Free resolutions
talk at the
ACAGM Summer School Leuven 2011
-Hans Sch ¨onemann
hannes@mathematik.uni-kl.de
Department of Mathematics University of Kaiserslautern
Syzygies
Definition
Let I = {g1, . . . , gq} ⊆ K[x]r.
The module of syzygies syz(I) is
ker(K[x]q → K[x]r),Xwiei 7→
X
wigi
Lemma The module of syzygies of I is
(g1(x) − er+1, . . . , gq(y) − er+q) ∩ {0}r × K[x]q
in (K[x1, . . . , xm]/J)q .
ring R=0,(x,y,z),(c,dp); ideal I=maxideal(1);
// the syzygies of the (x,y,z) syz(I);
Generalization of Gröbner bases to Modules
monomials mei ∈ Rr, i = 1..r, m monomial in R monomial orderings
L(f) leading monomial, L(M) leading submodule NF(f|M) normal form
GB(M) Gröbner basis Buchbergers algorithm
Generalization of Monomial Orderings to Modules
In what follows, let F be the free R-module F = Rs with its canonical basis e1, . . . , es.
Definition. A monomial in F is a monomial in R times a basis vector
of F, that is, an element of the form xαei. A term in F is a monomial
in F times a scalar, that is, an element of type axαei, where a ∈ K. A
submodule of F which is generated by monomials is called a
monomial submodule. A monomial order on F may be defined in the
same way as a monomial ordering on R. That is, it is a total order > on the set of monomials in F satisfying
xαei > xβej =⇒ xγxαei > xγxβej for each γ ∈ Nn.
We require in addition that
xαei > xβei ⇐⇒ xαej > xβej ,∀i, j = 1, . . . , s.
Generalization of Monomial Ordering to Modules
Important orderings: term over position
ring R=....,(dp,c); ring R=....,(dp,C);
position over term
ring R=....,(c,dp); ring R=....,(C,dp);
Capital C sorts generators in ascending order, i.e., gen(1) < gen(2) < ....
A small c sorts in descending order, i.e., gen(1) > gen(2) > .... Ordering, ..., C ) is the default.
Generalization of Gröbner Bases to Modules
Finally, given a monomial order on F, we define the leading term, the leading coefficient, the leading monomial, and the tail of an element of
F as we did for a polynomial in R.
With this basic notation, the whole concept of Gröbner bases including its fundamental algorithms extend.
Free Resolutions
Let I = (g1, ..., gs) ⊆ Rr and M = Rr/I. A free resolution of M is a
long exact sequence
... −→ F2 −→ F1 −→ F0 −→ M −→ 0,
where the columns of the matrix A1generate I,the columns of A2
generate the kernel of A1 etc. Note that resolutions need not to be
Hilbert Syzygy Theorem
Hilbert Syzygy Theorem
For an ideal I ∈ K[x1...xn] there exists a free resolution of length
smaller or equal than n.
For a module M = K[x1...xn]r/N there exists a free resolution of
length smaller or equal than n.
Free Resolutions
By iterating the process of computing syzygies, starting from a finitely generated R-module N, we get what is called a free resolution of M = Rr/N:
ring r=0,(x,y,z),dp; ideal I=x,y,z;
list Ir=res(I,0);
// print the results: Ir;
list Im=mres(I,0);
// print the results: Im;
list Is=sres(std(I),0); // print the results: Is;
Intersection of Modules
Syzygies can also be used to compute ideal intersections and ideal quotients.
Given ideals I = hf1, . . . , fri and J = hg1, . . . , gsi of R, compute the
syzygies on the columns of the matrix
1 f1 . . . fr 0 . . . 0
1 0 . . . 0 g1 . . . gs
! .
The entries of the first row of the resulting syzygy matrix generate I ∩ J.
The Geometry-Algebra Dictionary
Ideal quotient I
In general, given two algebraic sets, it is not their difference but the smallest algebraic set containing the difference which can be
described by polynomial equations.
Definition. If I, J are two ideals of a ring R, the set
I : J = {f ∈ R | f g ∈ I for all g ∈ J}
is an ideal of R containing I, the ideal quotient of I by J. Similarly, the
set I : J∞ := {f ∈ R | f Jm ⊂ I for some m ≥ 1} = ∞ [ m=1 (I : Jm)
The Geometry-Algebra Dictionary
Ideal quotient II
Theorem. If K = K, and I, J are ideals of R, then
V (I) \ V (J) = V (I : J∞
). If I is a radical ideal, then
V (I) \ V (J) = V (I : J).
The Geometry-Algebra Dictionary
Saturation
Since the polynomial ring R is Noetherian by Hilbert’s basis theorem, and since I : Jm = (I : Jm−1
) : J for any two ideals I, J ⊂ R, the computation of the saturation
I : J∞ = ∞ [ m=1 (I : Jm)
just means to iterate the computation of ideal quotients. Indeed, the ascending chain
I : J ⊂ I : J2 ⊂ · · · ⊂ I : Jm ⊂ . . .
Ideal Quotient: Computation
In the same way as with the intersection, we get generators for I : J from the matrix
g1 f1 . . . fr 0 . . . 0 g2 0 . . . 0 f1 . . . fr 0 . . . 0 .. . . .. gs 0 . . . 0 f1 . . . fr .
Ideal quotient: Example
Example > ring R = 2, (x,y,z), dp; > poly F = x5+y5+(x-y)ˆ2*xyz; > ideal J = jacob(F); > J;//-> J[1]=x4+x2yz+y3z J[2]=y4+x3z+xy2z J[3]=x3y+xy3 > maxideal(2);
//-> _[1]=z2 _[2]=yz _[3]=y2 _[4]=xz _[5]=xy _[6]=x2 > ideal K = quotient(J,maxideal(2));
> K;
//-> K[1]=y4+x3z+xy2z K[2]=x3y+xy3 K[3]=x4+x2yz+y3z //-> K[4]=x3z2+x2yz2+xy2z2+y3z2 K[5]=x2y2z+x2yz2+y3z2 //-> K[6]=x2y3
Ideal quotient: Example
> K = quotient(K,maxideal(2)); > K; K[1]=x3+x2y+xy2+y3 K[2]=y4+x2yz+y3z K[3]=x2y2+y4 > K = quotient(K,maxideal(2)); > K; K[1]=x3+x2y+xy2+y3 K[2]=y4+x2yz+y3z K[3]=x2y2+y4Saturation: Example
> LIB "elim.lib";
> int p = printlevel;
> printlevel = 2; // print more information while computing > sat(J,maxideal(2));
// compute quotient 1 // compute quotient 2 // compute quotient 3
// saturation becomes stable after 2 iteration(s) [1]: _[1]=x3+x2y+xy2+y3 _[2]=y4+x2yz+y3z _[3]=x2y2+y4 [2]: 2
Summary of Operations with Ideals
sum of ideals (intersection of algebraic sets)
Summary of Operations with Ideals
sum of ideals (intersection of algebraic sets) intersection of ideals (union of algebraic sets)
Summary of Operations with Ideals
sum of ideals (intersection of algebraic sets) intersection of ideals (union of algebraic sets)
elimination of variables (projection of algebraic sets)
Summary of Operations with Ideals
sum of ideals (intersection of algebraic sets) intersection of ideals (union of algebraic sets)
elimination of variables (projection of algebraic sets) ideal quotient/saturation (“difference“ of algebraic sets)
Summary of Operations with Ideals
sum of ideals (intersection of algebraic sets) intersection of ideals (union of algebraic sets)
elimination of variables (projection of algebraic sets) ideal quotient/saturation (“difference“ of algebraic sets) Hilbert function
dimension of the ideals (dimension of the algebraic set)