Modular invariant theory: Difference between revisions

Content deleted Content added
No edit summary
m added wikilinks
Line 1:
In [[mathematics]], a '''modular invariant''' of a [[group (mathematics)|group]] is an invariant of a [[finite group]] [[Group action (mathematics)|acting]] on a [[vector space]] of positive characteristic (usually dividing the [[order (group theory)|order]] of the group). The study of modular invariants was originated in about 1914 by {{harvtxt|Dickson|2004}}.
 
==Dickson invariant==
 
When ''G'' is the finite [[general linear group]] GL<sub>''n''</sub>('''F'''<sub>''q''</sub>) over the [[finite field]] '''F'''<sub>''q''</sub> of order a [[prime power]] ''q'' acting on the [[ring (mathematics)|ring]] '''F'''<sub>''q''</sub>[''X''<sub>1</sub>, ...,''X''<sub>''n''</sub>] in the natural way, {{harvtxt|Dickson|1911}} found a complete set of invariants as follows. Write [''e''<sub>1</sub>, ..., ''e''<sub>''n''</sub>] for the [[determinant]] of the [[matrix (mathematics)|matrix]] whose entries are ''X''{{su|b=''i''|p=''q''<sup>''e''<sub>''j''</sub></sup>}}, where ''e''<sub>1</sub>, ..., ''e''<sub>''n''</sub> are non-negative integers[[integer]]s. For example, the [[Moore determinant over a finite field|Moore determinant]] [0,1,2] of order 3 is
 
:<math>\begin{vmatrix} x_1 & x_1^q & x_1^{q^2}\\x_2 & x_2^q & x_2^{q^2}\\x_3 & x_3^q & x_3^{q^2} \end{vmatrix}</math>
 
Then under the action of an element ''g'' of GL<sub>''n''</sub>('''F'''<sub>''q''</sub>) these determinants are all multiplied by det(''g''), so they are all invariants of SL<sub>''n''</sub>('''F'''<sub>''q''</sub>) and the ratios [''e''<sub>1</sub>, ...,''e''<sub>''n''</sub>]&thinsp;/&thinsp;[0,&nbsp;1,&nbsp;..., ''n''&nbsp;&minus;&nbspthinsp;1] are invariants of GL<sub>''n''</sub>('''F'''<sub>''q''</sub>), called '''Dickson invariants'''. Dickson proved that the full ring of invariants '''F'''<sub>''q''</sub>[''X''<sub>1</sub>, ...,''X''<sub>''n''</sub>]<sup>GL<sub>''n''</sub>('''F'''<sub>''q''</sub>)</sup> is a polynomial algebra over the ''n'' Dickson invariants [0,&nbsp;1,&nbsp;..., ''i''&nbsp;&minus;&nbsp;1,&nbsp;''i''&nbsp;+&nbsp;1,&nbsp;...,&nbsp;''n'']&thinsp;/&thinsp;[0, 1, ..., ''n''&minusnbsp;−&thinsp;1] for ''i''&nbsp;=&nbsp;0, 1, ..., ''n''&nbsp;&minus;&nbsp;1.
{{harvtxt|Steinberg|1987}} gave a shorter proof of Dickson's theorem.
 
The matrices [''e''<sub>1</sub>, ..., ''e''<sub>''n''</sub>] are divisible by all non-zero linear forms in the variables ''X''<sub>''i''</sub> with coefficients in the finite field '''F'''<sub>''q''</sub>. In particular the [[Moore determinant over a finite field|Moore determinant]] [0,&nbsp;1,&nbsp;...,&nbsp;''n''&nbsp;&minus;&nbspthinsp;1] is a product of such linear forms, taken over 1&nbsp;+&nbsp;''q''&nbsp;+&nbsp;''q''<sup>2</sup>&nbsp;+&nbsp;...&nbsp;+&nbsp;''q''<sup>''n''&nbsp;–&nbsp;1</sup> representatives of (''n''&nbsp;–&nbspthinsp;1)-dimensional [[projective space]] over the field. This factorization is similar to the factorization of the [[Vandermonde determinant]] into linear factors.
 
==See also==