Download as pdf or txt
Download as pdf or txt
You are on page 1of 29

Lecture Notes on Grobner Basis

Theory

by

Prof Shiv Datt Kumar


Department of Mathematics

Motilal Nehru National Institute of Technology

Allahabad (UP), India, Pin-211004


E-mail: sdt@mnnit.ac.in
Shiv Datt Kumar Elementary Notions of Affine Variety

2
Chapter 1

Elementary Notions of

Affine Variety

1.1 Affine Variety

Definition 1.1.1. Let K be a field and let f1 , . . . , fs be polynomials in K[x1 , . . . , xn ].

Then V (f1 , . . . , fs ) = {(a1 , . . . , an ) ∈ K n |fi (a1 , . . . , an ) = 0, for all i = 1, 2 . . . , s}

is an affine variety defined by f1 , . . . , fs . This affine variety V (f1 , . . . , fs ) ⊆ K n

is the set of all solutions of the system of equations f1 (x1 , . . . , xn ) = . . . =

fs (x1 , . . . , xn ) = 0.

For example, graphs of polynomial functions are affine varieties.

(i) V (x2 + y 2 − 1) is circle of radius 1.


x3 − 1
(ii) V (xy − x3 + 1) = graph of y = .
x
Lemma 1.1.1. If V, W ⊆ K n are affine varieties, then V ∪ W and V ∩ W

are also affine varieties.

Proof. Suppose V = V (f1 , . . . , fr ) and W = V (g1 , . . . , gt ).

Claim: V ∩ W = V (f1 , . . . fr , g1 . . . , gt ) and V ∪ W = V (fi gj |1 ≤ i ≤ r, 1 ≤

j ≤ t).

3
Shiv Datt Kumar Elementary Notions of Affine Variety

(i) First equation is trivial to prove, being in V ∩ W means that both

f1 , . . . , fs and g1 , . . . , gt vanish, which is the same as f1 , . . . , fs , g1 , . . . , gt van-

ishing.

(ii) If (a1 , . . . , an ) ∈ V , then all of the fi , vanish at this point, which implies

that all of the fi gj also vanish at (a1 , . . . , an ). Therefore V ⊆ V (fi gj ) and

W ⊆ V (fi gj ) follows similarly. This proves V ∪ W ⊆ V (fi gj ). Now choose

(a1 , . . . , an ) ∈ V (fi gj ). If this lies in V , we are done and if not, there exists

fio (a1 , . . . , an ) 6= 0 for some i0 . Since fi0 gj vanish at (a1 , . . . , an ), ∀ j, then gj

must vanish at this point i.e. (a1 , . . . , an ) ∈ W . Hence V (fi gj ) ⊆ V ∪ W .

Example 1.1.2. Sketch the affine varieties:

(i) V (x2 − y 2 − 1)

(ii) V (x2 + y 2 + 22 − 1)

(iii) V (x2 y + xy 2 − x4 − y 4 )

(iv) V (x2 + y 2 − z 2 )

Remark 1.1.1. Finite intersection and finite union of affine varieties are again

affine varieties.

Ideal A subset I ⊆ K[x1 , . . . , xn ] is an ideal if it satisfies

(i) 0 ∈ I

(ii) If f, g ∈ I, then f + g ∈ I

(iii) If f ∈ I and h ∈ K[x1 . . . xn ], then hf ∈ I.

Definition 1.1.3. Let f1 , . . . , fs be polynomials in K[x1 , . . . , xn ]. Then we set


Ps
< f1 , . . . , fn >= { i=1 hi fi |h1 , . . . , hs ∈ K[x1 , . . . , xn ]}

Lemma 1.1.2. If f1 , . . . , fs ∈ K[x1 , . . . , xn ], then (f1 , . . . , fs ) is an ideal of

K[x1 , . . . , xn ]. We call (f1 , . . . , fs ) the ideal generated by f1 , . . . , fs .

Ps Ps
Proof. 0 ∈ (f1 , . . . , fs ), since 0 = i=1 0.fi . Note that f = i=1 pi fi & g =
Ps P
i=1 gi fi , and if h ∈ K[x1 , . . . , xn ]. Then f + g = i=1 (pi + qi )fi , hf =
P
i=1 (hpi )fi completes the proof that (f1 , . . . , fs ) is an ideal.

4
Shiv Datt Kumar Elementary Notions of Affine Variety

Proposition 1.1.3. If f1 , . . . , fs and g1 , . . . , gt are bases of the same ideal in

K[x1 , . . . , xn ], then (f1 , . . . , fs ) = (g1 , . . . , gt ), and

V (f1 , . . . , fs ) = V (g1 , . . . , gt ).

Example 1.1.4. Find V (2x2 + 3y 2 − 11, x2 − y 2 − 3). Since < (2x2 + 3y 2 −

11, x2 − y 2 − 3) >=< x2 − 4, y 2 − 1 >, then V (2x2 + 3y 2 − 11, x2 − y 2 − 3) =

V (x2 − 4, y 2 − 1) = {+2, +1}.

Remark 1.1.2. Ability to change the basis of the ideal without affecting the

variety is very important.

Question- How affine varieties give rise to an interesting class of ideals.

Definition 1.1.5. Let V ⊆ K n be an affine variety. Then the set I(V ) = {f ∈

K[x1 , . . . , xn ]|f (a1 , . . . , an ) = 0, ∀(a1 , . . . , an ) ∈ V }

Example 1.1.6. I(V ) is an ideal .

Proof. Let f, g ∈ I(V ), h ∈ K[x1 , . . . , xn ]. Let (a1 , . . . , an ) ∈ V . Thus

f (a1 , . . . , an ) + g(a1 , . . . , an ) = 0 + 0 = 0. h(a1 , . . . , an )f (a1 , . . . , an )

= h(a1 , . . . , an ).0 = 0.

Example 1.1.7. 1. I({(0, 0)}) =< x, y >.

2. I(K n ) = {0}, when K is an infinite field.

Note: Polynomials → variety → ideal.

f1 , . . . , fs → V (f1 , . . . , fs ) → I(V (f1 , . . . , fs ))

Lemma 1.1.4. If f1 , . . . , fs ∈ K[x1 , . . . , xn ], then (f1 . . . fs ) ⊂ I(V (f1 , . . . , fs )).

Although equality need not occur.

Ps
Proof. Let f ∈ (f1 , . . . , fs ). Then f = hi fi , where hi ∈ K[x1 , . . . , xn ].
i=1
Ps
Since f1 , . . . , fs vanish on V (f1 , . . . , fs ), so must i=1 hi fi . Thus f vanish on

V (f1 , . . . , fs ) which proves f ∈ I(V (f1 , . . . , fs )).

Example 1.1.8. (x2 , y 2 ) ⊆ I(V (x2 , y 2 )) but it is not equal. For if x2 = y 2 = 0,

then V (x2 , y 2 ) = {(0, 0)}. I(V (x2 , y 2 )) = (x, y) but x ∈


/ (x2 , y 2 ) since for

5
Shiv Datt Kumar Elementary Notions of Affine Variety

polynomial of the form h1 (x, y)x2 + h2 (x, y)y 2 , the degree of the monomial is at

least two.

Proposition 1.1.5. Let V and W be affine varieties in K n . Then

(i) V ⊆ W iff I(V ) ⊃ I(W ).

(ii) V = W iff I(V ) = I(W ).

Proposition 1.1.6. Let V be an algebraic set in K n . Then V is irreducible iff

I(X) is a prime ideal.

Proof. Assume V is irreducible and set V1 = V ∩V (f ), V2 = V ∩V (g). These are

affine varieties because intersection of affine varieties is again affine variety. Then

f g ∈ I(V ) ⇒ V = V1 ∪ V2 . Since V is irreducible, we have V = V1 or V = V2 .

Say V = V1 = V ∩ V (f ). Then f vanish on V so that f ∈ I(V ). Thus I(V )

is prime. Conversely assume I(V ) is a prime ideal. Let V = V1 ∪ V2 . Suppose

V 6= V1 . We claim that I(V ) = I(V2 ). To prove this note that I(V ) ⊆ I(V2 ),

since V2 ⊆ V . For the opposite inclusion, first note that I(V ) ( I(V1 ), since

V1 ( V . Thus we can pick f ∈ I(V1 ) − I(V ). Now take any g ∈ I(V2 ). Since

V = V1 ∪ V2 , it follows that f g vanishes on v and hence f g ∈ I(V ) but I(V )

is prime ideal. Thus f or g lies in I(V ), but f ∈


/ I(V ) so g ∈ I(V ) ⇒ I(V ) =

I(V2 ) ⇒ V = V2 .

1.2 Parametrization

Parametrization of affine variety Consider the system of equation

x + y + z = 1, x + 2y − z = 3 .........................(1)

Geometrically this represents the line in R3 . It follows that there are in-

finitely many solutions. Equation (1) can be written equivalently as

x + 3z = 1, y − 2z = 2 .............................(2)

6
Shiv Datt Kumar Elementary Notions of Affine Variety

Letting z = t, where t is arbitrary, equation (2) can be written as

x = −1 − 3t

y = 2 + 2t

z = t, where t ∈ R is a parameter. This is called is parametrization of (1).

Now consider x2 + y 2 = 1, common way to parametrize the circle is x = cost

y = sint. There is another way to parametrize the circle.


1−t2 2t
x= 1+t2 , y= 1+t2 . Check that x2 + y 2 = 1.

Remark 1.2.1. This parametrization does not describe the whole circle. Since
1−t2
x= 1+t2 can never be equal to −1. The point (−1, 0) is not covered.

Definition 1.2.1. Let K be field. A rational function in t1 , . . . , tn coefficient K


f h
is a quotient f /g, f.g ∈ K[x1 , . . . , xn ], g 6= 0. g = k if f k = gh. K(t1 , . . . , tn ) =

{ fg |f, g ∈ K[t1 , . . . , tn ], g 6= 0}

Definition 1.2.2. Let V = V (f1 , . . . , fs ) ⊂ K n be a variety. Then ratio-

nal parametric representation of V consists of rational function r1 , r2 , . . . rn ∈

K(t1 , . . . , tn ) such that the points given by x1 = r1 (t1 , . . . , tm ), . . . xn = rn (t1 , . . . , tm )

lies in V .

Remark 1.2.2. (1) Parametrization may not cover all points.

(2) In many situations r1 , . . . , rn are polynomials rather than rational func-

tions that is why we call a polynomials parametric representation.

Questions: (1) Does every affine variety has a rational parametric representa-

tion? Answer - No

(2) Given a parametric representation of an affine variety can we find defining

equation. (Can we find an implicit representation) Answer- Yes.

Implicit representation: The original defining equations f1 = 0, . . . , = fs = 0

of variety V are called implicit representation of V .

Example 1.2.3. Twisted cubic carve V (y − x2 , z − x3 ). This is a 3 dimensional

curve. Given one point on it we can draw the tangent line at that point. Now

imagine taking the tangent lines at all points on the twisted cubic. These tangent

lines generate the tangent space of the twisted cubic.

7
Shiv Datt Kumar Elementary Notions of Affine Variety

To convert this geometric description into more algebraic parametrization

set x = t in y − x2 = z − x3 = 0 which gives x = t, y = t2 , z = t3 , we write

this as r(t) = (t, t2 , t3 ). Now fix a particular value of t, which gives us a point

on the curve r0 (t) = (1, 2t, 3t2 ). It follows that tangent line is parametrized by

r(t) + ur0 (t) = (t, t2 , t3 ) + u(1, 2t, 3t2 ) = (t + u, t2 + 2tu, t3 + 3t2 u), where u is

a parameter that moves along the tangent line. If we now allow t to vary, then

we can parametrize the entire tangent surface by

x = t + u, y = t2 + 2tu, z = t3 + 2t2 u, where t, u have following the

interpretation: t tells position on the curve and u tells position on the tangent

line. This parametrization was used to draw the picture of the tangent surface

presented earlier.

1.3 Hilbert Nullstellensatz

Fundamental theorem of Algebra (Gauss): Establishes the basic link be-

tween Algebra and geometry: It says that a polynomial in one variable over C,

are algebraic object, is determined up to scalar factor by the set of its roots with

multiplicity, a geometric object. Nullstellensatz extend this link to certain ideals

of polynomials in many variables. It is a formal consequence of the fundamental

theorem of algebra in the sense that it holds for any algebraically closed field.

Polynomials and Affine spaces: A polynomial f ∈ K[x1 , . . . , xn ] with coeffi-

cient in field K defines a function f : K n → K by f (x1 , . . . , xn ) = f (a1 , . . . , an ).

The function f is called a polynomial function on n dimensional vector space

K n over K. If K is infinite, then no polynomial function other then zero can

vanish identically on K n .

It follows that if K is infinite, then distinct polynomials define distinct func-

tions. Thus we may regard the polynomial ring K[x1 , . . . , xn ] as the ring of poly-

nomial functions on K n is usually called affine n- space over K written An (K)

or simply An . Given a subset I ⊆ K[x1 , . . . , xn ], we defined a corresponding al-

gebraic subset of K n to be V (I) or Z(I) = {(a1 , . . . , an ) ∈ K n |f (a1 , . . . , an ) =

8
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

0, ∀f ∈ I}. Such algebraic sets are called affine algebraic sets.

Irreducible algebraic set A non empty algebraic set is called irreducible

if it not the union of two non empty smaller algebraic subsets. Irreducible

algebraic sets are called algebraic varieties. Given any set V ⊆ K n , we define

I(V ) = {f ∈ K[x1 , . . . , xn ]|f (a1 , . . . , an ) = 0, ∀(a1 , . . . , an ) ∈ V } is ideal of

K n.

Radical ideal If R is a ring and I ⊆ R is an ideal, then the set rad(I) = {f ∈

R|f m ∈ I, for some m} is an ideal, called radical of I.

If f m = 0 and g n = 0, then (af + bg)m+n = 0.

An ideal I called radical ideal if I = rad (I).

Example 1.3.1. R/I is a reduced ring iff I is a radical ideal.

Definition 1.3.2. A ring is called reduced if it’s only nilpotent element is 0.

Remark 1.3.1. Not every radical ideal can occur as I(V ), for example. I =
R[x] ∼
(x2 + 1)| ⊆ R[x] is radical because = C is reduced but V (I) = φ.
(x2 + 1)
So I is not of the form I(V ) for any V . If K is algebraically closed, then

situation is better. For examples every polynomial in one variable is a product

of linear factors and a polynomial f ∈ K[x] generates a radical ideal iff it has

no multiple roots. In this case if V is the set of roots of f , then I(V ) = (f ).

Hilbert Nullstellensatz (1893) extends this to the polynomial ring with many

variables.

Theorem 1.3.1. Let K be an algebraically closed field. If I ⊆ K[x1 , . . . , xn ] is



an ideal, then I(V (I)) = rad (I) = I.

The correspondence I → V (I) and V → I(V ) induces a bijection between

collection of algebraic subsets of An (k) = K n and radical ideals of K[x1 , . . . , xn ].



Proof. Suppose f ∈ I, then f m ∈ I for some m. Hence f m vanishes on

V (I) which implies that f ∈ V (I). Conversely, suppose f ∈ I(V (I)). Then by

definition f vanishes on V (I). Then there exists an integer m ≥ 1 such that


√ √
f m ∈ I ⇒ f ∈ I, since f is arbitrary, I(V (I)) ⊆ I.

9
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

10
Chapter 2

Basic Notions of Grobner

Basis Theory

2.1 Primary objectives to study Grobner basis

Let K be a field. Then primary objectives to study Grobner basis is


K[x1 , x2 , . . . , xn ]
(i) to understand factor rings .
I
(ii) to perform calculations in factor rings.

Since elements in factor rings are cosets of the type f + I, then first problem

is to represent the elements in unique way (normal form).

Example 2.1.1. If I = (g(x)), and f (x) ∈ K[x]. Then f (x) = q(x)g(x) + r(x),
K[x]
Here r(x) is unique. Thus each element in has representation f1 (x)+ <
g(x)
g(x) >, with deg(f1 (x)) < deg(g(x)).

K[x]
Example 2.1.2. Suppose R = . Then to find normal form of
(x2 − x + 1)
x4 + 2x + 3, we note that x2 − x + 1+ < x2 − x + 1 >= 0+ < x2 − x + 1 >.

Then x2 + < x2 − x + 1 >= (x − 1)+ < x2 − x + 1 >. So we can replace x2 by

x − 1. Similarly x2 h(x) can be replaced by (x − 1)h(x). Hence x4 + 2x + 3+ <

x2 − x + 1 >= (x − 1)(x − 1) + 2x + 3+ < x2 − x + 1 >= x2 + 4+ < x2 − x + 1 >=

(x + 3)+ < x2 − x + 1 >.

11
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

K[x, y]
Example 2.1.3. Let R = . Here there are two choices x2 → xy or
(x2 − xy)
xy → x2 . First choice gives normal form g1 (y) + xg2 (y)+ < x2 − xy >. Second

choice gives normal form g(y) + xf (x)+ < x2 − xy >, which one to use will

depend on which kind of questions we to want answer. In the first choice normal

form will be linear combination of 1, y d , xy d−1 , d ≥ 1. In the second case we get

the normal form as linear combination of 1, y d , xd , d ≥ 1. Number of monomials

used in each degree is same for both choices and this is not a coincidence.

2.1.1 Problems

1. Ideal description problem: Does every ideal I ⊆ K[x1 , . . . , xn ] has fi-

nite generating set i.e. can we write I =< f1 , . . . , fs > for some f1 , . . . , fs ∈

K[x1 , . . . , xn ].

2. Ideal membership problem: Given f ∈ K[x1 , . . . , xn ] and an ideal

I =< f1 , . . . , fs >, determine if f ∈ I. Geometrically this is closely related

to the problem of determining whether V (f1 , . . . , fs ) lies on V (f ). Ideal

membership problem follows from division algorithm.

3. Problem of solving polynomial equations: Find all common solu-

tions of polynomial equations

f1 (x1 , . . . , xn ) = 0, f2 (x1 , . . . , xn ) = 0, . . . , fs (x1 , . . . , xn ) = 0.

4. Implicitization: Let V be a subset of K n . Find a system of polynomial

equations in xi that determine the variety.

2.2 Monomial orderings:

Let K be a field and A = K[x1 , . . . , xn ]. An element x1 i1 . . . xn in is called a

monomial and an element cx1 i1 . . . xn in , where c ∈ K − {0} is called a term.

Let M be set of all monomials in A. An order < on M is called an admissible

order if it is a total order which is compatible with multiplication of monomials

i.e.

12
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

(i) For every pair of monomials m, n, we have m < n or n < m or m = n

(ii) If m1 < m2 and m2 < m3 , then m1 < m3

(iii) 1 < m, for all monomial m 6= 1.

(iv) If m1 < m2 , then mm1 < mm2 , for any monomial m ∈ M.

Example 2.2.1. (Admissible orderings)

1. Lexicographical ordering (lex): Here x1 i1 . . . xn in < x1 j1 . . . xn jn if

i1 = j1 , . . . , ik = jk , ik+1 < jk+1 for some k. In other words m < n if the

first variable with different exponents in m and n has lower exponent in

m. For e.g. x1 a xb2 < x1 c xd2 , if a < c or if a = c, then b < d.

2. Degree lexicographical ordering (deglex or grlex): Here

m = x1 i1 . . . xn in < x1 j1 . . . xn jn = n if deg(m) = i1 + . . . + in < j1 + . . . +

jn = deg(n) or if deg(m) = deg(n), then m < n in lex ordering.

3. Degree reverse lexicographical ordering (degrevlex or grevlex):

Here m = x1 i1 . . . xn in < x1 j1 . . . xn jn = n if deg(m) = i1 + . . . + in <

j1 + . . . + jn = deg(n) or if deg(m) = deg(n) and in = jn , in−1 =

jn−1 , . . . , ik = jk , ik−1 > jk−1 , for some k. In other words m < n if

deg(m) < deg(n) or if deg(m) = deg(n), then last variable with different

exponents in m and n has higher exponent in m.

Remark 2.2.1. : There are different ways to vary these ordering to get new

ones. In all above examples, we have x1 > x2 > . . . > xn of the variable. The

orderings of variables can be varied in n! ways.

Example 2.2.2. Let f = 4xy 2 z + 4z 2 − 5x3 + 7x2 z 2 ∈ K[x, y, z]. In lex order

f = −5x3 + 7x2 z 2 + 4xy 2 z + 4z 2 . In grlex order f = 7x2 z 2 + 4xy 2 z − 5x3 + 4z 2 .

In grevlex order f = 4xy 2 z + 7x2 z 2 − 5x3 + 4z 2 .

Remark 2.2.2. : Monomial x1 i1 . . . xn in form n-tupples of exponent (i1 , . . . , in ) ∈

Zn≥0 . This observation establishes correspondence between monomials in K[x1 , . . . , xn ]

and Zn≥0 .

13
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

aα xα be a non-zero polynomial in K[x1 , . . . , xn ]


P
Definition 2.2.3. Let f = α

and < be a monomial order. Then

1. Multidegree of f = multideg(f ) = max{α ∈ Zn≥0 |aα 6= 0}.

2. Leading coefficient of f = LC(f ) = amultideg(f ) ∈ K.

3. Leading monomial of f = LM (f ) = xmultideg(f ) .

4. Leading term of f = LT (f ) = LC(f ).LM (f )

Example 2.2.4. Let f = 4xy 2 z + 4z 2 − 5x3 + 7x2 z 2 ∈ K[x, y, z]. In lex order

>, multideg(f ) = (3, 0, 0) = 3 + 0 + 0 = 3, LC(f ) = −5, LM (f ) = x3 ,

LT (f ) = −5x3 .

2.3 Monomial ideals

Definition 2.3.1. An ideal I ⊆ K[x1 , . . . , xn ] is monomial ideal if there is a

subset A ⊂ Zn≥0 (possibly infinite) such that I consists of all polynomials which

are finite sums of the form α∈A hα xα , where hα ∈ K[x1 , . . . , xn ]. In this case
P

we write I =< xα : α ∈ A >.

Lemma 2.3.1. Let I =< xα : α ∈ A >, where A ⊂ Zn≥0 , is a monomial ideal.

Then a monomial xβ ∈ I iff xβ is divisible by xα , for some α ∈ A.

Proof. If xβ is divisible by xα , then xβ = xγ xα , for some γ ∈ Zn≥0 . Then by

definition of ideal, this implies xβ ∈ I =< xα : α ∈ A >. Conversely, if xβ ∈ I,


Ps
then xβ = i=1 hi xα(i) , hi ∈ K[x1 , . . . , xn ] and α(i) ∈ A. By expanding each

hi as linear combination of monomials, we see that every term on the right side

of the equation is divisible by xα(i) . Hence the left hand side xβ must have the

same property.

Lemma 2.3.2. Let I be a monomial ideal and let f ∈ K[x1 , . . . , xn ]. Then the

following are equivalent:

1. f ∈ I

14
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

2. Every term of f lies in I.

3. f is a K-linear combination of the monomials in I.

Proof. (3) ⇒ (2) ⇒ (1). Suppose f is a K-linear combination of the monomials

in I. Then f = α∈A hα xα , where hα ∈ K[x1 , . . . , xn ].Then every term of f


P

lies in I and hence f ∈ I.

(1) ⇒ (3). Suppose f ∈ I. Then by definition of the monomial ideal f =


α
P
α∈A hα x , where hα ∈ K[x1 , . . . , xn ], which is a K[x1 , . . . , xn ] linear combina-

tion of monomials. After expanding this can be treated as K-linear combination

of monomials in I.

Definition 2.3.2. Let R be ring and I, J are ideals. Then (I : J) = {α ∈

R|αJ ⊂ I} is called ideal quotient.

Lemma 2.3.3. (Dickson Lemma) A monomial ideal I ⊆ K[x1 , . . . , xn ] is finitely

generated.

Proof. By induction on n. Since every ideal in K[x] is principal, lemma is true

for n = 1. Suppose it is true for n − 1 variables. Let I be a monomial ideal in

K[x1 , . . . , xn ]. Let Ji = (I :< xin >) ∩ K[x1 , . . . , xn−1 ] =< Si >. Since Ji is an

ideal in K[x1 , . . . , xn−1 ], we can choose Si , to be finite. We have

J0 ⊆ J1 ⊆ J2 ⊆ . . .

It follows easily that ∪Ji is an ideal J in K[x1 , . . . , xn−1 ] and hence finitely

generated, let J =< S >, where S is a set of monomials in K[x1 , . . . , xn−1 ]. If


0
m ∈ I is a monomial, then m = m xkn for some monomial
0 0 0
m ∈ K[x1 , . . . , xn−1 ] and m xkn ∈ I for some k. Also m ∈ (I :< xkn >), so
0
m ∈< xkn Sk >. Thus S = S0 ∪ xn S1 ∪ x2n S2 ∪ . . . is a generating set for I. But

for some r, we have that Sk = Sr if k ≥ r, So S0 ∪ xn S1 ∪ x2n S2 ∪ . . . ∪ xrn Sr is

a finite generating set for I.

15
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

Example 2.3.3. Let I =< x4 y 2 , x3 y 4 , x2 y 5 >. Projection is J =< x2 >⊂

K[x]. Since x2 y 5 ∈ I, we have m = 5. Then we get slices in Jk , 0 ≤ k ≤

4 = m − 1 generated by monomials y k . Also Jk =< xα > such that xα y k ∈ I,

J0 = J1 = (0), J2 = J3 = (x4 ), J4 =< x3 >. These slices are the ideal slices.

Then by Dickson Lemma I =< x2 y 5 , x4 y 2 , x4 y 3 , x3 y 4 >.

2.4 Division in K[X1 , . . . , Xn ]

Example 2.4.1. 1. Find the remainder when f = xy 2 + 1, is divided by

f1 = xy + 1, f2 = y + 1 using lex order x > y. Note that f = xy 2 + 1 =

y(xy + 1) + (−1)(y + 1) + 2. Remainder= 2.

2. Divide f = x2 y + xy 2 + y 2 by f1 = xy − 1, f2 = y 2 − 1 with lex order.

With lex order x > y, we have x2 y + xy 2 + y 2 = (x + y)(xy − 1) + 1(y 2 −

1) + (y + x + 1). With lex order y > x, we have x2 y + xy 2 + y 2 =

(x + 1)(y 2 − 1) + x(xy − 1) + (2x + 1).

Theorem 2.4.1. (Division Algorithm) Fix a monomial order > on Zn≥0 and let

F = (f1 , . . . , fs ) be an ordered s-tupple of polynomials in K[x1 , . . . , xn ]. Then

for every f ∈ K[x1 , . . . , xn ], we have f = a1 f1 + . . . + as fs + r, where ai ∈

K[x1 , . . . , xn ] and either r = 0 or r is K-linear combination of monomials, none

of which are divisible by lt(f1 ), . . . , lt(fs ). Furthermore, if ai fi 6= 0, then we

have multideg(f ) > multideg(ai fi ).

Proof. We prove the existence of a1 , . . . , as and r by giving an algorithm for

their construction and it correctly operates on any input.

INPUT f1 , . . . , fs , f

OUTPUT a1 , . . . , as , r such that f = a1 f1 + . . . + as fs + r,

Initialization : a1 = 0, . . . , as = 0, h := f

WHILE h 6= 0 DO

i := 1

IF there exists i such that LM (fi ) divides LM (h)

16
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

While i ≤ s AND division occurred = false

DO

IF LT (fi ) divides LT (h) THEN


LT (h)
ai := ai +
LT (fi )
LT (h)
hi := h − fi
LT (fi )
division occurred:= true

Else i = i + 1

If division occurred = false. Then

r := r + LT (h)

h := h − LT (h)

Remark 2.4.1. Division algorithm solves the membership problem. Remainder

r = 0 is sufficient condition for the membership problem. For example: If

f1 = xy + 1, f2 = y 2 − 1, f = xy 2 − x, then with F = (f1 , f2 ), on division

of f by F , we get f = yf1 + 0f2 + (−x − y). With F = (f2 , f1 ), we have

xy 2 −x = x(y 2 −1)+0(xy+1)+0. The second calculation shows that f ∈ (f1 , f2 ).

The first calculation shows that even if f ∈ (f1 , f2 ), it is still possible to obtain

a nonzero remainder on division by F = (f1 , f2 ). The division algorithm for

many variables is imperfect generalization of its one variable. To remedy this

situation, we go for a different generating set for ideal I, so that on division by

these generators remainder is uniquely determined and condition r = 0 should

be equivalent to membership in the ideal. Grobner bases have exactly these

properties.

2.5 Grobner basis

Definition 2.5.1. ( Grobner basis) Fix a monomial order. A finite subset

G = {g1 , . . . , gt } of an ideal I is said to be a Grobner basis (or standard basis)

if < LT (g1 ), . . . , LT (gt ) >=< LT (I) >.

17
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

Equivalently, a set {g1 , . . . , gt } ⊆ I is a Grobner basis of I iff the leading

term of any element of I is divisible by one of LT (gi ) i.e. for all f ∈ I, f 6= 0,

there exists gi such that LT (gi ) divides LT (f ).

Example 2.5.2. Let I =< f1 , f2 >=< x3 − 2xy, x2 y − 2y 2 + x >. Then

{f1 , f2 } is not a Grobner basis with respect to grlex order for x2 ∈ LT (I) but

x2 ∈<
/ LT (f1 ), LT (f2 ) >, for xf2 − yf1 = x2 .

Example 2.5.3. Let I = (g1 , g2 ) = (x + z, y − z). Then g1 , g2 is a Grobner

basis using lex order in R[x, y, z]. (LT (g1 ), LT (g2 )) = (x, y). This is equivalent

to show that the initial form of any non-zero element of I is divisible by either

x or y. To prove this, consider f = ag1 + bg2 ∈ I. Suppose contrary that

f 6= 0 and neither x | LT (f ) nor y | LT (f ). Then by definition of lex order

f must be a polynomial in z alone. However f vanishes on linear subspace

L = V (x+z, y −z) ⊂ R3 , f ∈ I. It is easy to check that (x, y, z) = (−t, t, t) ∈ L,

for any real number t. The only polynomial in z alone that vanishes at all points

is a zero polynomial. Contradiction. It follows that (g1 , g2 ) is a Grobner basis.

Properties of Grobner basis

Proposition 2.5.1. Let G = {g1 , . . . gt } be a Grobner basis for an ideal I ⊆

K[x1 , . . . , xn ] and let f ∈ I. Then there is a unique r ∈ K[x1 , . . . , xn ] with the

following two properties.

1. No term of r is divisible by any one of the LT (g1 ), . . . , LT (gt ).

2. There exists a g ∈ I such that f = g + r.

Proof. The division algorithm gives f = a1 g1 +. . .+at gt +r. Let g = a1 g1 +. . .+

at gt . Then f = g + r. Uniqueness: Suppose f = G1 + r1 = G2 + r2 satisfy (1)

and (2), then r2 − r1 = G1 − G2 ∈ I. If r2 6= r1 , then LT (r2 − r1 ) ∈ LT (I) =<

LT (g1 ), . . . , LT (gt ) >. This implies that LT (gi )|LT (r2 − r1 ) for some i. This is

impossible since no term of r1 and r2 is divisible by LT (g1 ), . . . , LT (gt ). Thus

r2 − r1 must be zero. Thus r1 = r2 ⇒ G1 = G2 .

18
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

Corollary 2.5.2. Let G = {g1 , . . . , gt } be a Grobner basis for an ideal I ⊆

K[x1 , . . . , xn ] and let f ∈ K[x1 , . . . , xn ]. Then f ∈ I iff remainder of f on

division by G is zero.

Proof. Remainder r = 0 iff f ∈ I. Conversely let f ∈ I. Then f = f + 0

which satisfies two conditions of previous proposition. It follows that zero is the

remainder of f on division by G.

Definition 2.5.4. Let I be a non zero ideal of K[x1 , . . . , xn ]. Then LT (I) =

leading terms of I = {cxα |∃f ∈ I with LT (f ) = cxα )}.

Remark 2.5.1. If I =< f1 , . . . , fr >, then < LT (f1 ), . . . , LT (fs ) >⊆< LT (I) >.

However < LT (I) > can be strictly larger.

Example 2.5.5. Let I =< f1 , f2 >, where f1 = x3 −2xy, f2 = x2 y−2y 2 +x and

use grlex order in K[x, y]. Then x(x2 y − 2y 2 + x) − y(x3 − 2xy) = x2 ⇒ x2 ∈ I.

Thus x2 = LT (x2 ) ∈< LT (I) >. However x2 is not divisible by LT (f1 ) = x3

or LT (f2 ) = x2 y, so that x2 ∈<


/ LT (f1 ), LT (f2 ) >.

Proposition 2.5.3. Let I ⊆ K[x1 , . . . , xn ] be an ideal. Then

1. < LT (I) > is a monomial ideal.

2. There are g1 , . . . , gt ∈ I such that < LT (I) >=< LT (g1 ), . . . , LT (gt ) >.

Proof. (i) The leading monomials LM (g) of elements g ∈ I − {0} generate the

monomial ideal < LM (g) : g ∈ I − {0} >, Since LM (g) and LT (g) differ by

constants, the ideal is equal to < LT (g) : g ∈ I − {0} >. Thus LT (I) is a

monomial ideal.

(ii) Since monomial ideal is generated by LM (g), for g ∈ I − {0}. By

Dickson lemma there are finitely many g1 , g2 , . . . , gt such that < LT (I) >=<

LM (g1 ), . . . , LM (gt ) >. Since LM (g) and LT (g) differ by constants, it follows

that < LT (I) >=< LT (g1 ), . . . , LT (gt ) >.

Theorem 2.5.4. (Hilbert basis theorem) Every ideal I ⊆ K[x1 . . . xn ] has

finite generating set i.e. I =< g1 , . . . , gt >, for some by g1 , . . . , gt ∈ I.

19
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

Proof. If I =< 0 >, nothing to do. If I 6= (0), then by Proposition 2.5.3, there

exist g1 , . . . , gt ∈ I such that < LT (I) >=< LT (g1 ), . . . , LT (gt ) >.

Claim: I =< g1 , . . . , gt >. Note that < g1 , . . . , gt >⊆ I, for gi ∈ I. Now

let f ∈ I. Then by division algorithm f = a1 g1 + . . . + at gt + r, where each

term in r is not divisible by any term in LT (g1 ), . . . , LT (gt ). If r 6= 0, then

LT (r) ∈< LT (I) >=< LT (g1 ), . . . , LT (gt ) >. It follows that LT (r) must

be divisible by some LT (gi ). Contradiction. It follows that r = 0 and f =


Pt
i=1 ai gi ∈< g1 , . . . , gt >.

F
Definition 2.5.6. We write f for the remainder of f on division by the

ordered s-tupple F = (f1 , . . . , fs ). If F is a Grobner basis for (f1 , . . . , fs ),

then we regard F as a set (without partial order). Then there is a unique

r ∈ k[x1 , . . . , xn ] such that

1. no term of r is divisible by any leading term of the elements of F .

2. there is an element g ∈ I such that f = g + r.

Example 2.5.7. Let F = (x2 y − y 2 , x4 y 2 − y 2 ) ⊆ K[x, y], using lex order. We

have (x5 y)F = xy 3 , for division algorithm yields x5 y = (x3 + xy)(x2 y − y 2 ) +

0(x4 y 2 − y 2 ) + xy 3 .

2.6 S-polynomial

Definition 2.6.1. Let f, g ∈ K[x1 , . . . , xn ] be nonzero polynomials. Then the

S-polynomial of f and g is
f g
S(f, g) = LCM (LM (f ), LM (g))[ − ].
LT (f ) LT (g)
Example 2.6.2. Let f = x3 y 2 − x2 y 3 + x, g = 3x4 y + y 2 ∈ R[x, y] with

grlex order LCM (LM (f ), LM (g)) = LCM (x3 y 2 , x4 y) = x4 y 2 , LT (f ) = x3 y 2 ,

LT (g) = 3x4 y. Then


x4 y 2 x4 y 2
S(f, g) = 3 2 f − 4 g = xf − 13 y.g = −x3 y 3 + x2 − 31 y 3 .
x y 3x y
Remark 2.6.1. Note that S-polynomial S(f, g) is designed to produce cancella-

tion of leading terms.

20
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

Pt αi
Lemma 2.6.1. Suppose we have a sum of the form i=1 ci x gi , where c1 , c2 , . . . , ct

are constants and αi + multideg(gi )δ ∈ Zn≥0 , whenever ci 6= 0. If multideg


Pt Pt
( i=1 ci xαi gi ) < δ, then there are constants cjk such that αi
i=1 ci x gi =
δ−yjk
S(gj , gk ), where xγjk = LCM (LM (gj )), LM (gk )). Further more,
P
j,k Cjk x

each xδ−γjk S(gj , gk ) has multi degree < δ.

Proof. Let di = LC(gi ), so that ci di is the leading coefficient of ci xαi gi . Since

ci xαi gi have multi degree δ and their sum has strictly smaller multi degree, it
P xαi gi
follows that ci di = 0. Define pi = and that pi has leading coefficient
di
Pt Pt
1. Consider the telescopic sum i=1 ci xαi gi = i=1 ci di pi = c1 d1 (p1 − p2 ) +

(c1 d1 + c2 d2 )(p2 − p3 ) + (c1 d1 + . . . + ct−1 dt−1 )(p(t−1) − pt ) + (c1 d1 + . . . + ct dt )pt .

Now let LT (gi ) = di xβi . By assumption, we have αi + βi = δ for all i. Then

LM (gi ) = xβi divides xδ and consequently xγjk = LCM (LM (gj ), LM (gk )) also

divides xδ . Thus xδ−yik is a monomial and


xγjk xγjk
xδ−yjk S(gj , gk ) = xδ−yjk ( gj − gk )
LT (gj ) LT (gk )
xδ xδ
= β
gj − gk (∗)
dj x j dk xβk
xαj xαk Pt
= gj − gk = pj − pk . Using this equation and i=1 ci di = 0, the above
dj dk
Pt
telescopic sum can be written as i=1 ci xαi gi = c1 d1 xδ−γ12 S(g1 , g2 ) + (c1 d1 +

c2 d2 )xδ−γ23 S(g2 , g3 ) + . . . + (c1 d1 + . . . + ct−1 dt−1 )xδ−γ(t−1)t S(gt−1 , gt ), which is

a sum of the desired form. Since pj and pk have multideg δ and leading term

=1, the difference pj − pk has multideg < δ. By equation (∗) same is true of

xδ−yjk S(gj , gk ). This proves the lemma.

Theorem 2.6.2. Let I be a non-zero ideal of K[x1 , . . . , xn ]. Then for the set of

non-zero polynomials G = {g1 , g2 , . . . , gt } ⊆ I the following are equivalent.

1. G is a Grobner basis for I.

G
2. f ∈ I iff f −
→+ 0.

Pt
3. f ∈ I iff f = i=1 hi gi with lp(f ) = max1≤i≤t (lp(hi )lp(gi )).

4. LT (G) = LT (I)

21
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

Proof. (1) ⇒ (2) Let f ∈ K[x1 , . . . , xn ]. Then by division algorithm ∃ r ∈


G
K[x1 , . . . , xn ] reduced with respect to G such that f −
→+ r. Thus f − r ∈ I and

thus f ∈ I iff r ∈ I. Clearly if r = 0, then f ∈ I. Conversely if f ∈ I and r 6= 0,

then r ∈ I and by (1) there exists i ∈ {1, 2, . . . , t} such that lp(gi ) | lp(r). This

is a contradiction to the fact that r is reduced with respect to G. Thus r = 0


G
and f −
→+ 0.
G
(2) ⇒ (3) For f ∈ I, we know by hypothesis that f −
→+ 0 and since the process

of reduction is same as the division algorithm, (3) follows.

(3) ⇒ (4) Clearly LT (G) ⊆ LT (I). For the reverse inclusion, let f ∈ I, then by
Pt
(3), f = i=1 hi gi with lp(f ) = max1≤i≤t (lp(hi )lp(gi )). Thus LT (f ) ∈ LT (G)

and
X
LT (f ) = LT (hi )LT (gi ),
i

where sum is over all i such that lp(f ) = lp(hi )lp(gi ). Hence LT (I) ⊆ LT (G).

Therefore LT (G) = LT (I).

(4) ⇒ (1) Let f ∈ I. Then LT (f ) ∈ LT (G).

t
X
LT (f ) = hi LT (gi ), f or some hi ∈ K[x1 , . . . , xn ]. (2.1)
i=1

Expanding RHS of 2.1, we observe that each term is divisible by some lp(gi ).

Hence LHS LT (f ) is also divisible by some lp(gi ) as required.

Lemma 2.6.3. Let f1 , . . . , fs ∈ K[x1 , . . . , xn ] be such that lp(fi ) = X 6= 0 for


Ps
all i = 1, 2, . . . , s. Let f = i=1 ci fi , with ci ∈ K, i = 1, 2, . . . , s. If lp(f ) < X,

then f is linear combination of S(fi , fj ) with coefficient in K, 1 ≤ i < j ≤ s.

Proof. Write fi = ai X+ lower terms, ai ∈ K. Then the hypothesis says that


Ps 1 1
i=1 ci ai = 0, for the ci are in K. Now by definition S(fi , fj ) = fi − fj ,
ai aj
for lp(fi ) = lp(fj ) = X. Thus
1 1
f = c1 f1 + . . . + cs fs = c1 a1 ( f1 ) + . . . + cs as ( fs )
a1 as
1 1 1 1
= c1 a1 ( f1 − f2 ) + (c1 a1 + c2 a2 )( f2 − f3 ) + . . . + (c1 a1 + . . . +
a1 a2 a2 a3
1 1 1
cs−1 as−1 )( fs−1 − fs ) + (c1 a1 + . . . + cs as ) fs .
as−1 as as

22
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

= c1 a1 S(f1 , f2 )+(c1 a1 +c1 a2 )S(f2 , f3 )+. . .+(c1 a1 +. . .+cs−1 as−1 )S(fs−1 , fs ),

for c1 a1 + . . . + cs as = 0.

Theorem 2.6.4. Let I be a polynomial ideal. Then the basis G = {g1 , . . . , gt }

is a Grobner basis for I iff for all pairs (i, j), i 6= j the remainder on division of

S(gi , gj ) by G (listed in some order) is zero.

Proof. If G is a Grobner basis, then the remainder on division by G is zero,

for S(gi , gj ) ∈ I. Conversely, let f ∈ I be a non-zero polynomial. To show if

all the S-polynomials have zero remainders on division by G, then LT (f ) ∈<

LT (g1 ), . . . , LT (gt ) >. Given f ∈ I =< g1 , . . . , gt >, there are polynomials hi ∈


Pt
K[x1 , . . . , xn ] such that f = i=1 hi gi . Let mi = multideg (hi gi ) and define X

= max (m1 , . . . , ms ). Then multideg (f ) ≤ max (multideg (hi gi )) = X. If

X = lp(f ), we are done. If X < lp(f ), we find a representation of f with smaller

X and this will be a contradiction. Let S = {i|lp(hi )lp(gi ) = X}. For i ∈ S,


P
write hi = ci Xi + lower terms. Set gi = i∈S ci Xi gi . Then lp(Xi gi ) = X, for

all i ∈ S, but lp(g) < X. By Lemma 2.6.3, there exist dij ∈ K such that

X
g= dij S(Xi gi , Xj gj ).
i,j∈S,i6=j

Now X = lcm(LM (Xi gi , Xj gj ), Thus

X X
S(Xi gi , Xj gj ) = Xi gi − Xj gj
LT (Xi gi ) LT (Xj gj )

X X X
S(Xi gi , Xj gj ) = gi − gj = S(gi , gj ),
LT (gi ) LT (gj ) Xij
G
where Xij = lcm(LM (gi ), LM (gj )). By hypothesis S(gi , gj ) −
→+ 0 and then
G
from the last equation we have S(Xi gi , Xj gj ) −
→+ 0. This gives representation

t
X
S(Xi gi , Xj gj ) = hijr gr ,
i=1

23
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

where by division algorithm theorem,

max1≤r≤t (lp(hijr )lp(gr )) = lp(S(Xi gi , Xj gj )) < max(lp(Xi gi ), lp(Xj gj )) = X.

Pt
Substituting these expressions above and g into f , we get f = i=1 hi 0 gi , with
0
max1≤r≤t (lp(Xi gi ), lp(hi ) < X. This is contradiction.

Example 2.6.3. Let I = (y−x2 , z−x3 ) be twisted cubic curve in R3 . Then G =

{y − x2 , z − x3 } is a Grobner basis for lex order y > z > x. S-polynomial, S(y −


yz
x2 , z −x3 ) = yz 2 3 2 3
y (y −x )− z (z −x ) = −zx +yx . Using the division algorithm
G
we find −zx2 +yx3 = x3 (y −x2 )+(−x2 )(z −x3 )+0. So S(y − x2 , z − x3 ) = 0.

Then G is a Grobner basis for I.

Remark 2.6.2. G is not a Grobner basis in the above example when x > y > z

lex order is used.

2.7 Buchberger Algorithm

Theorem 2.7.1. Let I =< f1 , . . . , fs >6= (0) be a polynomial ideal. Then Grob-

ner basis for I can be constructed in a finite number of steps by the following

algorithm:

Input F = (f1 , . . . , fs )

Output : a Grobner basis G = (g1 , . . . , gt ) for I, with F ⊂ G.

G := F

REPEAT:
0
G := G
0
For each pair {p, q}, p 6= q in G ,

DO
0
G
S := S(p, q)

IF S 6= 0, THEN G := G ∪ {S}

24
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

UNTIL
0
G=G.

Proof. Note that G ⊂ I at every stage of the algorithm. This is initially true
0
G
and whenever we enlarge G, we do so by adding the remainder S = S(p, q) for
0
p, q ∈ G . Thus if G ⊂ I, then p, q and hence S(p, q) are in I and since we are
0
dividing by G ⊂ I, we get G∪{S} ⊂ I. Also note that G contains the given basis
0
F of I, so that G is actually a basis of I. The algorithm terminates when G = G
0
G 0
which means that S(p, q) = 0 for all p, q ∈ G . Hence G is a Grobner basis

of I. To prove that algorithm terminates. We need to consider what happens


0
after each step when we pass through the main loop. The set G consists of G
0
(old G) together with non zero remainder of S-polynomial elements of G . Then
0 0 0
< LT (G ) >⊆< LT (G) >, for G ⊆ G. Furthermore, if G 6= G, then we claim

that < LT (G0 ) >(< LT (G) >. Suppose remainder r(6= 0) has been attached
0 0
to G . Since r is the remainder on division by elements of G , LT (r) is not
0 0
divisible by leading terms of elements of G , and thus LT (r) ∈<
/ LT (G ) >. Yet
0
LT (r) ∈< LT (G) >, which proves our claim. Since < LT (G ) >⊆< LT (G) >,
0
the ideals < LT (G ) > from successive iterations of the loop form an ascending

chain of ideals in K[x1 , . . . , xn ]. Then ACC implies that after finite number
0
of iterations the chain will stabilize, so that < LT (G ) >=< LT (G) > must
0
happen eventually. By Theorem 2.6.2, G = G, so that the algorithm terminates.

Example 2.7.1. Let I =< f1 , f2 >=< x3 − 2xy, x2 y − 2y 2 + x >. Note

that < f1 , f2 > is not Grobner basis for I, for LT (S(f1 , f2 )) = −x2 ∈<
/

LT (f1 ), LT (f2 ) >. S(f1 , f2 ) = −x2 ∈ I and its remainder on division by

F = (f1 , f2 ) is −x2 6= 0. Hence we add new generator f3 = −x2 . We set

F = (f1 , f2 , f3 ) and check if this new set is a Grobner basis. S(f1 , f2 ) =


F
f3 , so S(f1 , f2 ) = 0. S(f1 , f3 ) = (x3 − 2xy) − (−x)(−x2 ) = −2xy but
F
S(f1 , f3 ) = −2xy 6= 0 we must add f4 = −2xy to our generating set. If
F F
we let F = (f1 , f2 , f3 , f4 ). Then S(f1 , f2 ) = S(f1 , f3 ) = 0. S(f1 , f4 ) =

25
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

−1 2 F
y(x3 − 2xy) − ( )x (−2xy) = −2xy 2 = yf4 , so S(f1 , f4 ) = 0. S(f2 , f3 ) =
2
F
(x2 y −2y 2 +x)−(−y)(−x2 ) = −2y 2 +x but S(f2 , f3 ) = −2y 2 +x 6= 0. Thus we

must also add f5 = −2y 2 +x to our generating set. Setting F = (f1 , f2 , f3 , f4 , f5 )


F
one can compute s(fi , fj ) = 0, ∀ all 1 ≤ i < j ≤ 5. Thus Grobner basis of

I = {f1 , f2 , f3 , f4 , f5 } = {x3 − 2xy, x2 y − 2y 2 + x, −x2 , −2xy, −2y 2 + x}.

26
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

2.8 Exercises

1. Let f = 10x5 z 2 − 15x4 y 5 − 21x3 y 3 z 4 + 16xy 4 z 4 . Find the leading term,

leading coefficient and leading monomial of f with respect to the lex, grlex

and grevlex order. Reorder the terms in decreasing order.

2. Let f = y 2 x + 4xy − 3x2 and g = 2y + x + 1 in Q[x, y] with grlex order

with y > x. Show that f reduces to (1/4)x3 − (9/2)x2 − (7/4)x modulo g.

3. Let g = 2y + x + 1 and f = y 2 x + 4yx − 3x2 . Follow the DIVISION

ALGORITHM method to show that f = qg + r where q = (1/2)yx −

(1/4)x2 + 7/4x and r = (1/4)x3 − (9/2)x2 − (7/4)x.

4. Find the quotient and remainder when f = x3 y 3 + 2y 2 is divided by

f1 = 2xy 2 + 3x + 4y 2 and f2 = y 2 − 2y − 1.

5. Suppose g1 = xy 2 − xz + y, g2 = xy − z 2 , g3 = x − yz 4 . Let I = (g1 , g2 , g3 ).

Using the lex order show that < LT (I) >6=< LT (g1 ), LT (g2 ), LT (g3 ) >.

6. Find the Grobner basis for the ideal I = (y 5 +x2 +z 2 −4, y 2 +2x2 −5, yz−1)

with different monomial orderings.

7. Compute a Grobner basis for the ideal (x3 y 2 − 4x2 y 3 + 5y 5 , x6 − 7xy 5 ) ⊂

Q[x, y, z]. Is xy 9 ∈ I?

8. Determine whether the following sets G are Grobner basis for the ideal

they generate

• G = {x2 − y, x3 − z} - GLex order

• G = {y − x2 , z − x3 } - Lex order with x > y > z.

• G = {xy 2 − xz + y, xy − z 2 , x − yz 4 } - Lex order

9. Let R = Q[x, y, z] and f = 4x2 yz+3z 2 −5x3 +7xy 2 −7y 4 . Find the leading

monomial, leading term and leading exponent and leading coefficient for

f with respect to the different orderings.

10. Find the GCD and LCM of f1 and f2 where f1 = x3 −3x+2 and f2 = x2 −1.

27
Shiv Datt Kumar Basic Notions of Grobner Basis Theory

11. Let f1 = xy−x, f2 = x2 −y x < y with lex order. Show that G = f1 , f2 , f3

is a GB for I = (f1 , f2 ) where f3 = x3 − x.

12. Let f1 = x2 y + z, f2 = xz + y. Show that G = {f1 , f2 , xy 2 − z 2 , y 3 + z 3 }

is a GB for I = (f1 , f2 ) with grlex order with x > y > z.

13. Use Buchberger’s algorithm to compute Grobner basis of the ideal I =

hy − z 2 .z − x3 i ∈ Q[x, y, z] with grevlex and lex orders.

14. Solve the linear equations

3x + 4y − z + w = 0,

x − 3y + 3z − 4w = 0,

x−y+z−w =0

by computing a Grobner basis of the ideal generated by the polynomials

f1 = 3x + 4y − z + w, f2 = x − 3y + 3z − 4w, f3 = x − y + z − w.

15. Show that the polynomials f1 = x − y 2 w, f2 = y − zw, f3 = z − w3 ,

f4 = w3 − w ∈ Q[x, y, z, w] with lex ordering where x > y > z > w form

a Grobner basis for the ideal they generate. Show that they do not form

a lex Grobner basis if w > x > y > z.

16. Divide f = xy 2 + 1 by f1 = xy + 1 and f2 = y + 1 using lex order with

x > y.

28
Bibliography

[1] Ramji Lal, Algebra 1, Algebra 2, Springer Publication, 2017.

[2] William, Adams, Philippe Loustaunau, An Introduction to Grobner

Bases, AMS 1995

[3] Hoffman and Kunze, Linear Algebra.

[4] Cox, Little and O’Shea, Ideals and Algorithms.

[5] Derek J.S. Robinson: An Introduction to Abstract Algebra.

[6] Lidl and Niederreiter, Finite Fields and Applications.

[7] Henry Helson, Linear Algebra.

[8] .B. Bhattacharya, S.K. Jain, S.R. Nagpaul: Basic Abstract Algebra

29

You might also like