eigenvalue
Let be a vector space over a field , and let be anendomorphism of (meaning a linear mapping of into itself).A scalar is said to be aneigenvalue
of if there is a nonzero for which
(1) |
Geometrically, one thinks of a vector whose direction is unchangedby the action of , but whose magnitude is multiplied by .
If is finite dimensional, elementary linear algebra shows thatthere are several equivalent definitions of an eigenvalue:
(2) The linear mapping
i.e. , has no inverse.
(3) is not injective.
(4) is not surjective.
(5) , i.e. .
But if is of infinite dimension
, (5) has no meaning and theconditions (2) and (4) are not equivalent to (1).A scalar satisfying (2) (called a spectral value of) need not be an eigenvalue. Consider for example the complexvector space of all sequences
of complexnumbers
with the obvious operations
, and the map given by
Zero is a spectral value of , but clearly not an eigenvalue.
Now suppose again that is of finite dimension, say .The function
is a polynomial of degree over in thevariable , called the characteristic polynomial
of theendomorphism . (Note that some writers define the characteristicpolynomial as rather than , but thetwo have the same zeros.)
If is or any other algebraically closed field, or if and is odd, then has at least one zero, meaning that has at least one eigenvalue. In no case does have more than eigenvalues.
Although we didn’t need to do so here, one can compute the coefficientsof by introducing a basis of and the corresponding matrix for. Unfortunately, computing determinants and finding rootsof polynomials of degree are computationally messy proceduresfor even moderately large , so for most practical purposesvariations on this naive scheme are needed. See the eigenvalueproblem for more information.
If but the coefficients of are real (and in particular if has a basis for which the matrix of has only real entries), thenthe non-real eigenvalues of appear in conjugate pairs. For example,if and, for some basis, has the matrix
then , with the two zeros .
Eigenvalues are of relatively little importance in connection withan infinite-dimensional vector space, unless that space is endowed withsome additional structure, typically that of a Banach space
or Hilbert space
. But in those cases the notion is of great value inphysics, engineering, and mathematics proper. Look for “spectral theory”for more on that subject.
Title | eigenvalue |
Canonical name | Eigenvalue |
Date of creation | 2013-03-22 12:11:52 |
Last modified on | 2013-03-22 12:11:52 |
Owner | Koro (127) |
Last modified by | Koro (127) |
Numerical id | 15 |
Author | Koro (127) |
Entry type | Definition |
Classification | msc 15A18 |
Related topic | EigenvalueProblem |
Related topic | SimilarMatrix |
Related topic | Eigenvector![]() |
Related topic | SingularValueDecomposition |
Defines | eigenvalue |
Defines | spectral value |