Tutorial 1: Eigenvalues and Diagonalization

Course: Introduction to AI and Data Science

15.07.2025

Course Outcome: Apply advanced mathematical concepts such as matrix opera-

tions, singular values, and principal component analysis to analyze and solve engineering
problems.

Apply Level Questions (KL3)

1. Compute the eigenvalues and eigenvectors of the following matrices:
 
1 2
(a)
2 1

Solution: Step 1: Characteristic Polynomial

To find eigenvalues, solve the characteristic equation:

det(A − λI) = 0

 
1−λ 2
A − λI =
2 1−λ

det(A − λI) = (1 − λ)2 − 4 = λ2 − 2λ − 3

Step 2: Solve the characteristic equation

λ2 − 2λ − 3 = 0 ⇒ (λ − 3)(λ + 1) = 0
So, the eigenvalues are:
λ1 = 3, λ2 = −1
Step 3: Find eigenvectors
For λ1 = 3:
    
−2 2 −2 2 x
A − 3I = ⇒ =0
2 −2 2 −2 y

1
Mathematical Foundations Tutorial 1

 
1
−2x + 2y = 0 ⇒ y = x ⇒ x1 =
1
For λ2 = −1:
    
2 2 2 2 x
A+I= ⇒ =0
2 2 2 2 y
 
1
2x + 2y = 0 ⇒ y = −x ⇒ x2 =
−1
Final Answer:
   
1 1
Eigenvalues: λ = 3, −1 Eigenvectors: x1 = , x2 =
1 −1
 
2 0 0
(b) 0 3 4
0 4 9

Solution: Step 1: Characteristic Polynomial

2−λ 0 0
det(B − λI) = 0 3−λ 4
0 4 9−λ
Expand along the first row:

3−λ 4
= (2 − λ) = (2 − λ)[(3 − λ)(9 − λ) − 16]
4 9−λ

= (2 − λ)(λ2 − 12λ + 11)

Step 2: Solve the characteristic equation
√ √
12 ± 144 − 44 12 ± 100 12 ± 10
λ2 − 12λ + 11 = 0 ⇒ λ = = =
2 2 2
So the eigenvalues are:

λ1 = 2, λ2 = 11, λ3 = 1

Step 3: Find Eigenvectors

For λ1 = 2:
 
0 0 0
B − 2I = 0 1 4
0 4 7
From second row: y + 4z = 0 ⇒ y = −4z, and x is free

2
Mathematical Foundations Tutorial 1

So, eigenvector:  
1
x1 = 0
0
For λ2 = 11:
 
−9 0 0
B − 11I =  0 −8 4  ⇒ −8y + 4z = 0 ⇒ z = 2y, x=0
0 4 −2
 
0
x2 = 1

2
For λ3 = 1:
 
1 0 0
B − I = 0 2 4 ⇒ 2y + 4z = 0 ⇒ y = −2z, x=0
0 4 8
 
0
x3 = −2
1
Final Answer:
     
1 0 0
Eigenvalues: λ = 2, 11, 1 Eigenvectors: x1 = 0 ,
 x2 = 1 ,
 x3 = −2
0 2 1

2. The electrical impedance matrix in a two-node circuit is given by:

 
5 −2
Z=
−2 5

Compute its eigenvalues and comment on whether the matrix is diagonalizable.

Solution: We are given the matrix:

 
5 −2
Z=
−2 5

Step 1: Compute the eigenvalues

To find the eigenvalues, solve the characteristic equation:

det(Z − λI) = 0

 
5 − λ −2
det = (5 − λ)2 − (−2)2 = (5 − λ)2 − 4 = 0
−2 5 − λ

3
Mathematical Foundations Tutorial 1

(5 − λ)2 = 4 ⇒ 5 − λ = ±2 ⇒ λ = 3, 7

So, the eigenvalues are λ1 = 3 and λ2 = 7, both real and distinct.

Step 2: Comment on diagonalizability

The matrix Z is a real symmetric matrix with distinct eigenvalues. Hence:

• It has two linearly independent eigenvectors,

• It is diagonalizable,
• Moreover, it is orthogonally diagonalizable.

Conclusion: The matrix Z has real and distinct eigenvalues λ = 3 and λ = 7, and
is therefore diagonalizable.

3. Consider a structural matrix used in bridge design:

 
4 1
A=
0 4

Compute its algebraic and geometric multiplicity and determine if the matrix is
diagonalizable.
Step 1: Find the eigenvalues (Algebraic Multiplicity)
Compute the characteristic polynomial:
 
4−λ 1
det(A − λI) = det = (4 − λ)2
0 4−λ

Hence, the eigenvalue is:

λ = 4 with algebraic multiplicity is2

Step 2: Find the geometric multiplicity

To find the geometric multiplicity, we solve:
    
0 1 x1 0
(A − 4I)x = 0 ⇒ =
0 0 x2 0

From the first equation: x2 = 0

The second equation imposes no new condition, so x1 is free.
Thus, the general solution is:
   
x1 1
x= ⇒ The eigenvectors are scalar multiples of
0 0

Therefore, the geometric multiplicity is 1.

Step 3: Diagonalizability

4
Mathematical Foundations Tutorial 1

A matrix is diagonalizable if and only if the geometric multiplicity of each eigenvalue

equals its algebraic multiplicity.
Here:

Algebraic multiplicity = 2, Geometric multiplicity = 1 ⇒ A is not diagonalizable .

Analyze Level Questions (KL4)

1. Prove or disprove: “A matrix with repeated eigenvalues is always diagonalizable.”

Solution: Statement: “A matrix with repeated eigenvalues is always diagonal-

izable.”
We aim to prove or disprove this statement.

Step 1: Key Concepts

• A matrix is diagonalizable if there exists an invertible matrix P such that:

A = P DP −1

where D is a diagonal matrix consisting of the eigenvalues of A.

• An eigenvalue’s algebraic multiplicity is the number of times it appears as
a root of the characteristic polynomial.
• Its geometric multiplicity is the dimension of the eigenspace corresponding
to that eigenvalue.
• A matrix is diagonalizable if and only if the sum of the geometric multiplicities
of all its eigenvalues equals the matrix size n.

Step 2: Disproof by Counterexample

Let us consider the matrix:  
7 1
A=
0 7

Step 2.1: Find eigenvalues

7−λ 1
det(A − λI) = = (7 − λ)2 = 0 ⇒ λ = 7 (algebraic multiplicity = 2)
0 7−λ

Step 2.2: Find eigenvectors

Solve (A − 7I)x = 0:
      
0 1 x1 0 1
= ⇒ x2 = 0 ⇒ x = x1
0 0 x2 0 0

5
Mathematical Foundations Tutorial 1

So, the eigenspace is 1-dimensional.

Geometric multiplicity = 1

Step 2.3: Conclusion

• Algebraic multiplicity of λ = 7 is 2.
• Geometric multiplicity is 1.
• Since geometric multiplicity < algebraic multiplicity, the matrix is not diag-
onalizable.

Final Answer: The statement is false. A matrix with repeated eigenvalues is

not necessarily diagonalizable.

Note: Real symmetric matrices are always diagonalizable, even if they have re-
peated eigenvalues.

2. Prove: “If a real symmetric matrix has distinct eigenvalues, then its eigenvectors
are orthogonal.”

Solution: Let A ∈ Rn×n be a real symmetric matrix, i.e., AT = A, and suppose

λ1 and λ2 are two distinct eigenvalues of A, with corresponding eigenvectors v1 and
v2 , respectively.
That is,
Av1 = λ1 v1 and Av2 = λ2 v2

We want to show that v1 and v2 are orthogonal:

v1T v2 = 0

Step 1: Use symmetry of A

Since A is symmetric, we have:

(Av1 )T v2 = v1T Av2

Step 2: Substitute eigenvalue expressions

Left-hand side:
(Av1 )T v2 = (λ1 v1 )T v2 = λ1 v1T v2

Right-hand side:
v1T (Av2 ) = v1T (λ2 v2 ) = λ2 v1T v2

Step 3: Equating both sides

λ1 v1T v2 = λ2 v1T v2

Subtracting both sides:

(λ1 − λ2 )v1T v2 = 0

6
Mathematical Foundations Tutorial 1

Since λ1 ̸= λ2 , we must have:

v1T v2 = 0

Conclusion: The eigenvectors v1 and v2 are orthogonal. Therefore, if a real

symmetric matrix has distinct eigenvalues, then its eigenvectors are or-
thogonal.

3. Prove: “Every real symmetric matrix is diagonalizable.” Also, show that all its
eigenvalues are real.

Solution: Let A ∈ Rn×n be a real symmetric matrix, i.e., AT = A.

We aim to prove:

(a) All eigenvalues of A are real.

(b) A is diagonalizable.

(1) All eigenvalues are real:

Let λ ∈ C be an eigenvalue of A, with corresponding eigenvector v ∈ Cn , v ̸= 0,
such that:
Av = λv

Take the complex conjugate transpose of both sides:

vH Av = λvH v (1)

Since A is real and symmetric, we also have:

vH Av = (Av)H v = (λv)H v = λ̄vH v (2)

Equating (1) and (2):

λvH v = λ̄vH v ⇒ (λ − λ̄)vH v = 0

Since vH v > 0, it follows that:

λ = λ̄ ⇒ λ ∈ R

Hence, all eigenvalues of a real symmetric matrix are real.

(2) A is diagonalizable:
From the result above, all eigenvalues of A are real.
Let us consider two eigenvalues λ1 ̸= λ2 with corresponding eigenvectors v1 , v2 . As
shown in an earlier proof, eigenvectors of real symmetric matrices corresponding to
distinct eigenvalues are orthogonal.
Therefore:

• There exists a set of n linearly independent eigenvectors (including repeated

eigenvalues),

7
Mathematical Foundations Tutorial 1

• These can be used to form a matrix P, whose columns are orthonormal eigen-
vectors of A,
• So A = PDP−1 , with D diagonal and P−1 = PT ,
• That is, A = PDPT , i.e., A is orthogonally diagonalizable.

Conclusion: Every real symmetric matrix is diagonalizable, and all its eigenvalues
are real.
4. Prove that a matrix A is diagonalizable if and only if the sum of the geometric
multiplicities of all its eigenvalues is equal to n.

Solution: Let A ∈ Rn×n , and suppose the eigenvalues of A are

λ1 , λ2 , . . . , λk
(each repeated according to its algebraic multiplicity).
Geometric multiplicity of an eigenvalue refers to the number of linearly inde-
pendent eigenvectors associated with that eigenvalue.
Forward direction (⇒):
Suppose A is diagonalizable. Then there exists an invertible matrix P such that
A = PDP−1
where D is a diagonal matrix whose diagonal entries are the eigenvalues of A.
This means:
• A has n linearly independent eigenvectors.
• These eigenvectors must correspond to the eigenvalues of A.
• Therefore, the total number of linearly independent eigenvectors (i.e., the sum
of the geometric multiplicities) is n.

Reverse direction (⇐):

Now suppose that the sum of the geometric multiplicities of all eigenvalues of A is
equal to n.
This implies:
• There exist n linearly independent eigenvectors of A.
• These can be assembled into a matrix P, whose columns are the eigenvectors.
• Then A can be written as:
A = PDP−1
where D is a diagonal matrix of eigenvalues.
• Hence, A is diagonalizable.

Conclusion: A matrix A is diagonalizable if and only if the total number of linearly

independent eigenvectors is equal to n, which is equivalent to the statement that
the sum of geometric multiplicities of all its eigenvalues is n.