Similar Matrices Have the Same Eigenvalues
Introduction
In the world of linear algebra, eigenvalues and eigenvectors are among the most fundamental concepts that every student of mathematics, physics, computer science, or engineering must understand. They serve as powerful tools for analyzing the behavior of systems, from the oscillations of a mechanical structure to the stability of a financial model. On the flip side, one of the most elegant and important results in this field is the theorem that similar matrices always share the same eigenvalues. Simply put, if you take a square matrix and transform it into another matrix that is "similar" to it—meaning they represent the same linear transformation in different coordinate systems—you can be absolutely certain that the eigenvalues of both matrices are identical Worth keeping that in mind. Nothing fancy..
And yeah — that's actually more nuanced than it sounds.
This principle is not just a mathematical curiosity; it is a cornerstone of how we understand and work with linear transformations. In this article, we will explore what similar matrices are, why eigenvalues are so important, and how the theorem that similar matrices share the same eigenvalues is proven and applied in practice Practical, not theoretical..
What Are Similar Matrices?
Before diving into the eigenvalue theorem, it is essential to understand what it means for two matrices to be similar. Two square matrices A and B are said to be similar if there exists an invertible matrix P such that:
B = P⁻¹AP
The matrix P is called the change of basis matrix. In geometric terms, if A represents a linear transformation in one coordinate system, and B represents the same transformation in a different coordinate system, then B is similar to A. The matrix P encodes the relationship between the two coordinate systems Worth knowing..
The key insight is that similarity is an equivalence relation. - Symmetry: If A is similar to B, then B is similar to A. This means it satisfies three properties:
- Reflexivity: Every matrix is similar to itself (take P = I, the identity matrix).
- Transitivity: If A is similar to B and B is similar to C, then A is similar to C.
Honestly, this part trips people up more than it should Practical, not theoretical..
Because similarity is an equivalence relation, it captures the intrinsic properties of a linear transformation that are preserved regardless of the coordinate system used to describe it.
What Are Eigenvalues and Eigenvectors?
Before exploring the theorem, let's briefly recall what eigenvalues and eigenvectors are. Given a square matrix A, an eigenvector v is a non-zero vector that satisfies:
Av = λv
Here, λ is the corresponding eigenvalue. So in plain terms, an eigenvector is a vector that, when multiplied by the matrix A, simply gets scaled by a factor of λ—it does not change direction, only its magnitude. Eigenvalues and eigenvectors reveal the "inherent" properties of a linear transformation, such as the directions in which the transformation stretches or compresses space, and by how much Still holds up..
To give you an idea, if you have a rotation matrix in 3D space, the eigenvalues tell you whether the rotation preserves any particular axis, and the eigenvectors tell you the axes along which the transformation acts Easy to understand, harder to ignore..
The Theorem: Similar Matrices Share the Same Eigenvalues
The theorem states that if A and B are similar matrices, then they have the same set of eigenvalues (with the same multiplicities). Formally:
If B = P⁻¹AP, then λ is an eigenvalue of A if and only if λ is an eigenvalue of B.
This is a powerful result because it tells us that eigenvalues are invariant under similarity transformations. Basically, no matter how you change your coordinate system, the eigenvalues of a matrix remain unchanged. This makes eigenvalues a deeply intrinsic property of the linear transformation itself, not of any particular representation Small thing, real impact..
Why Is This True?
The proof of this theorem is elegant and relies on a simple algebraic manipulation. Suppose B = P⁻¹AP and λ is an eigenvalue of A with eigenvector v. We want to show that λ is also an eigenvalue of B with some eigenvector Not complicated — just consistent..
Starting from Av = λv, we multiply both sides on the left by P⁻¹:
P⁻¹Av = λP⁻¹v
Since B = P⁻¹AP, we can substitute:
BP⁻¹v = λP⁻¹v
Let w = P⁻¹v. Then:
Bw = λw
This shows that w is an eigenvector of B with the same eigenvalue λ. The converse can be shown similarly, establishing the "if and only if" nature of the theorem.
This proof also reveals why eigenvalues are invariant: the transformation P⁻¹AP is simply a change of basis, and eigenvalues are properties of the transformation itself, not of the basis But it adds up..
Step-by-Step Breakdown
Let's walk through the concept more systematically:
- Define the matrices: Start with two square matrices A and B that are related by similarity: B = P⁻¹AP.
- Recall the eigenvalue equation: For matrix A, an eigenvalue λ satisfies Av = λv.
- Apply the similarity transformation: Multiply both sides of the eigenvalue equation by P⁻¹.
- Substitute for B: Replace P⁻¹AP with B.
- Identify the new eigenvector: The resulting equation shows that w = P⁻¹v is an eigenvector of B with the same eigenvalue λ.
- Conclude: Since every eigenvalue of A has a corresponding eigenvector in B, and vice versa, the eigenvalues are identical.
This step-by-step process makes the theorem accessible and demonstrates that the invariance of eigenvalues under similarity is a direct consequence of the algebraic structure of the transformation.
Real-World Examples
Example 1: Rotation and Scaling
Consider a 2D linear transformation that first scales space by a factor of 2 and then rotates it by 90 degrees. In one coordinate system, this might be represented by the matrix:
A = [[2, 0], [0, 2]]
If we change to a rotated coordinate system using a rotation matrix P, the same transformation is represented by:
B = P⁻¹AP
Both A and B will have the same eigenvalues: λ = 2 (with multiplicity 2). The eigenvalues tell us that the transformation scales space uniformly by a factor of 2, regardless of the coordinate system.
Example 2: Diagonalization
If a matrix A is diagonalizable, it can be written as A = PDP⁻¹, where D is a diagonal matrix of eigenvalues. Plus, any matrix B similar to A will also be expressible as B = QDQ⁻¹ for some invertible Q. Since D contains the eigenvalues, B has the same eigenvalues as A.
Example 3: Practical Application in Quantum Mechanics
In quantum mechanics, the state of a physical system is described by a vector in a Hilbert space, and observables are represented by Hermitian matrices. But when we change the basis of the Hilbert space (a similarity transformation), the eigenvalues of the observable matrix remain the same. This is why, for instance, the energy levels of a quantum system are invariant under a change of basis—they are intrinsic properties of the system And that's really what it comes down to..
Scientific and Theoretical Perspective
The theorem that similar matrices share the
The theorem that similar matrices share the same eigenvalues, determinant, trace, and characteristic polynomial.
From a theoretical standpoint, similarity captures the idea that a linear operator is independent of the particular basis chosen to represent it. When we replace the basis vectors with a new, invertible set, the matrix that describes the operator changes by the similarity transformation (B = P^{-1}AP). Which means all quantities that are defined solely in terms of the operator—such as its spectrum, the product of its eigenvalues (the determinant), the sum of its eigenvalues (the trace), and the coefficients of its characteristic polynomial—remain unchanged. So naturally, two matrices that are similar belong to the same similarity class and encode the same intrinsic properties of the underlying transformation Took long enough..
In practice, this invariance underpins many computational and conceptual tools. Consider this: likewise, in numerical linear algebra, algorithms that compute eigenvalues or decompose a matrix often rely on similarity transformations to improve conditioning or to exploit special structures (e. Still, the Jordan canonical form, for instance, provides a unique representative for each similarity class; any matrix similar to a given one can be reduced to this form, making it possible to classify matrices up to essential structure rather than superficial changes of basis. Worth adding: in control theory, the eigenvalues of the system matrix dictate stability, so similarity transformations—used when redesigning coordinate systems—do not alter the system’s qualitative behavior. Consider this: g. , tridiagonalization) Small thing, real impact..
No fluff here — just what actually works.
Beyond the immediate algebraic consequences, the preservation of these invariants offers a powerful lens for interpreting physical and mathematical phenomena. In quantum mechanics, as mentioned earlier, changing the basis of the state space corresponds to a similarity transformation of the observable’s matrix; the unchanged eigenvalues correspond to measurable quantities such as energy levels, which must be basis‑independent. In differential equations, similarity can be used to simplify the coefficient matrix while preserving the solution space, allowing one to focus on the essential dynamics rather than on an arbitrary representation Surprisingly effective..
Boiling it down, the fact that similar matrices share a complete set of invariant characteristics—eigenvalues, determinant, trace, and characteristic polynomial—highlights the fundamental role of similarity as a bridge between different representations of the same linear operator. This invariance not only consolidates theoretical results across diverse fields but also provides a reliable foundation for practical applications where basis‑free insight is essential.
