% Examples of use of Groebner code.
% In the Examples 1 - 3 the polynomial ring for the ideal operations
% (variable sequence, term order mode) is defined globally in advance.
% Example 1, Linz 85.
torder ({q1,q2,q3,q4,q5,q6},lex)$
groebner {q1,
q2**2 + q3**2 + q4**2,
q4*q3*q2,
q3**2*q2**2 + q4**2*q2**2 + q4**2*q3**2,
q6**2 + 1/3*q5**2,
q6**3 - q5**2*q6,
2*q2**2*q6 - q3**2*q6 - q4**2*q6 + q3**2*q5 - q4**2*q5,
2*q2**2*q6**2 - q3**2*q6**2 - q4**2*q6**2 - 2*q3**2*q5*q6
+ 2*q4**2*q5*q6 - 2/3*q2**2*q5**2 + 1/3*q3**2*q5**2
+ 1/3*q4**2*q5**2,
- q3**2*q2**2*q6 - q4**2*q2**2*q6 + 2*q4**2*q3**2*q6 -
q3**2*q2**2*q5 + q4**2*q2**2*q5,
- q3**2*q2**2*q6**2 - q4**2*q2**2*q6**2 + 2*q4**2*q3**2*q6**2
+ 2*q3**2*q2**2*q5*q6 - 2*q4**2*q2**2*q5*q6 + 1/3*q3**2*q2**2
*q5**2 + 1/3*q4**2*q2**2*q5**2 - 2/3*q4**2*q3**2*q5**2,
- 3*q3**2*q2**4*q5*q6**2 + 3*q4**2*q2**4*q5*q6**2
+ 3*q3**4*q2**2*q5*q6**2 - 3*q4**4*q2**2*q5*q6**2
- 3*q4**2*q3**4*q5*q6**2 + 3*q4**4*q3**2*q5*q6**2
+ 1/3*q3**2*q2**4*q5**3 - 1/3*q4**2*q2**4*q5**3
- 1/3*q3**4*q2**2*q5**3 + 1/3*q4**4*q2**2*q5**3 + 1/3*q4**2
*q3**4*q5**3 - 1/3*q4**4*q3**2*q5**3};
% Example 2. (Little) Trinks problem with 7 polynomials in 6 variables.
trinkspolys := {45*p + 35*s - 165*b - 36,
35*p + 40*z + 25*t - 27*s,
15*w + 25*p*s + 30*z - 18*t - 165*b**2,
- 9*w + 15*p*t + 20*z*s,
w*p + 2*z*t - 11*b**3,
99*w - 11*s*b + 3*b**2,
b**2 + 33/50*b + 2673/10000}$
trinksvars := {w,p,z,t,s,b}$
torder(trinksvars,lex)$
groebner trinkspolys;
groesolve ws;
% Example 3. Hairer, Runge-Kutta 1, 6 polynomials 8 variables.
torder({c2,c3,b3,b2,b1,a21,a32,a31},lex);
groebnerf {c2 - a21,
c3 - a31 - a32,
b1 + b2 + b3 - 1,
b2*c2 + b3*c3 - 1/2,
b2*c2**2 + b3*c3**2 - 1/3,
b3*a32*c2 - 1/6};
% The examples 4 and 5 use automatic variable extraction.
% Example 4.
torder gradlex$
g4 :=
groebner({b + e + f - 1,
c + d + 2*e - 3,
b + d + 2*f - 1,
a - b - c - d - e - f,
d*e*a**2 - 1569/31250*b*c**3,
c*f - 587/15625*b*d});
hilbertpolynomial g4;
glexconvert(g4,gvarslast,newvars={e},maxdeg=8);
% Example 5.
torder({u0,u2,u3,u1},lex)$
groesolve({u0**2 - u0 + 2*u1**2 + 2*u2**2 + 2*u3**2,
2*u0*u1 + 2*u1*u2 + 2*u2*u3 - u1,
2*u0*u2 + u1**2 + 2*u1*u3 - u2,
u0 + 2*u1 + 2*u2 + 2*u3 - 1},
{u0,u2,u3,u1});
% Example 6. (Big) Trinks problem with 6 polynomials in 6 variables.
torder(trinksvars,lex)$
btbas :=
groebner {45*p + 35*s - 165*b - 36,
35*p + 40*z + 25*t - 27*s,
15*w + 25*p*s + 30*z - 18*t - 165*b**2,
-9*w + 15*p*t + 20*z*s,
w*p + 2*z*t - 11*b**3,
99*w - 11*b*s + 3*b**2};
% The above system has dimension zero. Therefore its Hilbert polynomial
% is a constant which is the number of zero points (including complex
% zeros and multipliticities);
hilbertpolynomial ws;
% Example of Groebner with numerical postprocessing.
on rounded;
groesolve(trinkspolys,trinksvars);
off rounded;
% Additional groebner operators.
% Reduce one polynomial wrt the basis of big Trinks. The result 0
% is a proof for the ideal membership of the polynomial.
torder(trinksvars,lex)$
preduce(45*p + 35*s - 165*b - 36,btbas);
% The following examples show how to work with the distributive
% form of polynomials.
torder({u0,u1,u2,u3},gradlex)$
gsplit(2*u0*u2 + u1**2 + 2*u1*u3 - u2,{u0,u1,u2,u3});
torder(trinksvars,lex)$
gsort trinkspolys;
gspoly(first trinkspolys, second trinkspolys);
gvars trinkspolys;
% Tagged basis and reduction trace. A tagged basis is a basis where
% each polynomial is equated to a linear combination of the input
% set. A tagged reduction shows how the result is computed by using
% the basis polynomials.
% First example for tagged polynomials: show how a polynomial is
% represented as linear combination of the basis polynomials.
% First I set up an environment for the computation.
torder(trinksvars,lex)$
% Then I compute an ordinary Groebner basis.
bas := groebner trinkspolys$
% Next I assign a tag to each basis polynomial.
taggedbas := for i:= 1:length bas collect
mkid(p,i) = part(bas,i);
% And finally I reduce a (tagged) polynomial wrt the tagged basis.
preducet(new=w*p + 2*z*t - 11*b**3,taggedbas);
% Second example for tagged polynomials: representing a Groebner basis
% as a combination of the input polynomials, here in a simple geometric
% problem.
torder({x,y},lex)$
groebnert {circle=x**2 + y**2 - r**2,line = a*x + b*y};
% In the third example I enter two polynomials that have no common zero.
% Consequently the basis is {1}. The tagged computation gives me a proof
% for the inconsistency of the system which is independent of the
% Groebner formalism.
groebnert {circle1=x**2 + y**2 - 10,circle2=x**2 + y**2 - 2};
% Solve a special elimination task by using a blockwise elimination
% order defined by a matrix. The equation set goes back to A.M.H.
% Levelt (Nijmegen). The question is whether there is a member in the
% ideal which depends only on two variables. Here we select x4 and y1.
% The existence of such a polynomial proves that the system has exactly
% one degree of freedom.
% The first two rows of the term order matrix define the groupwise
% elimination. The remaining lines define a secondary local
% lexicographical behavior which is needed to construct an admissible
% ordering.
f1 := y1^2 + z1^2 -1;
f2 := x2^2 + y2^2 + z2^2 -1;
f3 := x3^2 + y3^2 + z3^2 -1;
f4 := x4^2 + z4^2 -1;
f5 := y1*y2 + z1*z2;
f6 := x2*x3 + y2*y3 + z2*z3;
f7 := x3*x4 + z3*z4;
f8 := x2 + x3 + x4 + 1;
f9 := y1 + y2 + y3 - 1;
f10:= z1 + z2 + z3 + z4;
eqns := {f1,f2,f3,f4,f5,f6,f7,f8,f9,f10}$
vars := {x2,x3,y2,y3,z1,z2,z3,z4,x4,y1}$
torder(vars,matrix,
mat(
(1,1,1,1,1,1,1,1,0,0),
(0,0,0,0,0,0,0,0,1,1),
(1,0,0,0,0,0,0,0,0,0),
(0,1,0,0,0,0,0,0,0,0),
(0,0,1,0,0,0,0,0,0,0),
(0,0,0,1,0,0,0,0,0,0),
(0,0,0,0,1,0,0,0,0,0),
(0,0,0,0,0,1,0,0,0,0),
(0,0,0,0,0,0,1,0,0,0),
(0,0,0,0,0,0,0,0,1,0)));
first reverse groebner(eqns,vars);
% For a faster execution we convert the matrix into a
% proper machine code routine. This step can be taken only
% if there is access to a compiler.
on comp;
torder_compile(levelt,mat(
(1,1,1,1,1,1,1,1,0,0),
(0,0,0,0,0,0,0,0,1,1),
(1,0,0,0,0,0,0,0,0,0),
(0,1,0,0,0,0,0,0,0,0),
(0,0,1,0,0,0,0,0,0,0),
(0,0,0,1,0,0,0,0,0,0),
(0,0,0,0,1,0,0,0,0,0),
(0,0,0,0,0,1,0,0,0,0),
(0,0,0,0,0,0,1,0,0,0),
(0,0,0,0,0,0,0,0,1,0)));
torder(vars,levelt)$
first reverse groebner(eqns,vars);
% For a homogeneous polynomial set we compute a graded Groebner
% basis with grade limits. We use the graded term order with lex
% as following order. As the grade vector has no zeros, this ordering
% is functionally equivalent to a weighted ordering.
torder({x,y,z},graded,{1,1,2},lex);
dd_groebner(0,10,{x^10*y + y*z^5, x*y^12 + y*z^6});
dd_groebner(0,50,{x^10*y + y*z^5, x*y^12 + y*z^6});
dd_groebner(0,infinity,{x^10*y + y*z^5, x*y^12 + y*z^6});
% Test groebner_walk
trinkspolys := {45*p + 35*s - 165*b - 36,
35*p + 40*z + 25*t - 27*s,
15*w + 25*p*s + 30*z - 18*t - 165*b**2,
- 9*w + 15*p*t + 20*z*s,
w*p + 2*z*t - 11*b**3,
99*w - 11*s*b + 3*b**2,
b**2 + 33/50*b + 2673/10000}$
trinksvars := {w,p,z,t,s,b}$
torder(trinksvars,gradlex)$
gg := groebner trinkspolys$
g := groebner_walk gg$
on div$
g;
clear g, gg, trinkspolys, trinksvars$
end;