Let A∈Fn×n be a square matrix. A scalar λ∈F is an eigenvalue of A if and only if there exists a non-zero vector v∈Fn such that multiplication by A scales v by λ.
Av=λv
Let A∈Fn×n be a square matrix and λ be an eigenvalue of A. A non-zero vector v∈Fn is an eigenvector of A corresponding to λ if and only if Av=λv. For an eigenvector v, the linear transformation v↦Av is the scalar transformation v↦λv.
Let A∈Fn×n be a square matrix. The characteristic polynomial of A is the polynomial of degree n in λ given by the determinant of A−λIn.
pA(λ):=∣A−λIn∣
The eigenvalues of A are the roots of the characteristic polynomial, since Av=λv has a non-zero solution v if and only if A−λIn is singular.
Let A∈Fn×n be a square matrix. The eigen-spectrum of A, written σA, is the set of distinct eigenvalues of A.
σA:={λ:∣A−λIn∣=0}
The characteristic polynomial of A has m distinct complex roots where 1≤m≤n, so taking eigenvalues over C the eigen-spectrum has cardinality m. A real matrix can have complex eigenvalues: the rotation matrix [01−10] has eigenvalues ±i and no real eigenvalues. The statements below that count n eigenvalues assume that all roots of the characteristic polynomial lie in F, which always holds for F=C.
Let A∈Fn×n be a square matrix and λ∈σA. The eigenvectors of A corresponding to λ are the non-zero solutions v of the equation
(A−λIn)v=0.
Let A∈Fn×n be a square matrix and λ∈σA. The eigen-space of A with respect to λ, written Eλ, is the kernel of A−λIn.
Eλ:={v∈Fn:(A−λIn)v=0}
The eigen-space is a subspace of Fn that contains the eigenvectors of A corresponding to λ together with the zero vector, and it is spanned by those eigenvectors.
Let A∈Fn×n be a square matrix and λ∈σA. The algebraic multiplicity of λ, written μA(λ), is the number of times λ appears as a root of the characteristic polynomial of A. The sum of the algebraic multiplicities of all eigenvalues of A is n.
Let A∈Fn×n be a square matrix and λ∈σA. The geometric multiplicity of λ, written γA(λ), is the dimension of the eigen-space Eλ. The geometric multiplicity is the maximum number of linearly independent eigenvectors of A corresponding to λ.
The algebraic and geometric multiplicities have the following properties:
1≤γA(λ)≤μA(λ)≤n∀λ∈σA
Eigenvectors corresponding to distinct eigenvalues are linearly independent. A repeated eigenvalue does not necessarily correspond to multiple linearly independent eigenvectors.
Let A∈Fn×n be a square matrix and λ1,…,λn∈F be its eigenvalues repeated according to algebraic multiplicity. The eigenvalues have the following properties:
i=1∑nλi=tr(A)i=1∏nλi=∣A∣
Let A∈Fn×n be a square matrix. The matrix A is defective if and only if the geometric multiplicity of some eigenvalue is strictly less than its algebraic multiplicity.
γA(λ)<μA(λ)forsomeλ∈σA
A matrix is defective if and only if it has fewer than n linearly independent eigenvectors. A matrix is non-defective if and only if γA(λ)=μA(λ) for every λ∈σA, in which case it has n linearly independent eigenvectors.
Let A∈Fn×n be a non-defective matrix. An eigen-basis of A is a basis of Fn consisting of n linearly independent eigenvectors of A.
Let A∈Fn×n be a non-defective matrix with eigen-basis v1,…,vn and corresponding eigenvalues λ1,…,λn. The eigenvector matrixP is the square matrix whose columns are the eigenvectors, and the eigenvalue matrixΛ is the diagonal matrix whose diagonal elements are the corresponding eigenvalues.
P:=↑v1↓…↑vn↓Λ:=λ1000⋱000λn
Let A∈Fn×n be a non-defective matrix with eigenvector matrix P and eigenvalue matrix Λ. The diagonalisation of A is the factorisation of A into the product of P,Λ and P−1.
A=PΛP−1
A matrix is diagonalisable if and only if it is non-defective. The diagonalisation has the property Ak=PΛkP−1 for every k∈N, where Λk is the diagonal matrix with diagonal elements λik.