Matrix types

Tridiagonal matrix

Non-zeroes on only main diagonal, and diagonals above and below it.

Hermitian matrix always reducable to tridiagonal.

  • Quick for linear algebraic solutions.
  • Cheap for storage.
  • Quick for eigenvector decomposition (see e.g. CUDA SDK).

(Semi-)definite matrix

Matrix equiv of “positive”, “nonnegative”, etc.

n \times n Hermitian matrix.

Equiv.: all eigenvalues are nonneg./positive/nonpos./negative.

Semidefinite positive: x^* M x \geq 0

Definite positive: x^* M x > 0

Hermitian matrix

m_{ij} = m^*_{ij}, where m^* is complex conjugate (imaginary portion negated).

  • \lambda_i are all real.
  • Necessarily normal.

Normal matrix

A A^* = A^* A.

Spectral theory generally applies to only normal matrices.

Conjugate ("Hermitian") transpose

(A^*)_{ij} = \bar{A_{ji}}; i.e. transpose A and take elementwise complex conjugates.

If A is a real matrix, A^* = A^T.

matrix_types.txt · Last modified: 2008/09/22 09:39 by jobriath
 
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Noncommercial-Share Alike 3.0 Unported
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki