Ain Shams University

Faculty of Engineering – Masters in Mechatronics Engineering

Linear Algebra

Report

Eigen values & Eigen vectors

All Eigenvalues Are Equal in a 3×3
Matrix

Name: Mohamed Maged Elsayed

ID: 2401845
1. Introduction
In linear algebra, eigenvalues and eigenvectors are essential tools in understanding
the behavior of linear transformations. For 3×3 matrix, there are three possibilities
for the nature of eigenvalues:
1. All eigenvalues are distinct.
2. Two eigenvalues are equal, and one is different.
3. All eigenvalues are equal.
This report focuses on the third case, where a matrix has three identical
eigenvalues. We will explore the implications of this situation on the matrix's
diagonalizability, the number of linearly independent eigenvectors, and the
possible matrix forms.

2. Definition and Theoretical Background

Let A be a 3×3 matrix. If all three eigenvalues of A are equal to a scalar λ\, then the
characteristic equation is:
𝑑𝑒𝑡(𝐴 − 𝜆𝐼) = 0 ⇒ (𝜆 − 𝜆)3 = 0
This means:
• The algebraic multiplicity of λ is 3.
• However, the geometric multiplicity (number of linearly independent
eigenvectors) can be 1, 2, or 3.
This has a major impact on whether the matrix can be diagonalized or not.

3. Matrix Types in This Case

3.1. Diagonal Matrix (Trivial Case)
Example:
2 0 0
𝐴 = [0 2 0 ]
0 0 2
 • All eigenvalues = 2
• 3 independent eigenvectors (standard basis)
• Diagonalizable
3.2. Non-Diagonalizable Matrix (Defective Case)
Example:
2 1 0
𝐴 = [0 2 1 ]
0 0 2
• All eigenvalues = 2
• Only 1 linearly independent eigenvector
• Not diagonalizable

4. Diagonalizability Conditions
A matrix is diagonalizable if and only if there are 3 linearly independent
eigenvectors.
For Case 3:
• Geometric multiplicity = 3 → Diagonalizable.
• Geometric multiplicity < 3 → Not diagonalizable.
This depends on the structure of the matrix, not just the eigenvalues.

5. How to Find Eigenvectors in This Case

Let’s say the matrix 𝐴 ∈ 𝑅3×3 has all eigenvalues equal to λ to find the
eigenvectors, follow these steps:

Step 1: Solve the Characteristic Equation

You already know that the eigenvalue λ appears three times (algebraic multiplicity
= 3), so:

𝑑𝑒𝑡(𝐴 − 𝜆𝐼) = 0

This confirms the eigenvalue and tells us we need to find the null space of 𝐴 − 𝜆𝐼.
Step 2: Set Up and Solve the System
Let:
(𝐴 − 𝜆𝐼)𝑥 = 0
You now solve this homogeneous system to find the eigenvectors x

• The number of linearly independent solutions (dimension of null space)

gives you geometric multiplicity.

• You do row reduction (Gaussian elimination) on (𝐴 − 𝜆𝐼) find the null

space.

Example 1:
Let’s consider:
3 1 0
𝐴 = [0 3 1 ]
0 0 3
All eigenvalues are λ=2 so compute:

0 1 0
A − 3I = [0 0 1]
0 0 0
Now solve:
(A − 3I)x = 0
𝑥
𝑙𝑒𝑡 𝑋 = [𝑦]
𝑧
0 1 0 𝑥 𝑦 0
[0 0 1] [𝑦] = [ 𝑧 ] = [0]
0 0 0 𝑧 0 0
𝑦=0
𝑧=0
 𝑥 1
𝑋𝐸𝑖𝑔𝑒𝑛 𝑉𝑒𝑐𝑡𝑜𝑟 = [ 0 ] = 𝑥 [ 0]
0 0
So, the only eigenvector is in the direction of [𝟏 𝟎 𝟎]𝑻 .
This means the geometric multiplicity is 1, and the matrix is not diagonalizable.

So, we can diagonalize the matrix using Jordan form.

Jordan form:
is an upper triangular matrix of a certain type known as a Jordan matrix that, with
respect to some base, represents a linear operator on a finite-dimensional vector
space. This type of matrix contains identical diagonal entries to the left and below
it, with each non-zero off-diagonal entry equal to 1 just above the main diagonal (on
the superdiagonal).
(𝐴 − 3𝐼)𝑣⃗1 = 0 (𝑟𝑒𝑔𝑢𝑙𝑎𝑟 𝑒𝑖𝑔𝑒𝑛𝑣𝑒𝑐𝑡𝑜𝑟)
(𝐴 − 3𝐼)𝑣⃗2 = 𝑣⃗1
(𝐴 − 3𝐼)𝑣⃗3 = 𝑣⃗2

Find 𝑣⃗2 :
We want:
1
(𝐴 − 3𝐼)𝑣⃗2 = 𝑣⃗1 = [0]
0
𝑆𝑒𝑡:
0 1 0 𝑥 𝑦
(𝐴 − 3𝐼) = [0 0 1] , 𝑣⃗2 = [𝑦] ⇒ (𝐴 − 3𝐼) 𝑣⃗2 = [ 𝑧 ]
0 0 0 𝑧 0
𝑦 1
[ 𝑧 ] = [0] ⇒ 𝑦 = 1, 𝑧 = 0, 𝑥 = 𝑤ℎ𝑎𝑡𝑠𝑜𝑒𝑣𝑒𝑟
0 0
 0
𝑃𝑖𝑐𝑘 𝑥 = 0 𝑠𝑜: 𝑣⃗2 = [1]
0
Find 𝑣⃗3 :
We want:
0
(𝐴 − 3𝐼)𝑣⃗3 = 𝑣⃗2 = [1]
0
𝑆𝑒𝑡:
𝑦 0
(𝐴 − 3𝐼)𝑣⃗3 = [ 𝑧 ] = [1] ⇒ 𝑦 = 0, 𝑧 = 1, 𝑥 = 𝑤ℎ𝑎𝑡𝑠𝑜𝑒𝑣𝑒𝑟
0 0
0
𝑃𝑖𝑐𝑘 𝑥 = 0 𝑠𝑜: 𝑣⃗3 = [0]
1

Build Generalized Eigenvector Matrix & write Jordan form:

1 0 0
𝑃 = [𝑣⃗1 𝑣⃗2 𝑣⃗3 ] = [0 1 0]
0 0 1
3 1 0
𝐽 = [ 0 3 1]
0 0 3
⸫ Jordan canonical form of A → 𝐴 = 𝑃𝐽𝑃−1
2

𝐴𝑛 = 𝐽𝑛 = (𝜆𝐼 + 𝑁)𝑛 = ∑ (𝑘 𝑛)𝜆𝑛−𝑘 𝑁 𝑘

𝑘=0

𝑛(𝑛 − 1) 𝑛−2 2
𝐴𝑛 = 3𝑛 𝐼 + 𝑛3𝑛−1 𝑁 + 3 𝑁
2
0 1 0 0 0 1
𝑁 = [0 0 1] , 𝑁 = [0 0 0] and solve for 𝐴𝑛
2

0 0 0 0 0 0
 𝑑𝑥
Solve the = 𝐴𝑋 :
𝑑𝑡

𝑥(𝑡) = 𝑒 𝐴𝑡 𝑥0 = 𝑃𝑒 𝐽𝑡 𝑃−1 𝑥0
𝑆𝑖𝑛𝑐𝑒 𝑃 = 𝐼 𝑎𝑛𝑑 𝐴 = 𝐽 , 𝑡ℎ𝑖𝑠 𝑠𝑖𝑚𝑝𝑙𝑖𝑓𝑖𝑒𝑠 𝑡𝑜:
𝑥⃗(𝑡) = 𝑒 𝐽𝑡 𝑥⃗0
Getting 𝑒 𝑗𝑡 :

𝐽𝑡 𝜆𝑡
𝑡2 2
𝑒 =𝑒 (𝐼 + 𝑡𝑁 + 𝑁 )
2
1 0 0 0 1 0 0 0 1
2
𝐼 = [0 1 0] , 𝑁 = [0 0 1] , 𝑁 = [0 0 0]
0 0 1 0 0 0 0 0 0
𝑡2
1 𝑡
2
So 𝑒 𝑗𝑡 3𝑡
= 𝑒 [0 1 𝑡]
0 0 1

𝑡2
1 𝑡
3𝑡 2
⸫ General solution: 𝑥⃗(𝑡) = 𝑒 [0 1 𝑡 ] 𝑥⃗0
0 0 1

𝑡2
1 𝑡 1 1
[ 2 ] . [0] = [0]
0 1 𝑡 0 0
0 0 1

𝑒 3𝑡 0 𝑡
⸫𝑥⃗(𝑡) = [ 0 ] 𝑎𝑛𝑑 𝑤ℎ𝑒𝑛 𝑥0 = [1] → 𝑥⃗(𝑡) = 𝑒 [1]
3𝑡

0 0 0

𝑡2
0
& 𝑥0 = [0] → 𝑥⃗(𝑡) = 𝑒 3𝑡 [ 2 ]
𝑡
1
1
⸫ The general solution will be:

𝑡 𝑡2
1
𝑥⃗(𝑡) = 𝑐1 𝑒 3𝑡 [0] + 𝑐2 𝑒 3𝑡 [1] + 𝑐3 𝑒 3𝑡 [ 2 ]
0 𝑡
0
1
General Tip:
• If after solving (𝐴 − 𝜆𝐼)𝑥 = 0 you get 3 pivot columns, the null space is trivial
→ no eigenvector.

• If you get 1 or 2 free variables, the null space has dimension 1 or 2 → you get 1
or 2 eigenvectors.

Example 2:
Let:
4 0 0
𝐴 = [ 0 4 0]
0 0 4
This is a diagonal matrix, so we already know the eigenvalues and eigenvectors:

Step 1: Find Eigenvalues:

Since it's diagonal, the eigenvalues are just the diagonal entries:
𝜆1 = 𝜆2 = 𝜆3 = 4
So, all eigenvalues are equal to 4 → this is Case 3.

Step 2: Find Eigenvectors:

To find the eigenvectors, solve:
(𝐴 − 4𝐼)𝑥 = 0
Calculate 𝐴 − 4𝐼:
 0 0 0
𝐴 − 4𝐼 = [0 0 0]
0 0 0
This is the zero matrix, so:
0 ⋅ 𝑣 = 0 ⇒ Any nonzero vector is a solution
So every vector in 𝑅3 is an eigenvector — we can choose a basis of 3 linearly
independent eigenvectors:
1 0 0
𝑥1 = [0] , 𝑥2 = [1] , 𝑥3 = [0]
0 0 1
These form a basis for 𝑅3 → so A is diagonalizable.

6. Applications and Relevance

Matrices with repeated eigenvalues appear in:

• Stability analysis in control systems

• Solutions of systems of differential equations

• Jordan canonical form theory

• Simplified transformations in graphics

Understanding whether such matrices are diagonalizable helps in simplifying matrix

powers, exponentials, and system behavior predictions.

7. Conclusion
Case 3, where all eigenvalues of a 3×3 matrix are equal, offers both simplicity and
subtlety. While some matrices in this category are trivially diagonalizable, others are
not, depending on the number of linearly independent eigenvectors.

→ Equal eigenvalues do not guarantee diagonalizability.