Matrix Diagonalization: Complete Guide to A = PDP⁻¹

Quick Navigation

• What is Matrix Diagonalization?
• When is a Matrix Diagonalizable?
• How to Diagonalize a Matrix Step by Step
• Properties of Diagonalizable Matrices
• Real-World Applications
• Step-by-Step Examples
• Defective Matrices (Not Diagonalizable)
• Frequently Asked Questions
• Practice Problems
• Related Topics

---

What is Matrix Diagonalization?

Definition

Matrix diagonalization is the process of decomposing a square matrix $A$ into three specific matrices: an invertible matrix $P$, a diagonal matrix $D$, and the inverse of $P$. The relationship is defined as:

$$ \boxed{A = P D P^{-1}} $$

Where:

• $A$ is the original $n \times n$ square matrix.
• $P$ is an invertible matrix whose columns are the linearly independent eigenvectors of $A$.
• $D$ is a diagonal matrix whose diagonal entries are the corresponding eigenvalues of $A$.
• $P^{-1}$ is the inverse of matrix $P$.

Simple Example

Consider the matrix: $$ A = \left[\begin{array}{cc} 4 & 1 \\ 2 & 3 \end{array}\right] $$

Its diagonalization consists of: $$ P = \left[\begin{array}{cc} 1 & 1 \\ 1 & -2 \end{array}\right], \quad D = \left[\begin{array}{cc} 5 & 0 \\ 0 & 2 \end{array}\right] $$

Such that $A = P D P^{-1}$.

Why Is This Important?

Diagonalization reveals the intrinsic structure of a linear transformation. In the coordinate system defined by the eigenvectors (the columns of $P$), the transformation acts as simple scaling along each axis (represented by $D$). This simplifies complex operations like computing high powers of matrices or solving systems of differential equations.

---

When is a Matrix Diagonalizable?

Necessary and Sufficient Condition

An $n \times n$ matrix $A$ is diagonalizable if and only if it has $n$ linearly independent eigenvectors.

Quick Checks

Condition	Diagonalizable?	Examples
All eigenvalues are distinct	✅ Always	Most random matrices
Matrix is symmetric	✅ Always	Covariance matrices, Hessians
Matrix is diagonal	✅ Always	Identity, scalar matrices
Repeated eigenvalues with enough eigenvectors	✅ Yes	Identity matrix ($\lambda=1$ repeated)
Repeated eigenvalues with missing eigenvectors	❌ No (Defective)	Jordan blocks

The Diagonalizability Test

For each distinct eigenvalue $\lambda$:

1. Algebraic Multiplicity ($m_a$): How many times $\lambda$ appears as a root of the characteristic polynomial.
2. Geometric Multiplicity ($m_g$): The dimension of the eigenspace (number of linearly independent eigenvectors for $\lambda$).

Rule: $A$ is diagonalizable if and only if $m_g = m_a$ for every eigenvalue.

---

How to Diagonalize a Matrix Step by Step

Step 1: Find the Eigenvalues

Compute the characteristic polynomial and solve for roots: $$ \det(A - \lambda I) = 0 $$

Step 2: Find the Eigenvectors

For each eigenvalue $\lambda_i$, solve the homogeneous system: $$ (A - \lambda_i I)\vec{v} = \vec{0} $$ Find the basis vectors for the null space (eigenspace).

Step 3: Verify Diagonalizability

Count the total number of linearly independent eigenvectors found.

• If total $= n$ → Proceed.
• If total $< n$ → Matrix is not diagonalizable.

Step 4: Build Matrix $P$

Construct $P$ by placing the eigenvectors as columns: $$ P = \left[\begin{array}{cccc} \vec{v}_1 & \vec{v}_2 & \cdots & \vec{v}_n \end{array}\right] $$

Step 5: Build Matrix $D$

Construct $D$ by placing the corresponding eigenvalues on the diagonal in the same order as the eigenvectors in $P$: $$ D = \left[\begin{array}{cccc} \lambda_1 & 0 & \cdots & 0 \\ 0 & \lambda_2 & \cdots & 0 \\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & \cdots & \lambda_n \end{array}\right] $$

Step 6: Compute $P^{-1}$

Calculate the inverse of $P$.

Step 7: Verify

Confirm that $P D P^{-1} = A$.

---

Properties of Diagonalizable Matrices

1. 🔑 Powers of $A$ Are Easy

If $A = P D P^{-1}$, then raising $A$ to the power $k$ is simply: $$ A^k = P D^k P^{-1} $$ Since $D$ is diagonal, $D^k$ is just the diagonal entries raised to the $k$-th power: $$ D^k = \left[\begin{array}{cccc} \lambda_1^k & 0 & \cdots & 0 \\ 0 & \lambda_2^k & \cdots & 0 \\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & \cdots & \lambda_n^k \end{array}\right] $$

2. 📈 Matrix Exponential

For solving systems of linear ODEs ($\dot{x} = Ax$): $$ e^{At} = P e^{Dt} P^{-1} $$ Where $e^{Dt}$ has diagonal entries $e^{\lambda_1 t}, e^{\lambda_2 t}, \dots$.

3. 🔄 Change of Basis

Diagonalization represents a change of basis to the eigenvector basis. In this new basis, the linear transformation is decoupled.

4. 📊 Determinant and Trace

The determinant and trace are invariant under similarity transformations: $$ \det(A) = \det(D) = \prod_{i=1}^n \lambda_i $$ $$ \text{trace}(A) = \text{trace}(D) = \sum_{i=1}^n \lambda_i $$

---

Real-World Applications

📈 Population Dynamics (Leslie Matrix)

In population models, diagonalization helps predict long-term growth: $$ \mathbf{x}(t) = A^t \mathbf{x}(0) = P D^t P^{-1} \mathbf{x}(0) $$ The dominant eigenvalue determines the asymptotic growth rate.

📉 Markov Chains

For Markov chains, diagonalization reveals the stationary distribution (the eigenvector corresponding to $\lambda=1$) and the rate of convergence.

🔄 Difference Equations (Fibonacci Sequence)

The Fibonacci recurrence can be written as a matrix system. Diagonalization yields Binet's Formula: $$ F_n = \frac{\phi^n - \psi^n}{\sqrt{5}} $$ Where $\phi$ and $\psi$ are the eigenvalues of the Fibonacci matrix.

⚛️ Quantum Mechanics

The Hamiltonian operator $H$ is diagonalized to find energy levels (eigenvalues) and stationary states (eigenvectors).

---

Step-by-Step Examples

Example 1: 2×2 Matrix with Distinct Eigenvalues

Problem: Diagonalize $A = \left[\begin{array}{cc} 4 & 1 \\ 2 & 3 \end{array}\right]$.

Step 1: Find Eigenvalues $$ \det(A - \lambda I) = (4-\lambda)(3-\lambda) - 2 = \lambda^2 - 7\lambda + 10 = 0 $$ $$ (\lambda - 5)(\lambda - 2) = 0 \implies \lambda_1 = 5, \lambda_2 = 2 $$

Step 2: Find Eigenvectors For $\lambda_1 = 5$: $$ (A - 5I)\vec{v} = \left[\begin{array}{cc} -1 & 1 \\ 2 & -2 \end{array}\right] \vec{v} = 0 \implies v_1 = \left[\begin{array}{c} 1 \\ 1 \end{array}\right] $$ For $\lambda_2 = 2$: $$ (A - 2I)\vec{v} = \left[\begin{array}{cc} 2 & 1 \\ 2 & 1 \end{array}\right] \vec{v} = 0 \implies v_2 = \left[\begin{array}{c} 1 \\ -2 \end{array}\right] $$

Step 3: Construct P and D $$ P = \left[\begin{array}{cc} 1 & 1 \\ 1 & -2 \end{array}\right], \quad D = \left[\begin{array}{cc} 5 & 0 \\ 0 & 2 \end{array}\right] $$

Step 4: Verify $$ P^{-1} = \frac{1}{-3} \left[\begin{array}{cc} -2 & -1 \\ -1 & 1 \end{array}\right] = \left[\begin{array}{cc} 2/3 & 1/3 \\ 1/3 & -1/3 \end{array}\right] $$ Multiplying $P D P^{-1}$ returns the original matrix $A$. ✅

Example 2: Symmetric Matrix (Orthogonal Diagonalization)

Problem: Diagonalize $A = \left[\begin{array}{cc} 2 & 1 \\ 1 & 2 \end{array}\right]$.

Solution: Eigenvalues: $\lambda_1 = 3, \lambda_2 = 1$. Eigenvectors: $v_1 = \left[\begin{array}{c} 1 \\ 1 \end{array}\right], v_2 = \left[\begin{array}{c} 1 \\ -1 \end{array}\right]$.

$$ P = \left[\begin{array}{cc} 1 & 1 \\ 1 & -1 \end{array}\right], \quad D = \left[\begin{array}{cc} 3 & 0 \\ 0 & 1 \end{array}\right] $$

Note: Since $A$ is symmetric, we can normalize $P$ to make it orthogonal ($Q$), such that $A = Q D Q^T$.

Example 3: Defective Matrix (Not Diagonalizable)

Problem: Check if $A = \left[\begin{array}{cc} 1 & 1 \\ 0 & 1 \end{array}\right]$ is diagonalizable.

Solution:

1. Eigenvalues: $\det(A-\lambda I) = (1-\lambda)^2 = 0 \implies \lambda = 1$ (multiplicity 2).
2. Eigenvectors: Solve $(A-I)v=0$: $$ \left[\begin{array}{cc} 0 & 1 \\ 0 & 0 \end{array}\right] \left[\begin{array}{c} x \\ y \end{array}\right] = 0 \implies y=0 $$ Only one independent eigenvector: $v = \left[\begin{array}{c} 1 \\ 0 \end{array}\right]$.
3. Conclusion: We need 2 independent eigenvectors but only found 1. Not Diagonalizable.

---

Defective Matrices (Not Diagonalizable)

What Makes a Matrix Defective?

A matrix is defective if the geometric multiplicity of at least one eigenvalue is less than its algebraic multiplicity. These matrices cannot be diagonalized.

Jordan Normal Form

Instead of diagonalization, defective matrices can be put into Jordan Normal Form: $$ A = P J P^{-1} $$ Where $J$ is a block-diagonal matrix with eigenvalues on the diagonal and 1s on the superdiagonal.

Example of a Jordan Block: $$ J = \left[\begin{array}{ccc} \lambda & 1 & 0 \\ 0 & \lambda & 1 \\ 0 & 0 & \lambda \end{array}\right] $$

---

Frequently Asked Questions

Q: Is every square matrix diagonalizable?

A: No. Only matrices with $n$ linearly independent eigenvectors are diagonalizable. Defective matrices (like Jordan blocks) are not.

Q: Are all symmetric matrices diagonalizable?

A: Yes! This is guaranteed by the Spectral Theorem. Furthermore, they can be diagonalized by an orthogonal matrix ($P^{-1} = P^T$).

Q: Does diagonalization work for complex eigenvalues?

A: Yes. If $A$ is real but has complex eigenvalues, $P$ and $D$ will contain complex numbers.

Q: Why is diagonalization useful for computing $A^{100}$?

A: Computing $A^{100}$ directly requires 99 matrix multiplications. With diagonalization, you compute $D^{100}$ (just raise diagonal entries to the 100th power) and perform two matrix multiplications: $P D^{100} P^{-1}$.

Q: Can a singular matrix be diagonalizable?

A: Yes. A singular matrix has $\det(A)=0$, meaning at least one eigenvalue is 0. As long as it has enough eigenvectors, it is diagonalizable. Example: $\left[\begin{array}{cc} 1 & 0 \\ 0 & 0 \end{array}\right]$.

---

Practice Problems

Beginner Level

1. Diagonalize $A = \left[\begin{array}{cc} 3 & 0 \\ 0 & 5 \end{array}\right]$.
2. Find P and D for $A = \left[\begin{array}{cc} 2 & 1 \\ 0 & 3 \end{array}\right]$.

Intermediate Level

3. Diagonalize $A = \left[\begin{array}{cc} 2 & 1 \\ 1 & 2 \end{array}\right]$ and compute $A^5$.
4. Determine if diagonalizable: $A = \left[\begin{array}{ccc} 2 & 0 & 0 \\ 1 & 2 & 0 \\ 0 & 0 & 3 \end{array}\right]$.

Advanced Level

5. Prove that if $A$ and $B$ are similar ($B = P^{-1}AP$), they have the same eigenvalues.
6. For what values of $k$ is $A = \left[\begin{array}{cc} 1 & k \\ 0 & 1 \end{array}\right]$ diagonalizable?

(Solutions available in the interactive solver above)

---

Summary

Key Takeaways

Property	Diagonalizable	Not Diagonalizable
Eigenvectors	$n$ linearly independent	Fewer than $n$
Geometric Multiplicity	Equals Algebraic for all $\lambda$	Less than Algebraic for some $\lambda$
Power Computation	Easy ($A^k = P D^k P^{-1}$)	Complex (Jordan Form)
Examples	Symmetric, Distinct Eigenvalues	Jordan Blocks

Quick Reference Formula

$$ \boxed{A = P D P^{-1}} $$

• $P$: Matrix of eigenvectors.
• $D$: Diagonal matrix of eigenvalues.

---

Related Topics

Explore these related concepts in our Linear Algebra Toolbox:

• Eigenvalue/Eigenvector Calculator - Find eigenvalues and eigenvectors
• Symmetric Matrix Checker - All symmetric matrices are diagonalizable
• Positive Definite Checker - Positive definite matrices are diagonalizable
• SVD Calculator - Works for ALL matrices (even non-diagonalizable)
• Matrix Exponential Calculator - Uses diagonalization when possible
• Linear ODE System Solver - Solution uses diagonalization