13

EIGEN VALUES AND EIGEN

VECTORS OF A MATRIX

13.1 Let A = [aij] be a square matrix of dimension n × n. The equation

a11 − λ a12 L a1n
b
det A − λI =g a12
M
a 22
M
− λ L
M
a 2n
M
= 0, (1)
an1 an 2 L ann − λ

where λ is a parameter called the characteristic equation of A. It is of degree n. The roots of this
characteristic equation are called the characteristic roots or the Eigen values of the matrix A. A
square matrix of order n has always n eigen values.
When the determinant in (1) is expanded it becomes a polynomial of degree n, which is called
characteristic polynomial.
From (1) we get

b g b g
λn − a11 + a22 + ... + ann λn −1 + ... + −1
n
det A = 0
∴ λ 1 + λ 2 + ... + λ n = a11 + a22 + ... + ann = Trace A
λ1 λ 2 ... λ n = det A = A .
Consider the homogeneous equations
a11 x1 + a12 x2 + L + a1n xn = λx1 U|
a21 x1 + a22 x2 + L + a2 n xn = λx2
M
V| (2)
an1 x1 + an 2 x2 + L + ann xn = λxn W
where λ is an undetermined parameter. The n values of λ for which non-zero roots of the
homogeneous equations (2) exist, are called the Eigen values or Characteristic values of the parameter
λ . The non-zero column vector x satisfying (A— λI ) X = 0, is called Eigen vector of A. Corresponding
to each of the eigen value there is an eigen vector of A.
Consider the projection transformation in two dimensional space OX1X2 defined by the matrix

282
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 283

LM OP
A= 1
N Q
0
0 0 .
Here the eigen vectors are:
(i) The non-zero vectors directed along the X1-axis with eigen value λ 1 = 1 and
(ii) The non-zero vectors directed along the X2-axis with eigen value λ 2 = 1 (see Fig. 1)
X2

0 X1
x
Fig. 1
We state below some important properties of eigen values and eigen vectors of a matrix:
1. For each distinct eigen value λ , there exists at least one eigen vector corresponding to λ .
2. If A is a square matrix and λ 1 , λ 2 , ..., λ n are distinct eigen values of A with associated eigen
vectors v1, v2, …, vn respectively, then {v1, v2, …, vn} is a set of linearly independent vectors.
3. If B is a non-singular matrix then A and B–1 AB have same eigen values.
4. The eigen values of a Hermitian matrix are real.
5. The eigen values and eigen vectors of a real symmetric matrix are real.
6. The number of linearly independent eigen vectors corresponding to one and the same root
of the characteristic equation does not exceed the multiplicity of that root.
Note: In this chapter, we write

a11 − λ a12 ... a1n

a21 a22 − λ ... a2 n
| A = λI | = ... = p0λn + p1λn −1 + ... + pn (say)
... ... ...
an1 an 2 ... ann − λ

Example 13.1 Find the characteristic values (Eigen values) of the matrix
LM 3 −1 OP
A =
N2 0 Q
284 NUMERICAL ANALYSIS

Solution The characteristic equation of A is

| A − λI | = 0

3− λ −1
i.e., = 0
x −λ
⇒ (–λ) (3 – λ) + 2 = 0
λ2 – 3 + 2 = 0
⇒ (λ – 1) (λ – 2) = 0 ...(1)
The roots of are λ = 1, λ = 2
∴ The characteristic roots of values of A are 1, 2.
Example 13.2 Find the characteristic values and the corresponding characteristic vectors of

LM3 1 OP
A =
N1 3 Q
Solution The characteristic equation of A is
|A –λI| = 0

3− λ 1
i.e., = 0 ⇒ (3 – λ)2 – 1 = 0
1 3− λ

⇒ λ2 − 6λ + 9 − 1 = 0 ⇒ λ2 − 6λ + 8 = 0
⇒ (λ – 2) (λ – 4) = 0
∴ λ = 2, λ = 4 are the characteristic values of A.

LM x OP
1
Let x =
Nx Q
2

denote the characteristic vector corresponding to the value λ = 2; then

( A − λI ) x = 0
LM3 − 2 1 OP LM x OP LM0OP
1
⇒
N1 3− 2 = Q N x Q N0Q
2

LM1 1O L x O LM0OP
1PQ MN x PQ =
1
⇒
N1 2 N0Q
⇒ x1 + x2 = 0, x1 + x2 = 0
i.e., x1 + x2 = 0
⇒ x1 = –x2

x1 x2
⇒ = = k (say)
1 −1
∴ The characteristic vector corresponding to the value λ = z is

LM k OP
X1 =
N− k Q
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 285

LM 1 OP
or X1 =
N−1Q
when λ = 4 : we have
LM3 − 4 1 OP LM x OP = LM0OP
1
(A – λI) X = 0 ⇒
N1 3− 4 Q N x Q N0Q
2

⇒ − x1 + x2 = 0, x1 − x2 = 0
⇒ x1 = x2

x1 x2
⇒ = = k ′ (say)
1 1
∴ The characteristic vector corresponding to the value

λ = 4 is X 2 =
LMk ′OP or X = LM1OP
N k ′Q N1Q
2

The characteristic roots of A are λ = 2, λ = 4 and the characteristic vectors of A are X1 =

LM 1 OP, LM1OP
N−1Q X2 =
N1Q
Note: The characteristic vectors may be normalized and expressed as

LM 1 OP LM 1 OP
X1 = MM 21 PP, X2 =M
MM 1 PPP
2

MN− 2 PQ N 2Q
Example 13.3 Obtain the eigen values and eigen vectors of the symmetric matrix
0 1 1 LM OP
A= 1 0 1MM P
N1 1 0 QP
Solution The characteristic equation
–λ 1 1
A − λI = 0 gives 1 –λ 1 = 0,
1 1 –λ

b
⇒ λ +1 g b λ − 2g = 0 ,
2

hence eigen values are λ = –1, –1, 2.

Case 1 When λ = 2
The corresponding eigen vector is given by
(A – 2I)X = 0
⇒ –2x + y + z = 0
x – 2y + z = 0
x + y – 2z = 0
solving we get x = y = z = k (say), thus
X1 = [k k k]T.
286 NUMERICAL ANALYSIS

Case 2 When λ = –1, –1

The corresponding eigen vector is given by
(A – (–1)I ) X = 0, i.e., (A + I) X = 0,
⇒ x + y + z = 0
(all three equations are equivalent).
The trial solutions are
[1 0 –1]T and [1 + k –1 –k]T
since these are orthogonal, we have
1 + k + k = 0 ⇒ k = –1/2

X2 =
LM 1 0 −1 OP T
, X3 =
LM 1 −2 1 OP .
T

hence
N 2 2Q N6 6 6Q
Example 13.4 Determine the eigen values and the corresponding eigen vectors of the following system

10x1 + 2x2 + x3 = λx1

2x1 + 10x2 + x3 = λx2

2x1 + x2 + 10x3 = λx3

Solution We have

LM 10 2 1 OP
A= 2 MM 10 1 P
N2 1 10 QP

The characteristic equation

10 − λ 1 1
A − λI = 2 10 − λ 1 =0
2 1 10 − λ

b
⇒ 10 − λ g 3
b g
− 7 10 − λ + 6 = 0

⇒ λ = 13, λ = 9, λ = 8

i.e., λ 1 = 13, λ 2 = 9, λ 3 = 8 .

Case 1 When λ 1 = 13
∴ We get
–3x1 + 2x2 + x3 = 0,
2x1 – 3x2 + x3 = 0,
2x1 + x2 + 3x3 = 0,
solving the first two equations with x3 = 1, we get
x1 = 1, x2 = 1
these values satisfies all the three equations. The first eigen vector is [1 1 1]T.
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 287

Case 2 When λ 2 = 9
⇒ (A –9I)X = 0
⇒ x1 + 2x2 + x3 = 0
2x1 + x2 + x3 = 0
2x1 + x2 + x3 = 0
with x3 = 1, the second eigen vector becomes

LM− 1 1 1OP .
T

N 33Q
Case 3 When λ 3 = 8
The corresponding eigen vector is given by
(A –8I)X = 0
⇒ 2x1 + 2x2 + x3 = 0
2x1 + 2x2 + x3 = 0
2x1 + x2 + 2x3 = 0
with x3 = 1, the third eigen vector becomes

LM− 3 1 1OP .T

N 2 Q
Example 13.5 Find the Eigen values and eigen vectors of the matrix
LM 2 1 1 OP
MM
A= 1 2 1 PP
N
1 1 2 Q
Solution The characteristic equation of the matrix is
2−λ 1 1
A − λI = 2 2−λ 1 =0
2 1 2−λ

b
⇒ λ −1 g b4 − λ g = 0
2

we get λ 1 = λ 2 = 1, λ 3 = 4

Case 1 λ 1 = 1, we get

2−1 LM OP
b A − λI g X = 0
1 1 x1
⇒ 2
2
2−1
1
1
MN
x2 = 0
2 − 1 x3 PQ
⇒ x1 + x2 + x3 = 0 U|
x1 + x2 + x3 = 0 V| (3)
x1 + x2 + x3 = 0 W
The rank of the system (3) is one, therefore two of the equations are consequences of the third. It suffices to
solve the equation
288 NUMERICAL ANALYSIS

x1 + x2 + x3 = 0.
Putting
x1 = k1, x2 = k2,
we get x3 = –[k1 + k2],
where k1 and k2 are arbitrary scalars not simultaneously zero.
In particular choosing k1 = 1, k2 = 0, and then k1 = 0, k2 = 1, we get the solution consisting of two linearly
independent eigen vectors of matrix A
X1 = [1 0 –1]T, X2 = [0 1 –1]T.
All the other eigen vectors of A that correspond to the eigen value λ 1 = 1 are linear combinations of these basis
vectors and fill the plane spannel by the vectors X1 and X2.
Case 2 Now λ 3 = 4, gives

LM−12 1 1 OP LM xx OP = 0 ,
1

PQ MN x PQ
−2 1
MN 1 1 −2
2
3

or –2x1 + x2 + x3 = 0
x1 – 2x2 + x3 = 0
x1 + x2 – 2x3 = 0
The rank of the above system is 2. The third equation of the system is a consequence of the first two equations.
∴ Solving first two equations
–2x1 + x2 + x3 = 0
x1 – 2x2 + x3 = 0

x1 x2 x
we get = = 3
3 3 3
i.e., x1 = x2 = x3 = k
where k is a constant different from zero, putting k =1, we get the simplest solution that effects the eigen vector of
A.
T
∴ λ3 = 1 1 1 .
Example 13.6 If a + b + c = 0; find the characteristic roots of the matrix

LMa c b OP
MMc b a P
c PQ
A =
Nb a

Solution The characteristic matrix of A is

LMa − λ c b OP
MM c b−λ a P
c − λ QP
A – λI =
Nb a

applying c1 → c1 + c2 + c3
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 289

LM−λ c b OP
MM−λ b−λ a P
c − λ QP
~
N− λ a

R 2 → R 2 − R1
R 3 → R 3 − R1

LM−λ c OP
b
MM0 b−λ −c a −b P
c − b − λ PQ
~
N0 a−c

∴ The characteristic Equation of A is

−λ c b
|A − λI| = 0 ⇒ 0 b−λ−c a−b =0
0 a−c c−b−λ

⇒ ( − λ ) (b − c − λ ) (c − b − λ ) − (a − b) ( a − c) = 0

⇒ λ − (b − c) − λ (b − c + λ ) − a 2 + ac + ab − bc = 0

2 2 2
⇒ λ − (b − c) − λ − a + ac + ab − bc = 0

2 2 2 2
⇒ λ λ − b − c + 2bc − a + ac + ab − bc = 0

2 2 2 2
⇒ λ λ − ( a + b + c ) + ab + bc + ca = 0 ...(1)

but a + b + c = 0 ⇒ ( a + b + c)2 = 0

⇒ a 2 + b 2 + c2 + 2ab + 2bc + 2ca = 0

( a 2 + b 2 + c2 )
⇒ (ab + bc + ca) = − ...(2)
2
from (1) and (2), we get

(a 2 + b2 + c2 )
λ ( λ2 − ( a 2 + b 2 + c 2 ) − = 0
2

F 3( a 2 + b 2 + c2 ) I
⇒ GH
λ λ2 −
2 JK = 0

∴ The characteristic roots of A are

3 (a 2 + b 2 + c2 )
λ = 0, λ = ±
2
290 NUMERICAL ANALYSIS

13.2 METHOD FOR THE LARGEST EIGEN VALUE (POWER METHOD)

Let A be a given matrix whose (largest) eigen value is to be determined and X0 be an arbitrary
vector. We use X0 as the initial approximation to an eigen value of the matrix A.
Suppose we “normalize” the vector X0; by requiring that one component say the last by unity.
Compute the sequence
AX0 = λ1x1
AX1 = λ2x2
AX2 = λ3x3
... ...
AXi–1 = λi xi
... ...
In this sequence; all the vectors X1, X2, ... are to be normalized in whatever manner was chosen
originally. The iterative procedure converges, and we get a relation of this form
AX = λX
where X = lim xi, λ = lim λi
If the eigen value of A is real and unrepeated the above process will converge to give the largest
eigen value of the matrix A.
Example 1 Find the largest eigen value for the matrix

LM10 4 −1 OP
MM 4 2 3P
N −1 3 1 PQ

also find the eigen vector corresponding to the largest eigen vector.
Solution

LM10 4 −1 OP
Let A = MM 4 2 3 P
N−1 3 1 PQ

LM1OP
and X0 = MM1PP (Initial approximation of the eigen vector)
N1Q
LM10 4 −1 OP LM1OP LM13OP
MM 4 2 3 P M1P = M9 P
1 PQ MN1PQ MN3 PQ
AX 0 =
N −1 3

1. 0 LM OP
MM P
= 13 0. 62930 = 13X1 where
N0.23076PQ
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 291

LM1.0 OP
X1 = MM0.62930PP
N0.23076Q
LM10 4 OP LM
−1 1. 0 OP
MM 4 2 3 PM 0. 62930 P
1 PQ MN0. 23076PQ
AX 1 =
N −1 3

LM12.5384 OP LM1.0 OP
= MM 6. 07692 PP = 12. 5384 MM0.48466PP
N1.30796 Q N0.10429Q
1. 0 LM OP
= 12.5384 x2 where x2 = 0. 48466 MM P
N0.10429PQ
LM10 4 −1 OP LM1.0 OP
MM4 2 3 P M0.48466P
1 PQ MN0.10429 PQ
AX 2 =
N−1 3

LM11.83436OP LM1.0 OP
= MM5 . 28220 PP = 11. 834836 MM0.44634PP
N 0. 55828 Q N0.04717Q
1.0 LM OP
= 11.834836 x3 where x3 = 0. 44634
MM P
N0.04717PQ
LM10 4 −1O L1. 0
P M OP
MM 4 2 3 P M0. 44634 P
1 PQ MN0. 04717 PQ
AX 3 =
N−1 3
LM11.73821 OP LM1.0 OP
= MM5 . 03421 PP = 11. 73821 MM0.42887 PP
N 0. 386210 Q N0.032902Q
= 11.73821 x4

LM1.0 OP
where X4 =
MM0.42887 PP
N0.032902Q
LM10 4 −1OP LM1.0 OP
MM4 2 3 P M0.42887 P
1 PQ MN0. 032902 PQ
AX 4 =
N−1 3
292 NUMERICAL ANALYSIS

LM11.6826 OP LM1.0 OP
= MM4.95645 PP = 11.6826 MM0.424259PP
N0.319524Q N0.027350Q
LM1.0 OP
11. 6826 x where x = M0. 424259 P
MN0.027350PQ
= 5 5

LM10 4 −1O L1. 0

P M OP L11.66969 O
MM4 3 P M0. 424259 P = M4.93057 P
AX 5 = 2
1 PQ MN0. 027350 PQ
MM P
N−1 3 N0.300129PQ
LM1.0 OP
= 11. 66969 M0. 422510P
NM0.025718QP
LM1.0 OP
= 11. 66969 x6 where x6 = 0. 422510 MM P
N0.025718PQ
LM10 4 −1O L1. 0
P M OP
AX 6 = MM 4 2 3 PM 0. 422510 P
N −1 3 1 PQ MN0. 025718PQ

LM11.6643 OP LM1.0 OP
= MM4.92217PP = 11.6643 MM0.42198PP
N0.29325 Q N0.02514Q
LM1.0 OP
MM
= 11. 6643 x7 where x7 = 0. 42198 PP
N
0. 02514 Q
LM10 4 −1O L1. 0
P M OP
AX = M 4 2 3 PM 0. 42198 P
MN−1 3 1 PQ MN0. 02514 PQ
7

LM11.6628 OP LM1.0 OP
= M4. 91939 P = 11. 6628 M0. 42180P
MN0.291099PQ MN0.02495PQ
L1.0 O
= 11.6628 x where X = MM0. 42180PP
MN0.02495PQ
7 8

LM1.0 OP
The largest eign value is 11.6628 and corresponding eigen vector is M0. 42180P .
∴
MN0.02495PQ
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 293

Example 2 Find the largest eigen value and the corresponding eigen vector of the matrix.

LM1 6 1 OP LM1 OP
MM1 2 0 P taking x0 = 0 MM PP
3 PQ
A =
N0 0 N0 Q
LM1 6 1 OP LM1 OP LM1OP LM1OP
Solution We have AX 0 = MM1 2 0 P MM0PP = MM1PP = 1 . MM1PP
N0 0 3PQ N0Q N0Q N0Q
LM1 OP
= M1 P
= 1.X1, where X1
MN0PQ
LM1 6 1 OP LM1 OP LM7OP LM1 OP
MM1 2 0 P MM1 PP = MM3PP = 7MM0.4PP = 7X
3PQ
∴ AX 1 = 2

N0 0 N 0Q N 0 Q N 0 Q
LM1 OP
where X2 = MM0.4PP
N0 Q
LM1 6 1 OP LM1 OP LM3.4OP LM1 OP
MM1 2 0 P MM0.4PP MM PP
= 1. 4 = 3. 4 0.52MM PP
3PQ
Then AX 2 =
N0 0 N0 Q N0 Q N0 Q
LM1 OP
= 3.4X3, where X = M0.52 P
MN0 PQ
3

1O L1 O
LM1 6
MM1 2 P M P LM4.12 OP
0P M0.52 P = M2. 04 P = 4.12 X
3PQ MN0 PQ MN0 PQ
∴ AX 3 = 4

N0 0
LM1 OP
where X4 = MM0.49PP
N0 Q
LM1 6 1 OP LM1 OP LM3.94OP
Now AX 4 = MM1 2 0 PP MM0.49PP = MM1.98 PP
N0 0 3 Q N0 Q N 0 Q
1 LM OP
MM PP
= 3. 94 0.5 = 3. 94 X5
0 N Q
294 NUMERICAL ANALYSIS

LM1 OP
where X5 = MM0.5PP
N0 Q
LM1 6 OP
LM1 OP LM4OP LM1 OP
1
AX = MM1 2 PP
MM0.5PP = MM2PP = 4MM0.5PP
0
5
N0 0
N0 Q N0 Q N0 Q
3 Q
LM1 OP
= 4X , where X = M0.5P
6
MN0 PQ 6

LM1 OP
∴ The largest eigen value is 4 and the corresponding eigen vector is M0.5P .
MN0 PQ
13.3 CAYLEY-HAMILTON THEOREM
Theorem: Every square matrix A satisfies its own characteristic equation.
Proof: Let A = [aij]n×n
Then the characteristic matrix of A is A – λI and the cofactors of |A − λI| are of degree at most degree n–1.
Therefore the highest power of λ in the polynomial of Adj (A – λI) is n – 1.
We can write
n −1
Adj (A – λI) = B0λ + B1 λn − 2 + ... + Bn −1 ...(1)
where B0, B1, ..., Bn–1 are matrices of order n and whose elements are polynomials in the elements of A
∴ we have
(A − λI ) Adj (A − λI) = |A − λI| I ...(2)

or (A − λI) Adj (A − λI) = P0λn + P1λn −1 + P2 λn − 2 + ... + Pn I ; say ...(3)

(where |A − λI| = P0λn + P1λn −1 + ... + Pn )

Using (1), we can write

(A − λI) (B0λn −1 + B1λn − 2 + ... + Bn −1 ) = ( P0λn + P1λn −1 + ... + Pn ) I ...(4)

(4) is an identity in λ : therefore equating the coefficients of like powers of λ from both sides we obtain
–B0 = P0I
AB0 – B1 = P1I
AB1 – B2 = P2I ....(5)
. .
. .
. .
ABr–1 – Br = PrI
.
.
.
ABn–1 = PnI
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 295

By pre-multiplying these equations with An, An–1, ..., A, I respectively and adding we have
0 = P0An + P1An–1 + ... + PnI ...(6)
The matrix A satisfies its own characteristic equation.
Hence Proved.

Computation of the inverse of a non-singular Matrix

Cayley-Hamilton theorem can be used to compute the inverse of a non-singular matrix;
Let A be non-singular matrix of order. Then by Cayley-Hamilton theorem; we have

P0A n + P1A n −1 + P2A n − 2 + ... + Pn −1 A + PnI = 0 ...(i)

on multiplying (i) by A–1; we obtain

P0A n −1 + P1A n − 2 + P2 A n − 3 + ... + Pn −1 A + PnI = 0.A −1 ...(ii)

or P0A n −1 + P1A n − 2 + ... + Pn A −1 = 0

i.e., PnA–1 = −(P0A n −1 + P1A n − 2 + ... + Pn −1 I)

1 n −1
⇒ A–1 = − P ( P0A + P1A n − 2 + ... + Pn −1I) ...(iii)
n

Thus; inverse of A can be evaluated by putting the values of An–1; An–2, ... in (iii)
Remark : We can also apply cayley-Hamilton theorem to find A–2, A–3,...
Example 13.7 Verify cayley-Hamilton theorem for the matrix

LM 5 6 OP
N1 2 Q
Also; find the inverse of the matrix A.

LM5 6 OP
Solution Let A =
N1 2 Q
then the characteristic equation of A is
|A − λI| = 0

5− λ 6
i.e., = 0
1 2−λ

i.e., (5 − λ ) (2 − λ ) − 6 = 0

⇒ λ2 − 7 λ + 4 = 0 ...(1)
we have to show that A satisfies (1)

LM5 6 OP LM5 OP LM
6
=
31 42 OP
now A2 =
N1 2 Q N1 2Q N7 10 Q
∴ A2 – 7A + 4I
LM31 42
−7
5 OP LM 6OP LM
+4
1 0 OP
=
N7 10 1 Q N 2 Q N0 1 Q
296 NUMERICAL ANALYSIS

LM31 − 35 + 4 42 − 42 + 0
=
0 OP LM 0 OP
=
N7 − 7 + 0 10 − 14 + 4 0 Q N 0 Q
∴ A2 – 7A + 4I = 0 ...(2)
Hence, the matrix A satisfies its own characteristic equation.
Cayley-Hamilton theorem is verified
from (2), we have
A2 – 7A + 4I = 0

⇒ A −1 (A 2 − 7A + 4I) = 0

⇒ A − 7I + 4IA −1 = 0

⇒ 4 A −1 = 7 I − A = 7 LM1 0 OP − LM5 6 OP
N0 1 Q N1 2 Q
1 2 LM −6 OP
⇒ A–1 =
4 −1 N 5 Q
Example 13.8 Show that the matrix

LM1 2 0 OP
A = MM2 −1 0 PP
N0 0 −1 Q
Satisfies its own characteristic equation and find A–1.
Solution The characteristic equation of A is

1− λ 2 0
A − λI = 0 ⇒ 2 −1 − λ 0 =0
0 0 −1 − λ

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

⇒ (1 − λ ) (1 + λ ) 2 + 4(1 + λ ) = 0

⇒ − λ3 − λ2 + 5λ + 5 = 0

⇒ x 3 + λ2 − 5λ − 5 = 0 ...(1)
We have to show that A satisfies the equation (1); i.e., A3 + A2 – 5A – 5I = 0

LM1 2 0 OP LM1 2 0 OP
A . A = M2 −1 0 PP MM2 −1 0 PP
NM0
now A2 =
0 −1 Q N0 0 −1 Q
LM 5 0 0 OP
= MM 0 5 0 PP
N0 0 1 Q
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 297

LM5 0 0 OP LM1 2 0 OP
A3 = A2 . A = M0 5 0P M 2 −1 0 P
NM0 0 1 QP MN0 0 −1QP

LM 5 10 OP
0
= MM10 −5 0P
N0 0 −1 PQ

∴ A3 + A2 – 5A – 5I

LM 5 10 0 OP LM5 0 0 OP LM 1 2 0 OP LM 1 0 0OP
= M10 −5 0 + 0 PP MM 5 PP
0 −5 2 MM −1 0 −5 0 PP MM 1 0PP
MN 0 0 −1 0 Q N 0 −1 0Q N 0 −1 0 Q N 0 −1 Q
LM5 + 5 − 5 − 5 10 + 0 − 10 + 0 0+0+0+0 OP
= MM10 + 0 − 10 + 0 −5 + 5 + 5 − 5 0+0+0+0 PP
N0 + 0 + 0 + 0 0+0+0+0 −1 + 1 + 5 − 5 Q
LM0 0 0 OP
= MM0 0 0P
N0 0 0PQ

∴ A satisfies its own characteristic equation

i.e., A 3 + A 2 − 5A − 5I = 0
Consider A3 + A2 – 5A – 5I = 0

⇒ A −1 (A 3 + A 2 − 5A − 5I) = 0

⇒ A 2 + A − 5I − 5A −1 = 0
⇒ 5A–1 = 5 – A – A2

5 LM 0 0 OP LM1 2 0 OP LM5 0 0 OP
∴ 5A–1 = 0
MM 5 P
0 − 2 MM −1 0 PP − MM0 5 0 P
N0 0 5 PQ N0 0 −1 Q N0 0 1 PQ

LM1 2 0OP
= 2 MM −1 0 P
0 N 0 −1 PQ

1 LM 2 0 OP
MM P
1
⇒ A–1 = 2 −1 0
N − 5 PQ
5
0 0
298 NUMERICAL ANALYSIS

Exercise 13.1

1. Determine the eigen values and the corresponding eigen vectors for the matrices
LM2 2 OP 1LM 4 OP 1 LM 2 OP
(a)
MN 2 1 PQ (b) 3
N 2 Q (c) 3
N 4 Q
2. Find the characteristic roots and the corresponding characteristic vectors of the matrices

LM 8 −6 2 OP LM1 2 3 OP LM 3 −1 OP
3
(c) M−1 PP
MM
(a) −6 7 −4 PP (b) 0
MM −4 2 PP MN 0
2 −1
N2 −4 3 Q N0 0 7 Q −1 3 Q
3. Find the largest eigen values and the corresponding eigen vector of the matrix
LM−15 4 3 OP
A= MM 10 −12 6 P
N 20 −4 2PQ

4. Determine the largest eigen value and the corresponding eigen vector of the matrix
2LM −1 0 OP
A = −1 MM 2 −1 PP
0 N −1 2 Q
5. Find the latent roots and the characteristic vectors of the matrices

L2 OP LM 6 −2 OP LM OP
(i) M0
0 1 2 1 2 3
A = M −2 P
MM 2 0 P ii)
MN 2
3 −1
3PQ
MM
(iii) 0 2 3 P
N1 0 2PQ −1 N0 0 2PQ
6. Find the latent roots and latest vectors of the matrix

a LM h g OP
A= o MM b o PP
o N o c Q
7. Verify Cayley-Hamilton’s theorem for the matrix

LM 0 0 1OP
A=M 3 1 0 PP
MN−2 1 4 Q
8. Show that the matrix A, satisfies the matrix equation A2 − 4 A − 5I = 0

LM1 2 2 OP
where A = M2 1 2 PP
MN2 2 1 Q
EIGEN VALUES AND EIGEN VECTORS OF A MATRIX 299

9. Show that the matrix A =

LM1 2 OP
N1 Q
satisfies Cayley-Hamilton’s theorem
1

L1 OP
A = MM3
1 2
10. Verify that the matrix
MN2
1 1 PP satisifies Cayley-Hamilton theorem.
3 1 Q
Answers

1. (a) λ1 = 0, λ 2 = 3,
LM −1 2 OP and LM
T
2 1 OP T

MN 3 3 PQ MN 3 3 PQ
= 5, M
L −4 1OP and 1 1
T
T
(b) λ1 = − 2, λ 2
N3 Q
LM0.46OP
N1Q
(c) 5. 38,

LM 1 1 1OP LM−1 −1 1OP

T T
T
2. (a) λ1 = 0, λ 2 = 3, λ 3 = 15
N2 Q N 2 Q , 2 −2 1

(b) [1 –4 7]T
T T T
(c) λ1 = 1, λ 2 = 3, λ 3 = 4, 1 2 1 , 1 0 − 1 , 1 − 1 1
3. [1 –0.5 –1]T
4. 3.41[0.74 –1 0.67]T
5. (i) 1, 2, 3, [1 0 –1]T, [0 1 0]T, [1 0 1]T
(ii) 8, 2, 2, [2 –1 1]T, [1 0 –2]T, [1 2 0]T.
(iii) [1 0 0]T, [2 10]T
6. [k, 0 0]T, [k2h k2(b – a)0]T
[k3 g 0 k3(c – a)]T