Next: codim standard basis [integer
Up: Computing numerical information
Previous: Computing numerical information
Computes the Hilbert series of the module M presented by the
given standard basis. By the theory of standard (Gröbner)
bases, the Hilbert series of the module generated by the columns
of the standard basis is the same as the Hilbert series of the
monomial submodule of lead monomials of each column of the standard
basis. Define the Hilbert series of a graded module M by:
It turns out that
can be written in two useful ways: First,
, where Q(t) is a polynomial in t, and n
is the number of variables in the polynomial ring. Q(t) is
intimately linked with the graded betti numbers of the module M.
Second,
, where P(t) is a polynomial,
and
.
The output of hilb is
- the polynomial Q(t),
- the polynomial P(t),
- the codimension of M, i.e.
, - the degree of M, i.e. P(1),
- the (sectional) genus of M. If
, this is the arithmetic
genus of M.
Some notes:
- The ring variables must all have weight 1, or the results mean
very little.
- By using interrupts (see section 7.4), it is possible
to stop a Hilbert series computation. It is not possible to
then restart it, except at the beginning.
- If the Macaulay parameter ``verbose'' is set to true, then
some feedback is given. Each time a new generator is considered,
a ``.'' is displayed. Bear in mind that the process
is not linear! The worst case is for each new element to take
about twice as long as the last one (roughly). Usually, the
process works nicely for less than 200 or so standard basis
elements, and for more than 1000 generators it almost never works
(at least in 30 or 40 variables). If you just need the codimension
or degree, use the two commands below, since they work well on any
(I think!) number of generators.
For example, for the rational quartic curve, j, in
,
% putstd j
; 1
; 4
; c3-bd2
; bc-ad
; b3-a2c
; ac2-b2d
% hilb j
; 1 t 0
; -1 t 2
; -3 t 3
; 4 t 4
; -1 t 5
; 1 t 0
; 2 t 1
; 2 t 2
; -1 t 3
; codimension = 2
; degree = 4
; genus = 0
In this example,
, and
.
Next: codim standard basis [integer
Up: Computing numerical information
Previous: Computing numerical information
Sorin Popescu
Fri Feb 14 17:37:19 EST 1997