Frobenius normal form

From Wikipedia, the free encyclopedia

This is an old revision of this page, as edited by Dysprosia (talk | contribs) at 12:57, 21 October 2004 (style). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Jump to navigation Jump to search

In linear algebra, the Frobenius normal form (also known as the rational canonical form) of a matrix is a normal form, that consists of the direct sum of companion matrices which in turn correspond to certain monic polynomials. One of these may be the minimal polynomial of the matrix, or the minimal polynomial may be the same as the characteristic polynomial for this matrix, which means that the Frobenius normal form of the matrix will be the companion matrix of the characteristic polynomial.

Since every matrix has a characteristic polynomial, then every matrix can be transformed by a similarity transformation to its Frobenius normal form.

Motivating example

For example, consider:

The characteristic polynomial of A is x6+6x4+12x2+8 = (x2 +2 )3 = p(x). The minimal polynomial of this matrix is x2 +2 .

We will then have one block in the Frobenius normal form as

The characteristic polynomial of A1 is indeed x2 +2.

Since p factors into solely x2 +2 terms, we expect the other two blocks comprising the normal form of the matrix to be identical to A1. So, we can simply write down the Frobenius normal form:

The characteristic and minimal polynomials of F are the same to that of A, which we would expect, since F can be obtained via a similarity transformation P−1AP=F, and determinants are similarity invariant. For this matrix A, P is

General case

In general, if M is some n×n matrix, there is an invertible matrix P such that P−1MP=F where

where the Mis are the companion matrices of what are known as the invariant polynomials of M. If Mi is the companion matrix of a polynomial fi(x), and the characteristic polynomial of M is p(x), then

Each fi divides fi+1, if one writes the companion matrices Mi such that as i increases, so does the number of rows/columns of Mi. The polynomial ft then is the minimal polynomial of M.

Algorithms