 3 R2 −3R1 0 2 0 2 1/2R2 0 1 0 1 R3 −2R2
Solution: A −−−−→  5 2 10 7 −
 −−−−→   −−−−→  −−−−−→
R3 −5R1 0 2 0 2 0 2 0 2 R4 −3R2
0 3 2 5 0 3 2 5 0 3 2 5
       
1 0 2 1 1 0 2 1 1 0 2 1 1 0 0 −1
0 1 0 1 R3 4 0 1 0 1 1/2R3 0 1 0 1 R1 −2R3 0 1 0 1
0 0 0 0 −−→ 0 0 2 2 −−−−→ 0 0 1 1 −−−−−→ 0 0
       .
1 1
0 0 2 2 0 0 0 0 0 0 0 0 0 0 0 0

Therefore, ρ(A) = 3.

 
1 1 2
Example 2.29 Find the inverse of A = 2 4 4 by elementary operation.
3 3 7
Solution:
   
1 1 2 : 1 00 1 1 2 : 1 0 0
R −2R1
[A : I3 ] = 2 4 4 : 0 10 −−2−−−→ 0 2 0 : −2 1 0
R3 −3R1
3 3 7 : 0 01 0 0 1 : −3 0 1
   
1 1 2 : 1 0 0 1 0 2 : 2 −1/2 0
1/2R2 R1 −R2
−−−−→ 0 1 0 : −1 1/2 0 −−− −−→ 0 1 0 : −1 1/2 0
0 0 1 : −3 0 1 0 0 1 : −3 0 1
 
1 0 0 : 8 −1/2 −2
R −2R3
−−1−−−→ 0 1 0 : −1 1/2 0  = [I3 : A−1 ].
0 0 1 : −3 0 1
 
8 −1/2 −2
Therefore, A−1 = −1 1/2 0 .
−3 0 1
Md Firoz Ali 12

2.7 System of Linear equations

A system of m linear equations in n unknowns x1 , x2 , . . . , xn is given as
a11 x1 + a12 x2 + · · · + a1n xn = b1
a21 x1 + a22 x2 + · · · + a2n xn = b2
··· ··· ··· ··· ···
am1 x1 + am2 x2 + · · · + amn xn = bm
where aij ’s and bi ’s are elements of a field F (in general, we take the field as the set of real numbers
R or as the set of complex numbers C). The above system of linear equations can be written in
matrix form as follows
   
x1 b1
 x2   b2 
AX = B, where, A = (aij )m×n , X =  · · · , B = · · · .
   (2.2)
xn bm
For the system (2.2), the matrix A = (aij )m×n is called the coefficient matrix and the matrix
 
a11 a12 · · · a1n b1
 a21 a22 · · · a2n b2 
A = Ab = ···

· · · · · · · · · · · ·
am1 am2 · · · amn bm
is called the augmented matrix. The system AX = B is called homogeneous system if B = O;
otherwise, a non-homogeneous system. An ordered pair (c1 , c2 , . . . , cn ) where ci ∈ F is said to be
a solution of the system (2.2) if each equation of the system (2.2) is satisfied for x1 = c1 , x2 =
c2 , . . . , xn = cn .

For a system of linear equations, there are three possibilities, (i) unique solution, (ii) no solution,
(iii) infinitely many solutions. Below we have illustrated the above three cases by suitable examples.
Example 2.30 Consider the system of equations
x1 + x2 = 3
x1 − x2 = 1.
For this system (2, 1) is a solution and there is no other solution of the system.
Example 2.31 Consider the system of equations
x1 + 2x2 = 5
3x1 + 6x2 = 10.
This system has no solution.
Example 2.32 Consider the system of equations
2x1 + x2 = 0
4x1 + 2x2 = 0.
For this system (1, −2) is a solution. Moreover, (2, −4), (3, −6) are also solutions. In fact, k(1, −2)
is a solution for each real number k. Thus this system has many solutions.
Definition 2.1 The system of equations (2.2) is said to be consistent if it has at least one
solution; otherwise, it is called inconsistent.
Existence and number of solution(s) of the system AX = B where A is an m × n
matrix:

There are two cases.

Case-1: The system is consistent if and only if rank of A = rank of A. In this case, the
following are possible.
Md Firoz Ali 13

(i) If rank of A = rank of A = n (number of unknowns) then the system has a unique solution.
(ii) If rank of A = rank of A < n then the system has infinitely many solutions.

Case-2: The system is inconsistent if and only if rank of A 6= rank of A.

Remark: The Homogeneous system AX = O, where A is an m × n matrix has

(i) always zero solution, called the trivial solution.

(ii) a unique solution (i.e., only zero solution) iff rank of A = n (the number of unknowns).
When A is square matrix, i.e., if number of unknowns is equal to number of equations, the
corresponding homogeneous system has unique solution iff det A 6= 0.
(iii) a non-zero solution, called non-trivial solution iff rank of A < n (then number of unknowns).
When A is square matrix, i.e., if number of unknowns is equal to number of equations, the
corresponding homogeneous system has non-trivial solution iff det A = 0.

Definition 2.2 Two systems AX = B and A0 X = B 0 are said to be equivalent systems if the
augmented matrices A and A0 are row equivalent.

Theorem 2.7 Let AX = B and A0 X = B 0 be two equivalent systems.

(i) If α is a solution of one of the systems then α is also a solution of the other system.
(ii) If one of the system is inconsistent then the other system is also inconsistent.

Theorem 2.8 Let AX = O be a homogeneous system in n unknowns, where A is an m×n matrix.

(i) The set X(A) of all solutions of the system form a vector space, called the solution space of
the system.

(ii) rank of A+ rank of X(A) = n (here, rank of X(A) = dim X(A)).

Remark: The solutions of a consistent non-homogeneous system AX = B do not form a vector

space because (0, 0, . . . , 0) is not a solution.

Theorem 2.9 If the non-homogeneous system AX = B possesses a solution X0 then all solution
of the system are obtained by adding X0 to the general solution of the associated homogeneous
system AX = O.

Example 2.33 Solve, if possible, the system of equations

x1 + x2 = 4
x2 − x3 = 1
2x1 + x2 + 4x3 = 7.

Solution: This is a system of non-homogeneous equations. Let us apply elementary row

operations on the augmented matrix A to reduce it to a row-echelon matrix.
     
1 1 0 4 1 1 0 4 1 0 1 3
R −2R R1 −R2
A = 0 1 −1 1 3−→ 1 0 1 −1 1  −→ 0 1 −1 1
R3 +R2
2 1 4 7 0 −1 4 −1 0 0 3 0
   
1 1 0 1 3 1 0 0 3
3 R3 1 −R3
−→ 0 1 −1 1 R−→ 0 1 0 1
R2 +R3
0 0 1 0 0 0 1 0

Since rank of A = rank of A = 3 = number of unknowns, the system is consistent and it has unique
solution. Hence the given system of equations is equivalent to x1 = 3, x2 = 1, x3 = 0 and therefore
the solution is (3, 1, 0).
Md Firoz Ali 14

Example 2.34 Solve, if possible, the system of equations

x1 + 3x2 + x3 = 0
2x1 − x2 + x3 = 0.
Solution: This is a system of homogeneous equations. Let us apply elementary row operations
on the augmented matrix A to reduce it to a row-echelon matrix.
     
1 3 1 0 R2 −2R1 1 3 1 0 − 71 R2 1 3 1 0
A= −→ −→
2 −1 1 0 0 −7 −1 0 0 1 1/7 0
 
R1 −3R2 1 0 4/7 0
−→
0 1 1/7 0

Since rank of A = 2 = rank of A < 3, the system is consistent and has infinitely many solutions.
Hence the system of equations is equivalent to
4
x1 + x3 = 0
7
1
x2 + x3 = 0.
7
Let x3 = c, where c is any real number. Then x1 = − 74 c and x2 = 17 c. Thus the solution is
(− 47 c, − 17 c, c) = − 17 c(4, 1, −7) i.e., k(4, 1, −7) where k is any real number.

Remark: In the previous example, the set X(A) = {k(4, 1, −7) ∈ R3 : k ∈ R} form a vector
space over R. Moreover, rank of A = 2 and rank of X(A) = dim X(A) = 1. Thus, rank of A+
rank of X(A) = 3.
Example 2.35 Solve, if possible, the system of equations
x1 + 2x2 − x3 = 10
−x1 + x2 + 2x3 = 2
2x1 + x2 − 3x3 = 2.
Solution: This is a system of non-homogeneous equations. Let us apply elementary row
operations on the augmented matrix A to reduce it to a row-echelon matrix.
     
1 2 −1 10 1 2 −1 10 1 R 1 2 −1 10
R2 +R1 3 2
A = −1 1 2 2  −→ 0 3 1 12  −→ 0 1 1/3 4 
R3 −2R1
2 1 −3 2 0 −3 −1 −18 0 −3 −1 −18
     
1 0 −5/3 2 − 1 R 1 0 −5/3 2 1 0 −5/3 0
R1 −2R2 −2R3
−→ 0 1 1/3 4  −→6 2
0 1 1/3 4 R1−→ 0 1 1/3 0 .
R3 +3R2 R2 −4R3
0 0 0 −6 0 0 0 1 0 0 0 1

Since rank of A(= 2) and rank of A(= 3) are not equal, the system is inconsistent.
Example 2.36 Solve, if possible, the system of equations
x1 + 2x2 − x3 = 10
−x1 + x2 + 2x3 = 2
2x1 + x2 − 3x3 = 8.
Solution: This is a system of non-homogeneous equations. Let us apply elementary row
operations on the augmented matrix A to reduce it to a row-echelon matrix.
     
1 2 −1 10 1 2 −1 10 1 R 1 2 −1 10
R2 +R1 3 2
A = −1 1 2 2  −→ 0 3 1 12  −→ 0 1 1/3 4 
R3 −2R1
2 1 −3 8 0 −3 −1 −12 0 −3 −1 −12
 
1 0 −5/3 2
R1 −2R2
−→ 0 1 1/3 4
R3 +3R2
0 0 0 0
Md Firoz Ali 15

Since rank of A = 2 = rank of A < 3, the system is consistent and has infinitely many solutions.
Hence the system of equations is equivalent to
5
x1 − x3 = 2
3
1
x2 + x3 = 4.
3
Let x3 = c, where c is any real number. Then x1 = 2 + 53 c and x2 = 4 − 13 c. Thus the solu-
tion is (2+ 35 c, 4− 13 c, c) = (2, 4, 0)+ 13 c(5, −1, 3) i.e., (2, 4, 0)+k(5, −1, 3) where k is any real number.

Remark: In previous Example, (2, 4, 0) is a particular solution of the system and k(5, −1, 3)
is the general solution of the associated homogeneous system AX = O.
Example 2.37 Solve, if possible, the system of equations
x1 + 2x2 + x3 − 3x4 = 1
2x1 + 4x2 + 3x3 + x4 = 3
3x1 + 6x2 + 4x3 − 2x4 = 4.
Solution: This is a system of non-homogeneous equations. Let us apply elementary row
operations on the augmented matrix A to reduce it to a row-echelon matrix.
     
1 2 1 −3 1 1 2 1 −3 1 1 2 0 −10 0
R −2R R1 −R2
A = 2 4 3 1 3 2−→ 1 0 0 1 7 1 −→ 0 0 1 7 1
R3 −3R1 R3 −R2
3 6 4 −2 4 0 0 1 7 1 0 0 0 0 0
Since rank of A = 2 = rank of A < 3, the system is consistent and has infinitely many solutions.
Hence the system of equations is equivalent to
x1 + 2x2 − 10x4 = 0
x3 + 7x4 = 1.
Let x4 = c, x2 = d, where c, d are any real numbers. Then x3 = 1 − 7c and x1 = 10c − 2d. Thus
the solution is (10c − 2d, d, 1 − 7c, c) = (0, 0, 1, 0) + c(10, 0, −7, 1) + d(−2, 1, 0, 0) where c, d are any
real numbers.

Remark: In previous Example, (0, 0, 1, 0) is a particular solution of the system and c(10, 0, −7, 1)+
d(−2, 1, 0, 0) is the general solution of the associated homogeneous system AX = O.
Example 2.38 Determine the conditions for which the system of equations
x+y+z =1
x + 2y − z = b
5x + 7y + az = b2 .
admits of (i) only one solution, (ii) no solution, (iii) infinitely many solutions.
Solution: The system has unique solution if rank of A = rank of A = 3. Now rank of A =
rank of A = 3 if det A 6= 0. Now,
1 1 1
det A = 1 2 −1 = a − 1
5 7 a
Thus, if a 6= 1 then the system has unique solution. If a = 1 then the system has either no solution
or infinitely many solutions. For a = 1, the augmented matrix A is
     
1 1 1 1 1 1 1 1 1 0 3 −b + 2
R2 −R1 1 −R2
A = 1 2 −1 b  −→ 0 1 −2 b − 1  R−→ 0 1 −2 b−1 
2 R3 −5R1 2 R3 −2R2
5 7 1 b 0 2 −4 b − 5 0 0 0 b2 − 2b − 3
If b2 − 2b − 3 6= 0 then rank of A = 2, rank of A = 3 and so the system is inconsistent. Thus,
if a = 1, b 6= −1, 3, the system has no solution.

If b2 − 2b − 3 = 0 then rank of A = rank of A = 2 and so the system is consistent. Thus, if

a = 1, b = −1 or if a = 1, b = 3, the system has many solutions.
Md Firoz Ali 16

2.8 Eigen values and eigen vectors

• Let A = (aij )n×n be a square matrix of order n over the field F .
• Then ΨA (λ) = det(A − λIn ) is called the characteristic polynomial of A.

• The equation ΨA (λ) = det(A − λIn ) = 0 is called the characteristic equation of A. On

expansion, we get

a11 − λ a12 ··· a1n

a21 a22 − λ ··· a2n
det(A − λIn ) = =0
··· ··· ··· ···
an1 an2 ··· ann − λ
=⇒ c0 λn + c1 λn−1 + . . . + cn = 0,

where

c0 = (−1)n , cr = (−1)n−r [sum of the principal minors of A of order r].

In particular, c1 = (−1)n−1 (a11 + a22 + . . . + ann ) = (−1)n−1 trace A and cn = det A.

• The roots of the characteristic equation ΨA (λ) = 0 of A are called the characteristic values
or eigen values of A.

• A root of ΨA (λ) = 0 of multiplicity r is said to be an r-fold eigen value of A.

• Note that eigen values of a matrix A over the field F may not belong to the field F .
• A non-zero vector X is said to be a characteristic vector or eigen vector of A if there
exists a scalar λ ∈ F such that AX = λX holds.

• Let X be an eigen vector of A. Then there exists a scalar λ ∈ F such that

AX = λX =⇒ (A − λIn )X = O.

This is homogeneous system of n equations in n unknowns. Since this system has a non-zero
solution, then det(A − λIn ) = 0. Thus λ is an eigen value of A.

Theorem 2.10 Let A be an n × n matrix over a field F .

(i) To an eigen vector of A there corresponds a unique eigen value of A.

(ii) To an eigen value λ ∈ F of A there corresponds at least one eigen vector of A. If an eigen
value λ 6∈ F then there exists no eigen vector corresponds to the eigen value λ.
 
1 1 1
Example 2.39 Find the eigen values and eigen vectors of the matrix A = −1 −1 −1.
0 0 1

Solution: The characteristic equation of A is det(A − λI3 ) = 0, i.e.,

1−λ 1 1
−1 −1 − λ −1 = 0 =⇒ λ2 (1 − λ) = 0 =⇒ λ = 0, 0, 1.
0 0 1−λ

Thus the eigen

  values are 0, 0, 1.
x
Let X = y  be an eigen vector corresponding to the eigen value λ = 0. Then AX = λX gives
z
AX = O, i.e.,

x+y+z =0
−x − y − z = 0
z = 0.
Md Firoz Ali 17

   
−c −1
Let y = c. Then x = −c, z = 0. Thus, the eigen vectors are X =  c  = c  1 , where c is
0 0
non-zero realnumber.

x
Let X = y  be an eigen vector corresponding to the eigen value λ = 1. Then AX = λX gives
z
(A − I3 )X = O, i.e.,

y+z =0
x + 2y + z = 0.

  
c 1
Let z = c. Then y = −c, x = c. Thus, the eigen vectors are X = −c = c −1 where c is
c 1
non-zero real number.
 
0 1 1
Example 2.40 Find the eigen values and eigen vectors of the matrix A = 1 0 1.
1 1 0

Solution: The characteristic equation of A is det(A − λI3 ) = 0, i.e.,

−λ 1 1
1 −λ 1 =0 =⇒ (λ + 1)2 (λ − 2) = 0 =⇒ λ = −1, −1, 2.
1 1 −λ

Thus the eigen values

 are −1, −1, 2.
x
Let X = y  be an eigen vector corresponding to the eigen value λ = −1. Then AX = λX
z
gives (A + I3 )X = O, i.e.,
x + y + z = 0.
     
−c − d −1 −1
Let z = c, y = d. Then x = −c − d. Thus, the eigen vectors are X =  d  = c  0  + d  1 
c 1 0
where c, d are any
 real numbers not simultaneously zero.
x
Let X = y  be an eigen vector corresponding to the eigen value λ = 2. Then AX = λX gives
z
(A − 2I3 )X = O, i.e.,

−2x + y + z = 0
x − 2y + z = 0
x + y − 2z = 0.
x y z
Solving the first two equations we get, = = = k (say). Then x = 3k, y = 3k, z = 3k which
3 3 3    
3k 1
also satisfies the third equation. Thus, the eigen vectors are X = 3k  = k1 1 where k1 = 3k
3k 1
is non-zero real number.
 
0 −1
Example 2.41 Find the eigen values and eigen vectors of the matrix A = .
1 0

Solution: The characteristic equation of A is det(A − λI2 ) = 0, i.e.,

−λ −1
=0 =⇒ λ2 + 1 = 0 =⇒ λ = i, −i.
1 −λ

Thus the eigen values are i, −i.

Md Firoz Ali 18

(Note: In this question, it is not mentioned whether the given matrix A is over the field R or,
over the field C. In general, it will be mentioned in the question when such situation arises. Below
we have shown both possibilities.)

Case-1: If we consider the given matrix A is over the field R, then the eigen values does not
belong to the field R. Thus, there is no eigen vector corresponding to the eigen values i, −i.

Case-2: If we consider the given matrix A is over the field C, then the eigen values belong to
the field C. Thus, corresponding to each eigen values of A there is an eigen vector.
 
x
Let X = be an eigen vector corresponding to the eigen value λ = i. Then AX = λX gives
y
(A + iI2 )X = O, i.e.,
ix − y = 0
x − iy = 0.
   
ic i
Let y = c. Then x = ic. Thus, the eigen vectors are X = =c where c is any non-zero
c 1
complex number.
 
1
Similarly, the eigen vectors corresponding to the eigen value λ = −i are X = c where c is
i
any non-zero complex number.

Theorem 2.11 (Properties of the eigen values and eigen vectors) (i) Any square matrix
A and its transpose At have the same eigen values.
(ii) The eigen values of a triangular matrix or a diagonal matrix are just the diagonal elements
of the matrix.
(iii) The sum of the eigen values of a matrix is the sum of the elements of the principal diagonal.
(iv) The product of the eigen values of a matrix A is equal to its determinant.
(v) If A is a singular matrix then 0 is an eigen value of A.
(vi) If λ is an eigen value of a non-singular matrix A, then 1/λ is an eigen value of A−1 .
(vii) If λ1 , λ2 , . . . , λn are the eigen values of a matrix A, then λm m m
1 , λ2 , . . . , λn are the eigen values
m
A (m being a positive integer).
(viii) If X1 , X2 , . . . , Xr are eigen vectors of an n × n matrix A corresponding to r distinct eigen
values λ1 , λ2 , . . . , λr respectively, then X1 , X2 , . . . , Xr are linearly independent.
(ix) The eigen vectors of an n × n matrix A over a field F corresponding to an eigen value
λ ∈ F , together with the zero-vector form a vector space, called the characteristic subspace
corresponding to λ.
(x) For an r-fold eigen value λ, r is called the algebraic multiplicity (A.M.) of λ and the rank
(dimension) of characteristic subspace corresponding to λ is called the geometric multiplicity
(G.M.) of λ.
(xi) It is easy to prove that for an eigen value λ, we have 1 ≤ geometric multiplicity
≤ algebraic multiplicity.
(xii) An eigen value λ, is said to be regular if algebraic multiplicity of λ is equal to its geometric
multiplicity.
Example 2.42 (i) In Example 2.39, A.M. of the eigen value 0 is 2, G.M. of the eigen value
0 is 1, A.M. of the eigen value 1 is 1, G.M. ofthe eigen value 1 is 1. The eigen vector
−1 1
 1  corresponding to λ = 0 and eigen vector −1 corresponding to λ = 1 are linearly
0 1
independent.
Md Firoz Ali 19

(ii) In Example 2.40, A.M. of the eigen value −1 is 2, G.M. of the eigen value −1 is 2, A.M. of
the eigen value 2 is 1, G.M. of the eigen value 2 is 1.

Theorem 2.12 (Cayley-Hamilton theorem) Every square matrix satisfies its own character-
istic equation, i.e., if the characteristic equation of an n × n matrix A is det(A − λIn ) = c0 λn +
c1 λn−1 + . . . + cn = 0 then
c0 An + c1 An−1 + . . . + cn In = O.
 
1 4
Example 2.43 Verify Cayley-Hamilton theorem for the matrix A = and then find A−1 .
2 3

Solution: The characteristic equation of A is det(A − λI2 ) = 0, i.e.,

1−λ 4
=0 =⇒ λ2 − 4λ − 5 = 0.
2 3−λ

Now,        
9 16 1 4 1 0 0 0
A2 − 4A − 5I = −4 −5 = = O.
8 17 2 3 0 1 0 0
Thus the matrix A satisfies its own characteristic equation. Hence Cayley-Hamilton theorem
verified. Again,
 
2 2 −1 −1 1 −3 4
A − 4A − 5I = O =⇒ 5I = A − 4A =⇒ 5A = A − 4I =⇒ A = .
5 2 −1
 
2 1 1
Example 2.44 By Cayley-Hamilton theorem find the inverse of the matrix A = 0 1 0.
1 1 2

Solution: The characteristic equation of A is det(A − λI3 ) = 0, i.e.,

2−λ 1 1
0 1−λ 0 =0 =⇒ λ3 − 5λ2 + 7λ − 3 = 0.
1 1 2−λ

By Cayley-Hamilton theorem ,

A3 − 5A2 + 7A − 3I = O
1 2
=⇒ 3I = A3 − 5A2 + 7A =⇒ 3A−1 = A2 − 5A + 7I =⇒ A−1 = (A − 5A + 7I).
3
       
5 4 4 12 1 1 0 0 2 −1 −1
Now, A2 − 5A + 7I = 0 1 0 − 5 0
0 + 7 0 1 0 =  0
1 3 0  and so
4 4 5 21 1 0 0 1 −1 −1 2
 
2 −1 −1
1
A−1 =  0 3 0 .
3
−1 −1 2
 
1 0 0
Example 2.45 By Cayley-Hamilton theorem find A100 where A = 1 0 1.
0 1 0

Solution: The characteristic equation of A is det(A − λI3 ) = 0, i.e.,

1−λ 0 0
1 −λ 1 =0 =⇒ λ3 − λ2 − λ + 1 = 0.
0 1 −λ

By Cayley-Hamilton theorem ,

A3 − A2 − A + I = O =⇒ A3 = A2 + A − I.
Md Firoz Ali 20

and so

A4 = A3 + A2 − A = (A2 + A − I) + A2 − A = 2A2 − I
A6 = A2 (2A2 − I) = 2A4 − A2 = 2(2A2 − I) − A2 = 3A2 − 2I
A8 = A2 (3A2 − 2I) = 3A4 − 2A2 = 3(2A2 − I) − 2A2 = 4A2 − 3I
A10 = A2 (4A2 − 3I) = 4A4 − 3A2 = 4(2A2 − I) − 3A2 = 5A2 − 4I.

In general, A2n = nA2 − (n − 1)I where n is a positive integer. Thus

     
1 0 0 1 0 0 1 0 0
A100 = 50A2 − 49I = 50 1 1 0 − 49 0 1 0 = 50 1 0 .
1 0 1 0 0 1 50 0 1

2.9 Diagonalization of a matrix

Definition 2.3 An n × n matrix A is said to be similar to an n × n matrix B if there exists a
non-singular n × n matrix P such that B = P −1 AP .

Since |P −1 AP − λI| = |P −1 AP − P −1 (λI)P | = |P −1 (A − λI)P | = |P −1 ||A − λI||P | = |A − λI|,

two similar matrices have same eigen values.

Definition 2.4 An n × n matrix A is said to be diagonalizable if A is similar to a diagonal matrix.

 
λ1 0 ··· 0
 0 λ2 · · · 0
If A is similar to a diagonal matrix D = diag (λ1 , λ2 , . . . , λn ) = 
· · ·
 then
· · · · · · · · ·
0 0 · · · λn
λ1 , λ2 , . . . , λn are the eigen values of A.

Theorem 2.13 Let A be an n × n matrix over a field F .

(i) A is said to be diagonalizable iff A has n linearly independent eigen vectors.

(ii) A is said to be diagonalizable iff all the eigen values of A are regular.
(iii) A is said to be diagonalizable if all the eigen values of A are distinct and belong to the field
F.
 
1 0
Example 2.46 Show that the matrix A = is not diagonalizable.
3 1

Solution: The eigen values of A are 1, 1 and the eigen vectors corresponding to the eigen value
0
λ = 1 are c where c is a non-zero real number (fill the details). Thus the eigen value 1 of A is
1
not regular (A.M6= G.M.). Therefore, A is not diagonalizable.
 
1 −3 3
Example 2.47 Diagonalize the matrix A = 3 −5 3.
6 −6 4

Solution: The eigen   of A are −2, −2, 4. The eigen vectors corresponding to the eigen
 values
1 0
value λ = −2 are c 1 + d 1 where c, d are real numbers with (c, d) 6= (0, 0). The eigen vectors
0 1
 
1
corresponding to the eigen value λ = 4 are k 1 where k is a non-zero real number. (fill the
2
details). Thus both the eigen values are regular and so A is diagonalizable.
Md Firoz Ali 21

     
1 0 1
Three linearly independent eigen vectors of A are X1 = 1 , X2 = 1 , X3 = 1. Consider
0 1 2 
  1 0 1
the matrix P whose columns are X1 , X2 , X3 respectively, i.e., P = X1 X2 X3 = 1 1 1.
0 1 2
Then P is non-singular. Thus,
  
    1 0 1 −2 0 0
AP = AX1 AX2 AX3 = −2X1 −2X2 4X3 = 1 1 1  0 −2 0 = P D,
0 1 2 0 0 4

where D = diag (−2, −2, 4) is a diagonal matrix. Thus P −1 AP = D.

 
1 1 −2
Example 2.48 Find a matrix P such that P −1 AP is a diagonal matrix, where A = −1 2 1 .
0 1 −1

Solution: The eigen values of A are   −1. The

1, 2,  eigen vectors corresponding to the eigen
3 1 1
values 1, 2, −1 are respectively c1 2 , c2 3 , c3 0 where c1 , c2 , c3 are arbitrary non-zero real
1 1 1
 and so A isdiagonalizable.
numbers (fill the details). Thus all the eigen values are regular  
3 1 1
Three linearly independent eigen vectors of A are X1 = 2 , X2 = 3 , X3 = 0. Consider
1 1 1 
  3 1 1
the matrix P whose columns are X1 , X2 , X3 respectively, i.e., P = X1 X2 X3 = 2 3 0.
1 1 1
Then P is non-singular. Thus,
  
    3 1 1 1 0 0
AP = AX1 AX2 AX3 = X1 2X2 −X3 = 2 3 0 0 2 0  = P D,
1 1 1 0 0 −1

where D = diag (1, 2, −1) is a diagonal matrix. Thus P −1 AP = D.