LINEAR ALGEBRA II

Chapter one
Characteristic Equation

by shimelis Ayele 1
1.1. Eigen values and Eigen
vectors

by shimelis Ayele 2
 Cont…
1 2 1 5 1
we have 𝐴𝑣 = = =5 = 𝜆𝑣 so 5 is an eigenvalue
4 3 2 10 2
1
of 𝐴 and is an eigenvector of 𝐴 corresponding to the eigenvalue
2
5
Definition1.1.2: Let 𝐴 be an 𝑛 × 𝑛 matrix. The set of all eigenvalues
of 𝐴 is called the spectrum of 𝐴.
Theorem 1.1.1: Let 𝐴 be an 𝑛 × 𝑛 matrix. A Number 𝜆 is an Eigen
value of 𝐴 if and only if 𝑑𝑒𝑡 𝐴 − 𝜆𝐼 = 0, where 𝐼 denotes the 𝑛 × 𝑛
identity matrix.

by shimelis Ayele 3
 Properties of Eigen Values
i. The sum of the eigen values of a matrix is the sum of the elements of
the principal diagonal.
ii. The product of the eigen values of a matrix A is equal to its
determinant.
iii. If 𝜆 is an eigen value of a matrix A, then 1Τ𝜆 is the eigen value of A-
1.

iv. If 𝜆 is an eigen value of an orthogonal matrix, then 1Τ𝜆 is also its

eigen value.
v. The eigen values of a triangular matrix are precisely the entries of
diagonal entries.

by shimelis Ayele 4
 Cont…

vi. If 𝜆1 , 𝜆2 ,…, 𝜆𝑛 are the eigenvalues of A, then

a. 𝑘𝜆1 , 𝑘𝜆2 ,…, 𝑘𝜆𝑛 are the eigenvalues of the matrix kA, where k is a
non – zero scalar.
1 1 1
b. , … are the eigenvalues of the inverse matrix 𝐴−1 .
𝜆1 𝜆2 𝜆𝑛

𝑝 𝑝 𝑝
c. 𝜆1 , 𝜆2 …𝜆𝑛 are the eigenvalues of 𝐴𝑃 , where 𝑝 is any positive
integer.

by shimelis Ayele 5
 1.2 Characteristic polynomial

• Definition 1.2.1:Polynomial of degree 𝑛 in 𝑥 is an expression of the form

𝑎𝑛 𝑥 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 + ⋯ + 𝑎1 𝑥 + 𝑎0 , where 𝑎0 , 𝑎1, …𝑎𝑛 ∈ 𝐹 𝑎𝑛𝑑 𝑎𝑛 ≠ 0
, 𝑛 𝑖𝑠 𝑛𝑜𝑛 𝑛𝑒𝑔𝑎𝑡𝑖𝑣𝑒 𝑖𝑛𝑡𝑒𝑔𝑒𝑟.

• We can write 𝑝 𝑥 = 𝑎𝑛 𝑥 𝑛 + 𝑎𝑛−1 𝑥 𝑛−1 + ⋯ + 𝑎1 𝑥 + 𝑎0 , if we replace x

every where by a given number 𝜆 and we obtain 𝑝 𝜆 = 𝑎𝑛 𝜆𝑛 +
𝑎𝑛−1 𝜆𝑛−1 + ⋯ + 𝑎1 𝜆 + 𝑎0

• Definition 1.2.2: Let 𝑝 𝑥 be a polynomial in x. A number 𝜆 is called a

root of 𝑝 𝑥 if 𝑝 𝜆 = 0.

by shimelis Ayele 6
 Cont.…

• Definition 1.2.3: Let 𝐴 = 𝑎𝑖𝑗 be an 𝑛 × 𝑛 matrix. Then the

polynomial

𝑎11 − 𝜆 𝑎12 ⋯ 𝑎1𝑛

𝑎21 𝑎22 − 𝜆 ⋯ 𝑎21
𝑑𝑒𝑡 𝐴 − 𝜆𝐼 = 𝑑𝑒𝑡 is called the
⋮ ⋮ ⋱ ⋮
𝑎𝑛1 𝑎𝑛2 … 𝑎𝑛𝑛 − 𝜆
characteristic polynomial of 𝐴.
• Denoted as 𝐶𝐴 𝑥 , the leading coefficient is −1 𝑛 𝑎𝑛𝑑 the constant
term is 𝑑𝑒𝑡𝐴.

by shimelis Ayele 7
 Cont…
• The equation 𝑑𝑒𝑡 𝐴 − 𝜆𝐼 = 0 or equivalently 𝑑𝑒𝑡 𝜆𝐼 − 𝐴 = 0 is
called characteristic equation of 𝐴.

• Definition 1.2.4: If an eigenvalue 𝜆 occur k times as a root of the

characteristic polynomial 𝐶𝐴 𝑥 , then k is called the multiplicity of the
eigenvalue .
3 −1 −1
• Example 1: Let 𝐴 = −12 0 5 then,
4 −2 −1

by shimelis Ayele 8
 Cont.
i. Find characteristic polynomial and characteristic equation of A.
ii. Eigenvalues and eigenvectors of A.
iii. Eigen values of 𝐴−1
iv. Eigen values of 𝐴10
v. Eigen values of 10A.
Solution:
Characteristic polynomial is
3 −1 −1 1 0 0
𝑑𝑒𝑡 𝐴 − 𝜆𝐼 = −12 0 5 −𝜆 0 1 0
4 −2 −1 0 0 1

by shimelis Ayele 9
 Cont.
3 − 𝜆 −1 −1
= −12 −𝜆 5
4 −2 −1 − 𝜆
= 3 − 𝜆 𝜆2 + 𝜆 + 10 + 1 12𝜆 + 12 − 20 − 1 24 + 4𝜆
characteristic polynomial is −𝜆3 + 2𝜆2 + 𝜆 − 2 and characteristic
equation of A is −𝜆3 + 2𝜆2 + 𝜆 − 2 = 0
ii. To find Eigenvalues of A we have to find root of
−𝜆3 + 2𝜆2 + 𝜆 − 2 = 0
⟹− 𝜆+1 𝜆−1 𝜆−2 =0
𝜆= -1 or 𝜆= 1 or 𝜆=2
by shimelis Ayele 10
 Cont.
So the eigenvalues are -1,1 and 2 .
To find eigenvector corresponding to eigenvalues
write 𝐴𝑣 = 𝜆𝑣
⟹ 𝐴 − 𝜆𝐼 𝑣 = 0 that is

3 − 𝜆 −1 −1 𝑥1 0
⟹ −12 −𝜆 5 𝑥2 = 0
4 −2 −1 − 𝜆 𝑥3 0
Then we can solve for 𝑥1 , 𝑥2 and 𝑥3 by appropriate method.
Let 𝜆= -1 ,then we have

by shimelis Ayele 11
 Cont.

4 −1 −1 𝑥1 0
−12 1 5 𝑥2 = 0
4 −2 0 𝑥3 0
Solve by using Gaussian elimination .
Since the constant matrix is zero matrix, reduce the coefficient matrix to row
echelon form
𝑅2 +3𝑅1
4 −1 −1 𝑅3 −𝑅1
4 −1 −1 𝑅3 −1ൗ2 𝑅2 4 −1 −1
−12 1 5 0 −2 2 0 −2 2 .
4 −2 0 0 −1 1 0 0 0

by shimelis Ayele 12
 Cont.
4 −1 −1 𝑥1 0
0 −2 2 𝑥2 = 0
0 0 0 𝑥3 0
Thus the linear system becomes
4𝑥1 − 𝑥2 − 𝑥3 = 0
−2𝑥2 + 2𝑥3 = 0
1
⇒ 𝑥2 = 𝑥3 , 𝑥1 = 𝑥3
2
 Let 𝑥3 = 𝑡 𝑡ℎ𝑒𝑛 𝑥2 = 𝑡 𝑎𝑛𝑑 𝑥1 = 1Τ2 𝑡 where t is arbitrary

by shimelis Ayele 13
 Cont.
𝑥1 1Τ
2 𝑡 1Τ
2
Thus v = 𝑥2 𝑡 =𝑡 1
𝑥3 𝑡 1
1Τ
2
So 1 is eigenvector corresponding to eigenvalue 𝜆 = -1
1
3
Similarly, we obtain −1 as an eigenvector corresponding to
7
1
eigenvalue 𝜆=1and −1 as an eigenvector corresponding to eigenvalue 2.
2
by shimelis Ayele 14
 Cont.
1
iii. Eigen values of 𝐴 . 1, -1 and are eigenvalues of 𝐴−1 by the
−1
2
property of eigenvalue.
iv. Eigen values of 𝐴10 are -1 , 1 and 210
v. Eigen values of 10A are -10,10 and 20
0 0 −2
Example 2: Let 𝐴 = 1 2 1
1 0 3
Solution:
• The characteristic equation of matrix A is 𝜆3 −5𝜆2 + 8𝜆 − 4 = 0, or,
in factored form 𝜆 − 1 𝜆 − 2 2 = 0
by shimelis Ayele 15
 Cont.
• Thus the eigenvalues of A are 𝜆 = 1and 𝜆2,3 = 2, so there
are two eigenvalues of A.
𝑥1
v = 𝑥2 ,is an eigenvector of A corresponding to 𝜆 if and
𝑥3
only v is a nontrivial solution of 𝐴 − 𝜆𝐼 𝑣 = 0

that is, of the form

−𝜆 0 −2 𝑥1 0
1 2−𝜆 1 𝑥2 = 0
1 0 3 − 𝜆 𝑥3 0
by shimelis Ayele 16
 cont.
−1 0 −2 𝑥1 0
• If 𝜆 = 1, then 1 1 1 𝑥2 = 0
1 0 2 𝑥3 0
Solving this system using Gaussian elimination yields 𝑥1 =
− 2𝑠, 𝑥2 = 𝑠 , 𝑥3 = 𝑠 (verify)
Thus the eigenvectors corresponding to 𝜆 = 1are the nonzero vectors
of the form
−2𝑠 −2 −2
𝑠 = 𝑠 1 so that 1 is the Eigen vector corresponding to
𝑠 1 1
eigenvalue 1
by shimelis Ayele 17
 Cont.
−2 0 −2 𝑥1 0
• If 𝜆 = 2, then it becomes 1 0 1 𝑥2 = 0
1 0 1 𝑥3 0
• Solving this system using Gaussian elimination yields
𝑥1 = −𝑠, 𝑥2 = 𝑡 , 𝑥3 = 𝑠(verify)
• Thus, the eigenvectors of A corresponding to 𝜆 = 2 are the nonzero
vectors of the form
Thus, the eigenvectors of A corresponding to 𝜆 = 2are the nonzero
vectors of the form
𝑥1 −𝑠 −1 0
𝑥2 = 𝑡 = 𝑠 0 + 𝑡 1
𝑥3 𝑠 1 0
by shimelis Ayele 18
 Cont…
−1 0
Since 0 and 1 are linearly independent, these vectors form a basis for
1 0
the Eigen space corresponding to 𝜆 = 2.
So the Eigen vectors of A corresponding Eigen values 1and 2 are
−2 −1 0
1 , 0 and 1
1 1 0
Exercise 1:
1. Find eigenvalues and eigenvectors of 1 1 2 3 4
0 1 1 2 1
A= 0 0 2 1 1
0 0 0 3 1
by shimelis Ayele
0 0 0 0 4 19
 Cont.
3 −2 2
2. Let A= 4 −4 6 ,then find
2 −3 5
a. Find characteristic polynomial and characteristic equation of A.
b. Eigenvalues and eigenvectors of A.
c. Eigen values of 𝐴−1
d. Eigen values of 𝐴11
30 −20 20
e. Eigen values of 40 −40 60 .
20 −30 50
by shimelis Ayele 20
 1.3.Similarity of matrix
Definition1.3.1: If A and B are n xn matrices, then A is similar to B if
there is an invertible matrix P such that P -1AP = B, or, equivalently
A = PBP-1
Changing A into B is called a similarity transformation.
Theorem 1.3.1: If nxn matrices A and B are similar, then they have
the same characteristic polynomial and hence the same eigenvalues
(with the same multiplicities).
Proof:
Suppose A and B are similar matrices of order n.

by shimelis Ayele 21
 Cont.
The characteristic polynomial of B is given by
det B − λI
= det P−1AP − λI
= det p−1 𝐴 − λI p
= det(p−1)det A − λI det(p )
= det(p−1 p) det A − λI
= det A − λI
Definition 1.3.2: The set of distinct eigenvalues, denoted by σ(A) , is
called the spectrum of A.
Definition 1.3.3:For square matrices A , the number ρ(A) =
max 𝜆 , λ∈ σ(A)is called the spectral radius of A.
by shimelis Ayele 22
 1.4. Diagonalization
• Definition1.4.1: A square matrix A is called diagonalizable if there
exist an invertible matrix P such that 𝑃−1 𝐴𝑃 is a diagonal matrix; the
matrix P is said to diagonalize A.
THEOREM 1.4.1 : If 𝑃1 , 𝑃2 , …, 𝑃𝑘 are Eigen vectors of A
corresponding to distinct eigenvalues
,λ1 , λ2 , …,λk then {P1 , P2 , …, Pk } is a linearly independent set.
• Proof
 Let , 𝑃1 , 𝑃2 , …, 𝑃𝑘 be eigenvectors of A corresponding to distinct
eigenvalues 𝜆1 , 𝜆2 , …,𝜆𝑘 .

by shimelis Ayele 23
 Cont.

• We shall assume that, 𝑃1 , 𝑃2, …, 𝑃𝑘 are linearly dependent and

obtain a contradiction. We can then conclude that , P1 , P2 , …, Pk
are linearly independent.
• Since an eigenvector is nonzero by definition, p1 is linearly
independent. Let r be the largest integer such that P1 , P2, …, Pr
is linearly independent. Since we are assuming that P1 , P2, …,
Pk is linearly dependent, r satisfies 1 ≤ r ≤ k .
More over, by definition of r, P1 , P2, …, Pr+1 is linearly
dependent.
by shimelis Ayele 24
 Cont.

Thus there are scalars c1 , c2 , …, cr+1 , not all zero, such that
P1 c1 + P2 c2 + …, Pr+1 cr+1 = 0 (4)
Multiplying both sides of 4 by A and using
AP1 =λ1 p1 , AP2 = λ2 P2 … , APn = λr+1 Pr+1
we obtain c1 λ1 p1 + c2 λ2 p2 + … + cr+1 λr+1 pr+1 = 0 … (5)
Multiplying both sides of 4 by λr+1 and subtracting the resulting
equation from 5 yields
c1 (λ1 −λr+1 )p1 + c2 (λ2 −λr+1 )p2 + … + cr (λr − λr+1 )pr = 0

by shimelis Ayele 25
 Cont.
• Since P1 , P2 , …, Pr is a linearly independent set, this equation
implies that c1 (λ1 −λr+1 ) + c2 (λ2 −λr+1 ) + … + cr (λr − λr+1 ) =
0and since , λ1 , λ2 , …, λr are distinct by hypothesis, it follows that
• c1 = c2 =…= cr = 0 (6)
Substituting these values in 4 yields
• Pr+1 cn+1 = 0 (7)
Since the eigenvector Pr+1 is nonzero, it follows that cr+1 = 0
Equations 6 and 7 contradict the fact that c1 , c2 , …, cr+1 are not all
zero; this completes the proof

by shimelis Ayele 26
 Cont.
Theorem 1.4.2: If A is an n × n matrix, then the following are
equivalent.
a. A is diagonalizable.
b. A has n linearly independent eigenvectors.
• Proof
𝑎 ⇒ 𝑏 Since A is assumed diagonalizable, there is an invertible matrix

𝑝11 𝑝12 ⋯ 𝑝1𝑛

𝑝21 𝑝22 … 𝑝2𝑛
⋮ ⋮ … ⋮
𝑝𝑛1 𝑝𝑛2 … 𝑝𝑛𝑛
such that 𝑃−1 𝐴𝑃 is diagonal, say 𝑃−1 𝐴𝑃 =𝐷 ,
by shimelis Ayele 27
 Cont.

𝜆1 0 ⋯ 0
0 𝜆2 … 0
where D =
⋮ ⋮ … ⋮
0 0 … 𝜆𝑛
It follows from the formula 𝑃−1 𝐴𝑃 = 𝐷 that 𝐴𝑃 = 𝑃𝐷 ; that is ,
𝑝11 𝑝12 ⋯ 𝑝1𝑛 𝜆1 0 ⋯ 0
𝑝21 𝑝22 … 𝑝2𝑛 0 𝜆2 … 0
AP = …
⋮ ⋮ ⋮ ⋮ ⋮ … ⋮
𝑝𝑛1 𝑝𝑛2 … 𝑝𝑛𝑛 0 0 … 𝜆𝑛

by shimelis Ayele 28
 Cont.

𝜆1 𝑝11 𝜆2 𝑝12 ⋯ 𝜆𝑛 𝑝1𝑛

𝜆1 𝑝21 𝜆2 𝑝22 … 𝜆𝑛 𝑝2𝑛
= … 1
⋮ ⋮ ⋮
𝜆1 𝑝𝑛1 𝜆2 𝑝𝑛2 … 𝜆𝑛 𝑝𝑛𝑛
If we denote the column vectors of P, 𝑏𝑦 𝑃1 ,𝑃2 , …, 𝑃𝑛 , then from 1,
the successive columns of PD are 𝜆1 𝑃1 , 𝜆2 𝑃2 , …, 𝜆𝑛 𝑃𝑛 .
The successive columns of AP are 𝐴𝑃1 , 𝐴𝑃2 , …, 𝐴𝑃𝑛 .
Thus we must have
𝐴𝑃1 = 𝜆1 𝑝1 , 𝐴𝑃2 = 𝜆2 𝑃2 … , 𝐴𝑃𝑛 = 𝜆𝑛 𝑃𝑛 2
by shimelis Ayele 29
 Cont.…
• Since P is invertible, its column vectors are all nonzero; thus, it
follows from 2 that 𝜆1 ,𝜆2 , …,𝜆𝑛 are eigenvalues of A, and 𝑃1 , 𝑃2 ,
…, 𝑃𝑛 are corresponding eigenvectors.

Since P is invertible, 𝑃1 , 𝑃2 , …,𝑃𝑛 are linearly independent.

 Thus A has n linearly independent eigenvectors.

𝑏 ⇒ 𝑎 Assume that A has n linearly independent eigenvectors 𝑃1 , 𝑃2 ,

…, 𝑃𝑛 with corresponding eigenvalues , 𝜆1 , 𝜆2 , …,𝜆𝑛

by shimelis Ayele 30
 Cont.…

𝑝11 𝑝12 ⋯ 𝑝1𝑛

𝑝21 𝑝22 … 𝑝2𝑛
• Let P = … be the matrix whose column vectors
⋮ ⋮ ⋮
𝑝𝑛1 𝑝𝑛2 … 𝑝𝑛𝑛
are
𝑃1 , 𝑃2 , …, 𝑃𝑛
• The column vectors of the product AP are 𝐴𝑃1 , 𝐴𝑃2 , …, 𝐴𝑃𝑛

But , 𝐴𝑃1 = 𝜆1 𝑝1 , 𝐴𝑃2 = 𝜆2 𝑃2 … , 𝐴𝑃𝑛 = 𝜆𝑛 𝑃𝑛 Why?

by shimelis Ayele 31
 Cont.

𝜆1 𝑝11 𝜆2 𝑝12 ⋯ 𝜆𝑛 𝑝1𝑛

𝜆1 𝑝21 𝜆2 𝑝22 … 𝜆2 𝑝2𝑛
AP = …
⋮ ⋮ ⋮
𝜆1 𝑝𝑛1 𝜆2 𝑝𝑛2 … 𝜆2 𝑝𝑛𝑛

𝑝11 𝑝12 ⋯ 𝑝1𝑛 𝜆1 0 ⋯ 0

𝑝21 𝑝22 … 𝑝2𝑛 0 𝜆2 … 0
= … … ⋮ = 𝑃𝐷 3
⋮ ⋮ ⋮ ⋮ ⋮
𝑝𝑛1 𝑝𝑛2 … 𝑝𝑛𝑛 0 0 … 𝜆𝑛

by shimelis Ayele 32
 Cont.
• Where D is the diagonal matrix having the eigenvalues, 𝜆1 , 𝜆2 , …,𝜆𝑛
on the main diagonal.
• Since the column vectors of P are linearly independent, P is invertible.
Thus 3 can be rewritten as 𝑃−1 𝐴𝑃 = 𝐷 ; that is, A is diagonalizable
• THEOREM 1.4.3: If an n × n matrix A has n distinct eigenvalues,λ1 ,
λ2 , …,λn then A is diagonalizable
• Proof: If P1 , P2 , …, Pn are eigenvectors corresponding to the distinct
eigenvalues ,λ1 , λ2 , …,λn
then by Theorem , P1 , P2 , …, Pn are linearly independent.
• Thus A is diagonalizable by Theorem .

by shimelis Ayele 33
Procedure to diagonalize a matrix
 Find the linearly independent eigenvectors
 Construct P from eigenvectors
 Construct 𝐷 = 𝑃−1 𝐴𝑃 from eigenvalues
• Example 1: Finding a Matrix P that diagonalize a matrix A

0 0 −2
A= 1 2 1
1 0 3
• Solution
2
• The characteristic equation of A is 𝜆 − 1 𝜆 − 2 = 0 and we found
the following bases for the Eigen spaces:

by shimelis Ayele 34
 Cont.

−1 0 −2
• 𝜆 = 2: 𝑝1 = 0 , 𝑝2 = 1 and 𝜆 = 1: 𝑝3 = 1
1 0 1

• There are three basis vectors in total, so the matrix A is diagonalizable

and

−1 0 −2
P= 0 1 1 diagonalize A.
1 0 1

by shimelis Ayele 35
 Cont.

Verify that 𝑃−1 𝐴𝑃 =

1 0 2 0 0 −2 1 0 2 2 0 0
1 1 1 1 2 1 1 1 1 = 0 2 1
−1 0 −1 1 0 3 −1 0 −1 0 0 1

• EXAMPLE 2:
1 0 0
• Find a matrix P that diagonalizable 𝐴 = 1 2 0
−3 5 2
by shimelis Ayele 36
 Cont.
Solution
• The characteristic polynomial of A is

1−𝜆 0 0
 det 𝐴 − 𝜆𝐼 = 1 2−𝜆 0
−3 5 2−𝜆

= 𝜆−1 𝜆−2 2
Thus the eigenvalues of A are 𝜆1 = 1 and 𝜆2,3 = 2 .

by shimelis Ayele 37
 Cont.
• Verify that eigenvalues are corresponding to
−1 0
𝜆1 = 1 𝑖𝑠 𝑝1 = 1 and corresponding to 𝜆2,3 = 2 𝑖𝑠 𝑝2 = 0
−8 1
• Since A is a 3 × 3 matrix and there are only two basis vectors in total,
A is not diagonalizable by theorem.

by shimelis Ayele 38
 1.5.Cayley-Hamilton Theorem
Theorem: If CA x is characteristic polynomial of A, then CA x = 0

a11 a12 … a1n

a21 a22 … a2n
• Let A = then characteristic polynomial of A
⋮ ⋮ ⋱ ⋮
an1 an2 … ann

a11 − λ a12 … a1n

a21 a22 − λ … a2n
CA x = det A − λI = det
⋮ ⋮ ⋱ ⋮
an1 an2 … ann − λ

by shimelis Ayele 39
 Cont.
The characteristic equation is given by
CA x = xn + an−1 x n−1 + an−2 xn−2 + ⋯ + a1 x1 + a0 = 0
• CA x = An + an−1 An−1 + an−2 An−2 + ⋯ + a1 A1 + Ia0 = 0
• Theorem: Let A be a non singular n × n matrix, and let its
characteristic polynomial be CA x = An + an−1 An−1 + an−2 An−2 +
1
⋯ + a1 A + Ia0 = 0,then A = − (An−1 + an−1 An−2 + ⋯ + Ia1 )
1 −1
a0
• Proof :
• By Cayley-Hamilton theorem An + an−1 An−1 + an−2 An−2 + ⋯ +
a1 A1 + Ia0 = 0
by shimelis Ayele 40
 Cont.
• a0 = −1 n detA ≠ 0, Since A is non singular .
• So the above equation can be written as

1 n
I = − (A + an−1 An−1 + an−2 An−2 + ⋯ + a1 A1
a0
1 n−1
= − (A + an−1 An−2 + ⋯ + Ia1 )A
a0

−1 1
A =− (An−1 + an−1 An−2 + ⋯ + Ia1 )
a0

by shimelis Ayele 41
 Cont.…

2 −1 1
Example 1:𝐿𝑒𝑡 A = −1 2 −1
1 −1 2
Verify Cayley -Hamilton theorem and compute A−1
• Solution
• The characteristic equation of A is 𝐴 − 𝜆𝐼 = 0
2 − λ −1 1
• −1 2 − λ −1 = 0
1 −1 2−λ

⇒ −(λ3 − 6λ2 + 9λ − 4) = 0
by shimelis Ayele 42
 Cont.
To verify Cayley – Hamilton theorem, we have to show that
A3 − 6A2 + 9A − 4I = 0

2 −1 1 2 −1 1 6 −5 5
A2 = −1 2 −1 −1 2 −1 = −5 6 −5
1 −1 2 1 −1 2 5 −5 6

6 −5 5 2 −1 1 22 −21 21
• A3 = A2 A = −5 6 −5 −1 2 −1 = −21 22 −21
5 −5 6 1 −1 2 22 −21 22

by shimelis Ayele 43
 Cont.
• Therefore A3 − 6A2 + 9A − 4I
22 −21 21 6 −5 5 2 −1 1
= −21 22 −21 − 6 −5 6 −5 + 9 −1 2 −1 −
22 −21 22 5 −5 6 1 −1 2
1 0 0 0 0 0
4 0 1 0 = 0 0 0
0 0 1 0 0 0
• Now multiplying both side of (1 ) by A−1 yields
• A2 − 6A1 + 9I − 4A−1 = 0
• 4A−1 = (A2 − 6A1 + 9I)
by shimelis Ayele 44
 Cont.

6 −5 5 2 −1 1 1 0 0
4𝐴−1 = −5 6 −5 − 6 −1 2 −1 + 9 0 1 0
5 −5 6 1 −1 2 0 0 1

3 1 −1
= 1 3 1
−1 1 3

1 3 1 −1
⟹ 𝐴−1 = 1 3 1
4
−1 1 3
by shimelis Ayele 45
 Cont.
1 2 −1
• Example 2: given A = 0 1 −1 find AdjA by using Cayley-
3 −1 1
Hamilton theorem.
• Solution
• The characteristic equation of a matrix A is
1−𝜆 2 −1
• 𝐴 − 𝜆𝐼 = 0 i.e., 0 1−𝜆 −1 = 0
3 −1 1−𝜆
• ⟹ 𝜆3 − 3𝜆2 + 5𝜆 + 3 = 0
• by Cayley-Hamilton theorem 𝐴3 − 3𝐴2 + 5𝐴 + 3𝐼 = 0
• Multiply both side by 𝐴−1 ,we get 𝐴2 − 3𝐴 + 5𝐼 + 3𝐴−1 = 0

by shimelis Ayele 46
 Cont.

−1 1
•𝐴 = − 𝐴2 − 3𝐴 + 5𝐼
3
−2 5 −4 3 6 −3 5 0 0
1
•=− −3 2 −2 − 0 3 −3 + 0 5 0
3
6 4 −1 9 −3 3 0 0 5
1 1
0
3 3
−1 −4 1
•𝐴 = 1 3
−
3
7 1
1 − −
3 3

by shimelis Ayele 47
 Cont.

−1 𝑎𝑑𝑗𝐴
• We know that , 𝐴 =
𝐴
1 1
0
3 3
−1 −4 1
• Therefore 𝑎𝑑𝑗𝐴 = 𝐴 𝐴 = −3 1 3
−
3
7 1
1 − −
3 3

0 −1 −1
= 1 4 1
1 7 1
by shimelis Ayele 48
 Cont.
• Another application of Cayley-Hamilton theorem is to compute power
of square matrix 𝐴
2 5
• Example: Let A = and then find A735
1 −2
• Solution
• The characteristic polynomial of A is x 2 − 9. Eigen values are -3,3 .
• Division algorithm applied to the polynomial x 735 , x 2 − 9 will give
equation of the form x735 = (x 2 − 9)q x + a0 + a1 x , (1)
Where a0 + a1 x is remainder obtained by dividing x 735 by x 2 − 9.

by shimelis Ayele 49
 Cont.

• Note that the degree of remainder is less than the degree of the
divisor x 2 − 9.
• By Cayley-Hamilton theorem A2 − 9I = 0 .
• Inserting A for x in equation (1), we get
• A735 = a0 + a1 A (2)
• Inserting eigen values 3 and -3 for x successively in equation (1), we
get
• 3735 = a0 + 3a1

by shimelis Ayele 50
 Cont.…
735
• −3 = a0 − 3a1

• This gives a0 = 0, a1 = 3734 .

• Then A735 = a0 + a1 A
• gives
• A735 = 3735 A.

by shimelis Ayele 51
 1.6.Minimal polynomial
• Definition 1.6.1: The minimal polynomial of a matrix A, denoted by 𝑚𝐴 𝜆
is the unique monic polynomial of least degree such that 𝑚𝐴 𝜆 = 0 .
• Theorem: a scalar 𝜆 is Eigen values of a matrix A if and only if is root of
minimal polynomial.
• proof(exercise)
2 2 −5
• Example: find minimal polynomial 𝑚𝐴 𝜆 of 𝐴 = 3 7 −15
1 2 −4
• Solution
• First find the characteristic polynomial of A.

by shimelis Ayele 52
 Exercise 3

1 2 −1
• Let 𝐴 = 0 1 −1 , then
3 −1 1
1. Verify Cayley-Hamilton theorem
2. Find 𝐴−1 and AdjA by using Cayley-Hamilton theorem.

by shimelis Ayele 53
 Cont.
p λ = −λ3 + 5λ2 − 7λ + 3 = − λ − 1 2 λ − 3 .
The minimal polynomial must divid𝑒
characteristic polynomial.
• Thus minimal polynomial is exactly one of the following
p λ = λ − 3 λ − 1 2or g λ = λ − 3 λ − 1
1 2 −5 −1 2 −5 0 0 0
g A = A − I A − 3I = 3 6 −15 3 4 −15 = 0 0 0
1 2 −5 1 2 −7 0 0 0

by shimelis Ayele 54
 Cont.
Thus 𝑔 𝑡 = (𝜆 − 1)(𝜆 − 3) = 𝜆2 − 4𝜆 + 3 is a minimal polynomial
of A
Exercise 3 : Find minimal polynomial m (λ ) of
3 −2 2
i. 𝐴 = 4 −4 6
2 −3 5

1 0 0 0
i. 𝐵= 0 1 1 0
0 0 1 0
0 0 0 1
by shimelis Ayele 55