A generalized eigenvalue problem asks for a nonzero vector and scalar satisfying . When is invertible, it is equivalent to the ordinary eigenvalue problem .
For a regular matrix pencil , an eigenvalue at infinity is represented by a homogeneous pair with and . The eigenvector relation then requires with .
For and , the first coordinate has finite eigenvalue one and the second has an infinite eigenvalue. More generally, if are and the pencil is regular, its infinite algebraic multiplicity is . A singular pencil, whose determinant is identically zero, requires a different structural analysis and cannot be treated as an ordinary list of finite and infinite eigenvalues.[ref]
A generalized eigenvalue is a scalar for which has a nonzero solution . For a regular square pencil this is equivalent, at finite , to .
For and , the generalized eigenvalues are two and three. If is invertible, these are the eigenvalues of , although explicitly forming that inverse is not necessary to solve the problem. Singular can introduce an eigenvalue at infinity, while a singular pencil is a separate case. For symmetric and positive-definite symmetric , transforming with gives a symmetric ordinary eigenproblem and hence real eigenvalues.[ref]
For square matrices of the same size, their generalized characteristic polynomial is . If it is not identically zero, the pair is a regular matrix pencil, and its finite generalized eigenvalues are precisely the roots of : singularity of gives a nonzero vector with .
If is invertible, has full degree and reduces to an ordinary characteristic polynomial. A singular can lower the degree, accounting for generalized eigenvalues at infinity in the projective description. For example, , gives , with one finite generalized eigenvalue and one eigenvalue at infinity. If is identically zero, a singular-pencil analysis is needed instead of treating this zero polynomial as an ordinary degree- characteristic polynomial.

Ancestors (6)

  1. Operator theory
  2. Linear algebra
  3. Algebra
  4. Area of mathematics
  5. Mathematics
  6. Home

Discussion (0) New discussion

There are no discussions about this article yet.

Articles by others on the same topic (0)

There are currently no matching articles.