1 Definition and basic concept

A right eigenvector is a nonzero vector that is transformed by a square matrix only by scaling. If a matrix \(A\) acts on a vector \(v\) and the result is the same vector multiplied by a scalar \(\lambda\), then \(v\) is a right eigenvector and \(\lambda\) is the corresponding eigenvalue. This idea captures directions that remain invariant under the action of the matrix, apart from a possible change in magnitude and sign.

1.1 Eigenvector equation

The defining equation for a right eigenvector is

\[ Av = \lambda v. \]

Here \(A\) is a square matrix, \(v\) is a vector, and \(\lambda\) is a scalar. The equation states that applying the matrix to \(v\) produces a vector parallel to \(v\). Such vectors identify special directions associated with the linear transformation represented by the matrix.

1.2 Eigenvalue and eigenvector pair

An eigenvalue and its eigenvector are usually considered together as a pair. The eigenvalue indicates the factor by which the vector is stretched, compressed, reversed, or left unchanged. Different eigenvectors can correspond to the same eigenvalue, and a single matrix may have several distinct eigenvalue-eigenvector pairs.

1.3 Nonzero vector requirement

The vector \(v\) must be nonzero. If the zero vector were allowed, the equation \(A0=\lambda 0\) would hold for every matrix and every scalar, making the concept uninformative. Requiring a nonzero vector ensures that the eigenvector represents a meaningful geometric or algebraic feature of the matrix.

2 Relation to matrix multiplication

Right eigenvectors arise from the standard convention of multiplying a matrix by a column vector on the left. This placement determines the form of the eigenvector equation and distinguishes right eigenvectors from related notions defined with row vectors.

2.1 Right multiplication convention

In linear algebra, vectors are commonly written as columns, and matrices act on them from the left. The term right eigenvector is used in this setting because the vector appears on the right side of the matrix in the product \(Av\). The terminology helps distinguish it from the left eigenvector, which is defined differently.

2.2 Column vector interpretation

When \(v\) is viewed as a column vector, the matrix \(A\) combines its coordinates into a new column vector. If \(v\) is a right eigenvector, the output remains aligned with the same column direction. This makes eigenvectors especially useful for describing invariant directions in geometry and transformations.

2.3 Comparison with left eigenvectors

A left eigenvector is typically written as a row vector \(w\) satisfying

\[ wA = \lambda w. \]

This is not the same as the right eigenvector condition, though the two are related. For many matrices, left and right eigenvectors differ unless the matrix has special symmetry properties. In applications, the right eigenvector is usually the one associated with the standard column-vector formulation.

3 Existence and properties

Not every matrix has eigenvectors over every number system, but many matrices do have them over the complex numbers. When they exist, right eigenvectors organize into subspaces tied to specific eigenvalues and reveal structural information about the matrix.

3.1 Characteristic equation

Eigenvalues are found by solving the characteristic equation

\[ \det(A-\lambda I)=0. \]

This determinant condition ensures that the matrix \(A-\lambda I\) is singular, so the homogeneous system \((A-\lambda I)v=0\) has nontrivial solutions. Those nontrivial solutions are precisely the right eigenvectors associated with \(\lambda\).

3.2 Eigenspaces

For each eigenvalue \(\lambda\), the set of all right eigenvectors associated with \(\lambda\), together with the zero vector, forms an eigenspace. This space consists of every vector that is scaled by the matrix by the same factor \(\lambda\). Eigenspaces are subspaces of the underlying vector space.

3.2.1 Basis of an eigenspace

Any eigenspace can be described by a basis, a smallest set of vectors that spans it. The basis vectors are themselves right eigenvectors. Once a basis is known, every vector in the eigenspace can be written as a linear combination of those basis vectors.

3.2.2 Dimension and geometric multiplicity

The dimension of an eigenspace is called the geometric multiplicity of the eigenvalue. It measures how many independent eigenvectors correspond to that eigenvalue. Geometric multiplicity is always at least 1 for an eigenvalue that exists, and it may be smaller than the algebraic multiplicity.

3.3 Linear independence of eigenvectors

Right eigenvectors corresponding to distinct eigenvalues are linearly independent. This property is central to diagonalization and to the construction of eigenbases. However, eigenvectors belonging to the same eigenvalue may be dependent or independent depending on the dimension of the associated eigenspace.

4 Computation of right eigenvectors

Right eigenvectors are usually computed after the eigenvalues have been identified. The process reduces to solving a homogeneous linear system for each eigenvalue.

4.1 Solving the homogeneous system

For a given eigenvalue \(\lambda\), one solves

\[ (A-\lambda I)v=0. \]

This is a system of linear equations in the components of \(v\). Any nonzero solution is a right eigenvector. The full set of solutions forms the eigenspace for \(\lambda\).

4.2 Characteristic polynomial

The characteristic polynomial is the determinant \(\det(A-\lambda I)\) expanded as a polynomial in \(\lambda\). Its roots are the eigenvalues of \(A\). Once a root is found, substitution into the homogeneous system yields the corresponding right eigenvectors.

4.3 Numerical methods

For large matrices, exact symbolic computation is often impractical. Numerical algorithms estimate eigenvalues and eigenvectors using iterative procedures, matrix factorizations, or related techniques. In applied work, these methods are favored for speed and stability, especially when the matrix is large or sparse.

5 Special cases

Certain matrices require extra care because their eigenstructure has repeated or incomplete sets of eigenvectors. In other cases, eigenvectors may be complex even when the matrix entries are real.

5.1 Repeated eigenvalues

A repeated eigenvalue occurs when a root of the characteristic polynomial appears more than once. Such an eigenvalue may have several independent right eigenvectors, or only a few. The number of independent eigenvectors is determined by the eigenspace dimension rather than the repetition count alone.

5.2 Defective matrices

A matrix is defective if it does not have enough linearly independent eigenvectors to form a basis of the whole space. In that case, it cannot be diagonalized. Defective matrices are often studied using generalized eigenvectors, which extend the eigenvector concept to complete the basis.

5.3 Complex eigenvalues and eigenvectors

A real matrix can have complex eigenvalues and complex right eigenvectors. This happens, for example, when the characteristic polynomial has no real roots. Complex eigenvectors are essential in understanding rotations, oscillations, and other transformations that do not preserve real directions.

6 Applications

Right eigenvectors appear throughout mathematics, science, and engineering because they identify invariant directions and simplify matrix actions. Their usefulness becomes especially clear in systems that evolve over time or can be decomposed into independent modes.

6.1 Matrix diagonalization

If a matrix has enough linearly independent right eigenvectors, it may be diagonalized. In that form, the matrix is represented by a diagonal matrix of eigenvalues in an eigenvector basis. Diagonalization makes powers of matrices and repeated transformations much easier to compute.

6.2 Systems of differential equations

Linear systems of differential equations often reduce to eigenvalue problems. Right eigenvectors describe the modes along which the system evolves independently. Each eigenvalue controls growth, decay, or oscillation along its corresponding eigenvector direction.

6.3 Markov chains and dynamical systems

In Markov chains, eigenvectors help describe steady-state behavior and long-term trends. In broader dynamical systems, right eigenvectors can indicate stable, unstable, or neutral directions. They are therefore useful for understanding repeated updates and equilibrium states.

6.4 Principal component analysis

In principal component analysis, eigenvectors of a covariance matrix identify directions of maximal variance. The leading right eigenvectors determine the principal components used for dimensionality reduction and data interpretation. This allows complex datasets to be summarized in a smaller number of meaningful directions.

Right eigenvectors are part of a larger family of spectral ideas in linear algebra. Several related notions help extend or complement their role in matrix theory.

7.1 Left eigenvector

A left eigenvector is a row vector satisfying \(wA=\lambda w\). It is related to the right eigenvector but arises from multiplication on the opposite side. Left eigenvectors are especially useful in dual-space formulations and in some numerical and probabilistic settings.

7.2 Generalized eigenvector

A generalized eigenvector is used when a matrix does not have enough ordinary eigenvectors. It satisfies a higher-order condition involving powers of \(A-\lambda I\). Generalized eigenvectors help build Jordan forms and complete bases for defective matrices.

7.3 Eigenbasis

An eigenbasis is a basis consisting entirely of eigenvectors. When such a basis exists, the matrix representation becomes especially simple. Not every matrix admits an eigenbasis, but when it does, many computations become straightforward.

7.4 Spectral decomposition

Spectral decomposition expresses a matrix in terms of its eigenvalues and eigenvectors. For matrices with suitable properties, this decomposition separates the matrix into components associated with distinct spectral modes. It provides both theoretical insight and practical computational tools.