Square b7 · Mathematics · Queenside

Eigenvalue

Eigenvalue equation: A sends v to a scalar multiple.

Meaning

An eigenvalue equation Av=λvAv=\lambda v states that a linear operator (or matrix) AA stretches a nonzero vector vv by a scalar factor λ\lambda, without changing its direction (up to sign or complex phase). The scalar λ\lambda is an eigenvalue of AA; the vector vv is an associated eigenvector. Equivalently, (A−λI)v=0(A-\lambda I)v=0 with v≠0v\neq 0, so λ\lambda is a root of the characteristic polynomial det⁡(A−λI)=0\det(A-\lambda I)=0.

Eigenvalues encode intrinsic scaling rates of AA: growth and decay rates, frequencies of oscillation, principal stretches of a deformation, and spectra of observables. The set of all eigenvalues is the spectrum of AA (in finite dimensions; infinite-dimensional operators require a broader definition).

Context

The eigenvalue idea emerged in the nineteenth century from several directions: principal axes of quadric surfaces and moments of inertia (Euler, Cauchy, Lagrange), systems of linear differential equations, and early matrix theory (Sylvester, Cayley). David Hilbert coined related spectral terminology in his work on integral equations; the German word eigen (“own” or “characteristic”) entered international usage through that tradition.

In finite dimensions, the spectral theorem for symmetric (or Hermitian) matrices guarantees an orthonormal basis of eigenvectors and real eigenvalues—the mathematical reason principal component analysis, singular value decompositions, and many physical normal-mode analyses work. For general matrices, Jordan form describes the possible failure of complete diagonalizability.

Applications are ubiquitous. In mechanics, eigenvalues of stiffness and mass matrices are squares of vibrational frequencies. In quantum mechanics, observables are self-adjoint operators and measured values are eigenvalues; the time-independent Schrödinger equation is an eigenvalue problem. In data science and network theory, eigenvalues of adjacency and Laplacian matrices describe connectivity and expansion. Google’s original PageRank algorithm is, at root, finding a dominant eigenvector of a carefully constructed matrix.

The square earns its place because Av=λvAv=\lambda v is the single most important equation of linear algebra in application: it extracts the “natural” scales and directions of a linear transformation, and those scales organize huge tracts of pure and applied mathematics.

Sources