Mathematics
Eigenvalues and Eigenvectors and Their Role in Diagonalization
Quick fact
German mathematician David Hilbert first used the term 'eigenvalue' in 1904, derived from the German word 'eigen', meaning 'own' or 'characteristic', because eigenvalues reveal the matrix's own intrinsic properties.
Why this is interesting
Imagine a rubber sheet that can be stretched, sheared, or rotated. Are there any directions where the sheet is only stretched or shrunk, without being turned? The answer to that question is the heart of eigenvalues and eigenvectors.
Read the full explanation
Understanding Eigenvalues and Eigenvectors and Their Role in Diagonalization
When a matrix acts on a vector, it typically changes both its direction and its length. But for certain special vectors, the transformation is simpler: the vector is only scaled by some factor, not twisted. These vectors are called eigenvectors, and the scaling factor is the corresponding eigenvalue. Mathematically, if A is a square matrix, a nonzero vector v such that A v = λ v is an eigenvector, with eigenvalue λ. To find the eigenvalues, we solve the equation det(A - λI) = 0, called the characteristic equation. Each eigenvalue has a set of eigenvectors, forming an eigenspace. If a matrix has enough independent eigenvectors to span its space, we can use them as a new basis, which often makes the matrix's action very simple to describe.
A deeper explanation
The power of eigenvalues and eigenvectors lies in the concept of diagonalization. Suppose we can find n linearly independent eigenvectors v₁, v₂, …, vₙ for an n×n matrix A. We assemble them as columns of a matrix P, and we let D be a diagonal matrix whose entries are the corresponding eigenvalues. Then we have the beautiful formula A = P D P⁻¹. This decomposition is called diagonalization. The reason it is so useful is that powers of A become simple: Aᵏ = P Dᵏ P⁻¹, and Dᵏ is just each diagonal entry raised to the power k. Moreover, exponentials and other functions can be defined similarly. Diagonalization also breaks a matrix into independent one-dimensional actions along each eigenvector, which is why systems of linear differential equations can be solved by decoupling them into scalar equations. However, not every matrix is diagonalizable; a matrix is diagonalizable if and only if it has a full set of linearly independent eigenvectors. The factorization A = P D P⁻¹ is a special case of the more general Jordan decomposition, which handles matrices lacking a full eigenbasis. This spectral decomposition reveals the core structure of linear transformations and is a cornerstone of linear algebra.