1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748
|
-*
Copyright 2014, Thomas Hawes.
You may redistribute this file under the terms of the GNU General Public
License as published by the Free Software Foundation, either version 2 of
the License, or any later version.
*-
-* Functions not ported:
o 'generateGroup' applied to a List of matrices is replaced by
'group' applied to FiniteGroupAction. The new function uses
schreierGraph which is more efficient than brute-force.
o a 'reynoldsOperator' was written separately for the new types.
o 'OrderBound' optional argument is not ported. It would need to
be reimplemented into our group generation function.
o 'IsGroup' optional argument is not ported because our methods
compute the group in advance.
*-
------------------------------------------------
-- molienSeries
-- Calculates the Molien (Hilbert) series of the invariant ring of a finite
-- group G.
-- Considers two cases: when the field the matrices are defined over is QQ
-- and when it is a general number field.
-- This is to get around the fact that frac for number fields is not yet
-- implemented in Macaulay2
-- The author thanks an anonymous referee for suggested improvements to this
-- method to ensure a consistent class of polynomial is returned
-- Reference:
-- [S08] Sturmfels, B. "Algorithms in Invariant Theory". 2nd Edition. Wien New
-- York: Springer-Verlag, 2008.
------------------------------------------------
-* Porting notes on molienSeries:
o Applies to FiniteGroupAction rather than RingOfInvariants
o Original code distinguished cases when coefficient ring K is QQ
from other fields because frac(K[U]) is not implemented when
K is a finite field extension of QQ created as quotient ring.
New code treats all fields as the most general one.
o Fred suggests caching the result of molienSeries in the
FiniteGroupAction because 1) it is computed independently of the
ring of invariants, and 2) it is called upon by the methods for
primary and secondary invariants.
o FUTURE TODO: extend to pos char in the non modular case
*-
molienSeries = method()
molienSeries FiniteGroupAction := { } >> opts -> (cacheValue (symbol molienSeries)) (G -> runHooks(FiniteGroupAction, symbol molienSeries, G) )
addHook(FiniteGroupAction, symbol molienSeries, G -> break (
--molienSeries FiniteGroupAction:= G -> (
K:=coefficientRing ring G;
if(isField K == false or char K =!= 0) then(
error "Action matrices must be defined over a field of characteristic zero"
);
R:= ring G;
if unique degrees R =!= {{1}} then
error "Only implemented for standard graded polynomial rings";
U:=symbol U;
-- frac of a general number field is not implemented yet in Macaulay2. So
-- need to calculate fractions and sum of fractions 'manually'.
n:= numgens R;
Ku:= K[U];
In:= id_(Ku^n);
L:=apply(group G, M->(det(In-U*sub(M,Ku))));
-- L is a list {p_1(U),...,p_m(U)}, say, (with m=#G), of reciprocals of
-- summands in the Molien series expansion.
numerat:=sum(#L,i->product(drop(L,{i,i})));
-- in above notation, (and using LaTeX syntax):
-- numerat=numerator of \frac{1}{p_1(U)}+...+\frac{1}{p_m(U)}.
denominat:=#(group G)*product(L);
A:=degreesRing(1);
-- A is the ring containing the numerator and denominator of the
-- Hilbert series
T:= first gens A; -- T is the variable of the Hilbert series
f:=map(A,Ku,{T}); -- Ring map K[U] -> A that sends U to T
-- Note that any non-integer elements of K would be sent to zero.
h:=new Divide from {f(numerat),factor(f(denominat))};
-- FG 6/17/2020, this code attempts a factorization
-- of the denominator of the Molien series that prioritizes
-- factors of the form (1-T^i)
-- if this is not desired, uncomment the next line
-- return reduceHilbert h;
d := first degree denominat;
h = reduceHilbert h;
den := value(h#1);
factors := {}; -- list of factors in the denominator
while d>0 do (
e := 0;
while (den % (1-T^d) == 0) do (
e = e+1;
den = den // (1-T^d);
);
if e != 0 then (
factors = {new Power from {1-T^d,e}} | factors;
);
d = d-1;
);
new Divide from {h#0,new Product from factors}
))
------------------------------------------------
-- primaryInvariants
-- Given a group and a polynomial ring over a number field, computes primary
-- invariants for the invariant ring
-- Uses the method optimalHSOP by default to do this, or dadeHSOP if optional
-- argument Dade is set to true
------------------------------------------------
-* Porting notes on primaryInvariants
o applies to FiniteGroupAction
*-
primaryInvariants=method(Options=>{Dade=>false, DegreeVector=>0})
primaryInvariants(FiniteGroupAction):=o->G->(
-- R - polynomial ring over a number field;
R := ring G;
-- V2.0 note: the optimal HSOP method only works with standard grading
-- It might be possible to make the Dade algorithm work with other
-- gradings but this will require further consideration.
-- For now, we throw an error if not in the standard graded case.
if unique degrees R =!= {{1}} then
error "Only implemented for standard graded polynomial rings";
n:=numgens R;
KK:=coefficientRing R;
if(isField KK == false or (o.Dade==false and char KK =!= 0)) then(
error "The polynomial ring must be defined over a field of
characteristic zero"
);
if o.Dade==true then(
if(
instance(KK,GaloisField) or (
instance(KK,QuotientRing) and (ambient KK===ZZ)
)==true
)
then(
if KK.order<=((#G)^(n-1)) then(
s:=read "Warning, your field may be too small for the Dade
algorithm to terminate successfully. Are you sure you want
to continue? (Enter `y' to continue, `n' to stop execution
of function.): ";
if s=!="y" then error "Use a bigger field."
)
);
if(
instance(KK,FractionField) and
(char KK=!=0) and
(dim(last(KK.baseRings))==0)
)==true then error(
"The ground field should be a field of characteristic zero
or a large finite field of type GaloisField or
QuotientRing."
);
return dadeHSOP G
);
if o.DegreeVector=!=0 then(
return optimalHSOP(G,DegreeVector=>(o.DegreeVector))
)
else return optimalHSOP G
);
------------------------------------------------
-- secondaryInvariants
-- Given a list of primary invariants for the invariant ring of a group, finds
-- secondary invariants
------------------------------------------------
-* Porting notes on secondaryInvariants
o applies to a list of primary invariants + FiniteGroupAction
*-
secondaryInvariants=method(Options=>{PrintDegreePolynomial=>false});
secondaryInvariants(List,FiniteGroupAction):=o->(P,G)->(
-- P - List of primary invariants, G - FiniteGroupAction
R := ring G;
if unique degrees R =!= {{1}} then
error "Only implemented for standard graded polynomial rings";
n:=numgens R;
KK:=coefficientRing R;
if(isField KK == false or char KK =!= 0) then(
error "Expected action over a field of characteristic zero"
);
for g in P do(if ring(g) =!= R then(
error "All polynomials must be in the ring on which the group acts"
)
);
molnum:=value numerator molienSeries G;
-- numerator of Molien series H(G,T)
moldenom:=value denominator molienSeries G;
-- denominator of Molien series H(G,T)
A:=ring moldenom;
-- the ring H(G,T) lives in. Note T is invertible in this ring.
t:=symbol t; B:=ZZ[t]; -- note t does not have an inverse in B
F:=map(B,A,{t});
-- F sends the variable T in A to t in B
-- Note that this will send negative powers of T to zero
-- This does not matter, because molnum and moldenom are concentrated in
-- non-negative degrees
pol:=(F(molnum)*(product apply(P,f->(1-t^((degree f)_0)))))//(F(moldenom));
-- pol = H(G,t)*(1-t^{d_1})*...*(1-t^{d_n}) where d_i are degrees of
-- primary invariants in P
-- pol is used to calculate the number of secondary invariants of
-- certain degrees.
degscoeffs:=apply(
apply(
flatten entries ((coefficients pol)#0),degree
),
D->{D,coefficient((sub(t,ZZ[t]))^(D_0),pol)}
);
-- outputs a list of lists of the form {{d},c} where d is a featuring
-- degree in pol and c is the corresponding coefficient. Thus there
-- are c secondary invariants of degree d.
I:=ideal P;
normforms:=new MutableHashTable;
secondary:=new MutableHashTable;
for dc in degscoeffs do(
d:=(dc#0)#0;
bas:=basis(d,R);
-- For this to work polynomials in P must have coefficients in a field
-- (and Macaulay 2 know it is a field)
mons:=flatten entries bas;
-- List of monomials of R of degree d
bas2:=sub(basis(d,R/I),R);
-- lifts a basis of degree d graded piece of R/I to elements of R
i:=1;
B:=matrix{{}};
for m in mons do(v:=(reynoldsOperator(m,G));
if((coefficientRing R)===QQ and v=!=sub(0,R)) then(
v=lcm(
apply(
flatten entries transpose((coefficients(v))#1),
l->(denominator(sub(l,QQ)))
)
)*v;
v=(1/gcd(
apply(
flatten entries transpose((coefficients v)#1),
l->sub(l,ZZ)
)
)
)*v
);
-- If coefficient ring of R is QQ and v nonzero, prune v to get
-- rid of fractions with denominators and leave coprime integer
-- coefficients.
nfv:=(v % I); -- find normal form of reynoldsOperator(m,G) mod I
if nfv==0 then continue else(
-- if zero, discard and go to next monomial
if i==1 then(
normforms#(d,i)=nfv;
secondary#(d,i)=v;
B=((coefficients(nfv,Monomials=>bas2))#1);
i=i+1
-- if nfv is first non zero normal form, then store
-- in normforms, store reynoldsOperator(m,G) in
-- secondary, add coefficients of nfv to B and
-- do i=i+1
) else(
A:=B|((coefficients(nfv,Monomials=>bas2))#1);
-- add coefficients of nfv to matrix of coefficients
-- of stored normal forms
if rank(A)==i then(
B=B|((coefficients(nfv,Monomials=>bas2))#1);
normforms#(d,i)=nfv;
secondary#(d,i)=v;
i=i+1
) else continue
-- a linear independence test. If nfv is linearly
-- dependent on previously stored normal forms, then
-- discard v. Else store v with the other secondary
-- invariants found.
);
);
if i==(1+(dc#1)) then break()
-- break loop once c secondary invariants of degree d
-- have been found
);
);
if o.PrintDegreePolynomial == true then print pol;
return sort values secondary
);
------------------------------------------------
-- invariantRing
-- Encorportates the primaryInvariants and secondaryInvariants methods into
-- one method.
------------------------------------------------
-* Porting notes on hironakaDecomposition
o this function is renamed from invariantRing to
hironakaDecomposition to provide consistent outputs for the
invariantRing method
o applies to a FiniteGroupAction, although would it be better
if it applied to the ring of invariants?
*-
hironakaDecomposition=method(
Options=>{
DegreeVector=>0,
PrintDegreePolynomial=>false
}
);
hironakaDecomposition(FiniteGroupAction):=o->G->(
L:=G;
-- if o.IsGroup==false then L=generateGroup(G,coefficientRing R);
P:=primaryInvariants(G,DegreeVector=>(o.DegreeVector));
S:=secondaryInvariants(P,G,PrintDegreePolynomial=>(
o.PrintDegreePolynomial
)
);
return((P,S))
);
------------------------------------------------
-- FUNCTIONS NOT EXPORTED --
-- (Can be viewed via listLocalSymbols command)
------------------------------------------------
-* Porting notes on unexported functions
o The functions groupAct, cart, vectors, isNonDecreasing,
degreeSequencesGivenProduct, nextBest, tryDegrees, optimalHSOP,
dadeHSOP below keep the original code. They have been minimally
modified to accept objects of type FiniteGroupAction as input.
*-
------------------------------------------------
-- groupAct
-- Outputs a list of polynomials, one element for each application of a member
-- of G to f.
-- Innermost apply computes a matrix: transpose of A*(variables in ring f).
-- Outermost apply evaluates f at these linear forms.
------------------------------------------------
groupAct=(f,G)->(
apply(
apply(
G, A->((vars ring f)*(transpose sub(A,ring f)))
),T->sub(f,T)
)
);
------------------------------------------------
-- cart
-- Takes two lists and outputs the Cartesian product of them as a list of lists
-- (the inner lists are the tuples).
-- E.g. ({1,2},{a,b})->{{1,a},{1,b},{2,a},{2,b}}
------------------------------------------------
cart=(A,B)->(
flatten(for u in A list(for v in B list(flatten({u,v}))))
);
------------------------------------------------
-- vectors
-- Inputs: n,k,i non-negative integers, with k<=i
-- Produces all n-tuples of natural numbers with total degrees from k to
-- i. Outputs a list of lists (vectors).
------------------------------------------------
vectors=(n,k,i)->(
x:=symbol x;
L:=ZZ[x_0..x_(n-1),MonomialOrder=>GLex];
out:=flatten flatten(
for j from k to i list(
apply(flatten entries basis(j,L),exponents)
)
);
return out
);
------------------------------------------------
-- nextBest
-- A 'next-best' function (see Kemper, G. An Algorithm to Compute Optimal
-- Homogeneous Systems of Parameters. Journal of Symbolic
-- Computation. 27. (1999) (hereon denoted by [kem99])).
-- Takes as input a list of positive integers {d_1,...,d_n}, with sum s and
-- product d, and outputs a list of positive integers {e_1,...,e_n}, with
-- e_1*...*e_n = d+1, or
-- e_1*...*e_n = d and e_1+...+e_n>s, or
-- e_1*...*e_n = d and e_1+...+e_n=s and {e_1,...,e_n}>{d_1,...,d_n}
-- (lexicographically).
-- This is the 'next-best function' suggested at the end of section 2 in
-- [kem99].
-- nextBest uses auxiliary functions isNonDecreasing and
-- degreeSequencesGivenProduct.
-- The author thanks an anonymous referee for suggesting the following
-- implementation.
------------------------------------------------
-- isNonDecreasing takes a list L and returns true if the list is (weakly)
-- increasing, false otherwise.
isNonDecreasing=L->(
for i to #L-2 do(
if L_i>L_(i+1) then return false;
);
return true;
);
-- degreeSequencesGivenProduct takes a pair of positive integers (n,d) and
-- outputs a list L of lists of positive integers {d_1,...,d_n} with
-- product({d_1,...,d_n})=d. The elements of L are ordered first according to
-- sum of entries (lowest to highest), then lexicographically.
degreeSequencesGivenProduct = (n,d)->(
if isPrime(d) then(
div:={1,d};
)
else(
div=for i from 1 to d list(
if d%i==0 then i else continue
);
); -- creates div - a list of divisors of d (including 1 and d)
L:=pack(1,div); -- L will ultimately be list of lists l of n elements
-- with product(l)=d
for i from 2 to n do(
L= flatten for l in L list(
q:=d//product(l);
div2:=select(div,x->q%x==0); -- div2 consists of divisors of
-- d that divide q=d//product(l)
select(apply(div2,x->prepend(x,l)),u->isNonDecreasing u)
-- To each element of div2, prepend to l and keep only if
-- resulting list is (weakly) increasing.
);
);
L=select(L,l->product(l)==d); -- discard lists l with
-- product(l) not equal to d
S:=partition(sum,L); -- S is a hashtable, with integer keys
-- s, and s=>all lists l with sum(l)=s
L=flatten for s in sort keys S list(
sort S#s
); -- sort the l in L first by increasing sum, then
-- lexicographically
return L
);
nextBest=vv->(
-- input is a list vv of non-negative integers
if instance(vv,List)==false then(
error "Input for nextBest must be a List of non-negative
integers"
);
for i in vv do(
if instance(i,ZZ)==false then(
error "Input for nextBest must be a List of
non-negative integers"
);
if i<=0 then(
error "Input for nextBest must be a List of
non-negative integers"
);
);
v:=sort vv;
n:=#v;
d:=product v;
degseqs:= degreeSequencesGivenProduct(n,d);
j:=position(degseqs,x->x==v); -- find position of v in the list of
-- degree sequences degseqs
if j==(#degseqs-1) then(
return first degreeSequencesGivenProduct(n,d+1);
-- If v is the last element of degseqs, then take the smallest degree
-- sequence with product d+1...
)
else(
return degseqs_(j+1); -- ... otherwise take the next degree
-- sequence in the list degseqs
)
);
------------------------------------------------
-- tryDegrees (Non-exported function)
-- The 'TryDegrees' function from [kem99].
-- Variable names mostly follow those in the TryDegrees algorithm in [kem99].
-- Is used to try and find an homogeneous system of parameters for the invariant
-- ring with the degrees input.
-- tryDegrees is the main computational process in the primaryInvariants method.
------------------------------------------------
tryDegrees=method();
tryDegrees(FiniteGroupAction,List):=(H,Li)->(
A := ring H;
nn := numgens A;
-- A - polynomial ring; H - FiniteGroupAction; nn - natural number, typically
-- dimension of A; Li - list of degrees
Bases:=new MutableHashTable;
-- stores spanning sets for the vector spaces (A^H)_l, where l is a degree
-- from a degree vector at various iterations of the computation method
computation:=(I,G,n,L)->(
-- 'computation' is the actual implementation of the TryDegrees
-- algorithm in [kem99] on the quotient ring A/I
k:=0; m:=#L; V:={sub(0,A)};
if m>0 then(
for l in L do(
Bases#(n,l)=toList set apply(
flatten entries basis(l,A),m->(reynoldsOperator(m,G))
)
); -- assigns to degree l a spanning set for (A^G)_l
V=(Bases#(n,(L#0))) -- V is the homogeneous piece of A^G
-- corresponding to the first degree in the degree vector L
);
if V=!={sub(0,A)} then(
-- only do the 'for f_1 \in V\0' part of loop in [kem99] if
-- V\0 is nonempty
V0:=delete(sub(0,A),V); breakfloop:=false; d:=1;
-- d is the degree of the 'coefficient vector'
while breakfloop==false do(
for v in apply(
vectors(#V0,d,d),u->(apply(u,l->(sub(l,A))))
) do( -- cycles through all vectors with degree d
-- 'coefficient vector'. This gives a way to
-- systematically try elements from the vector space
-- spanned by elements of V0.
-- Referring to p178, bullet point (a) of [Kem99], we
-- have chosen to include the natural numbers in the
-- ground field of A (which contains QQ) in the natural
-- way, since there seems to be no advantage to
-- choosing a different embedding.
f:=(((matrix{v})*transpose(matrix{V0}))_(0,0));
-- the f_1 in [kem99]
nextf:=false;
if k>0 then( -- beginning of k>0 condition in [kem99]
for M in(subsets toList(1..(k-1))) do(
union:=flatten apply(M,j->(Bases#(n,(L#j))));
if dim(A/(I+ideal({f}|union)))>(n-(#M)-1)
then(
break(nextf=true)
)
);
); -- end of k>0 condition in [kem99]
if nextf==true then continue;
R:=computation(I+ideal(f),G,n-1,drop(L,1));
-- 'R' is same notation as in [kem99]
if instance(R,List)==true then return({f}|R);
if R>=k then( -- beginning of the 'if R>=k then' loop
-- in [kem99]
k=R+1;
for M in(subsets toList(0..(k-1))) do(
union:=flatten apply(M,j->(Bases#(n,(L#j))));
-- corresponds to the union of homogeneous
-- pieces of A that appears in the algorithm
-- in [kem99]
if dim(A/(I+promote(ideal(union),A)))>(n-(#M)) then(
break(breakfloop=true)
)
);
-- end of the 'if R>=k then' loop in [kem99]
);
if breakfloop==true then break
); -- here is where the content of the 'for f_1 \in
-- V\0'
-- in [kem99] loop ends
d=d+1;
); -- here is where the 'for f_1 \in V\0' loop ACTUALLY ends
);
if k<=1 then(if dim(A/I)>n then k=0 else k=1);
-- the 'if k<=1' loop near end on [kem99] algorithm
if(m==0 and k==1) then k={};
return k
);
return computation(ideal(sub(0,A)),H,nn,Li)
);
------------------------------------------------
-- optimalHSOP (Non-exported function)
-- The 'HomogeneousParameterSystem' function from [kem99]
-- Uses the nextBest method to cycle through degree vectors ordered according to
-- rising values of the product of the degrees
-- Performs certain tests on a candidate degree vector to see if an hsop with
-- the degrees can't possibly exist
-- Then applies tryDegrees to each vector that passes these tests
------------------------------------------------
optimalHSOP=method(Options=>{DegreeVector=>0});
-- DegreeVector can be set to be a list of degrees if it is known an hsop with
-- these degrees exists. Bypasses the tests and applies tryDegrees to this list
-- straight away
optimalHSOP(FiniteGroupAction):= o->G->(
R:=ring G;
n:=numgens R;
if(o.DegreeVector=!=0 and instance(o.DegreeVector,List)==false) then(
error "DegreeVector must be zero or a List consisting of positive
integers"
);
if(o.DegreeVector=!=0 and (#(o.DegreeVector))>(numgens R)) then(
error "Length of DegreeVector must not exceed dimension of polynomial
ring"
);
L:=toList(n:1); -- L is list of degrees of primary invariants, initially
-- set to {1,...,1}. This will be changed by using nextBest and then
-- a set of primary invariants of degrees from L will be constructed
-- (subject to L passing various tests).
hsopfound:=false;
S:=0;
if o.DegreeVector=!=0 then(
S=tryDegrees(G,o.DegreeVector);
if instance(S,List)==true then hsopfound=true else(
error "There is no homogeneous system of parameters with degrees
given in DegreeVector"
)
);
molnum:=value numerator molienSeries G;
-- numerator of Molien series H(G,T)
moldenom:=value denominator molienSeries G;
-- denominator of Molien series H(G,T)
A:=ring moldenom;
-- the ring H(G,T) lives in. Note T is invertible in this ring.
t:=symbol t; B:=ZZ[t];
-- note t does not have an inverse in B
F:=map(B,A,{t});
-- F sends the variable T in A to t in B
-- Note that this will send negative powers of T to zero
-- This does not matter, because molnum and moldenom are concentrated in
-- non-negative degrees
a:=0;
while hsopfound==false do(
-- What follows are tests on the candidate degrees in L
if ((product L) % (#(group G)))=!=0 then(
L=(nextBest L);
continue;
);
-- order of group must divide product of the degrees
pol:=F(molnum)*(product apply(L,d->(1-t^d)));
if pol%(F(moldenom))=!=(sub(0,B)) then(
L=nextBest(L);
continue;
);
-- pol=F(molnum)*(1-t^{d_1})*...*(1-t^{d_n}) with d_i in L
-- pol should be divisible by the denominator of H(G,t)=F(H(G,T))
S=tryDegrees(G,L);
if instance(S,List)==true then hsopfound=true else(
L=(nextBest L);
)
-- tryDegrees returns either a list of invariants or a
-- positive integer
);
if (coefficientRing R)=!=QQ then return S;
SS:=for g in S list(
lcm(apply(
flatten entries transpose((coefficients(g))#1),l->(
denominator(sub(l,QQ))
)
)
)*g);
-- apply pruning to give integer coefficients if R is a polynomial
-- ring over QQ
return SS
);
------------------------------------------------
-- dadeHSOP (Non-exported function)
-- Finds an hsop for the invariant ring using the Dade algorithm.
-- (See Sturmfels, B., Algorithms in Invariant Theory. 2nd ed.
-- Wien New York: Springer-Verlag, 2008.)
-- Uses 'random' to produce candidate Dade basis vectors (DBVs) and tests
-- whether this is so using linear dependence tests, so is quite simple and
-- quick.
------------------------------------------------
dadeHSOP=method();
dadeHSOP(FiniteGroupAction):= G-> (
R := ring G;
n:=numgens R;
dadebasis:=new MutableHashTable;
-- integer keys, d=>d'th Dade basis vector
orbits:=new MutableHashTable;
-- integer keys, d=> orbit of d'th Dade basis vector
tuples:=new MutableHashTable;
-- integer keys, d=>orb(v_0) x ... x orb(v_{d-1}), d=0...n-1
d:=0;
dadebasis#d=((random(R^1,R^{-1}))_(0,0)); -- make a random linear form
-- V2.0 note: this way of making linear forms requires standard grading
orbits#d=toList(set groupAct(dadebasis#d,group G));
-- assigns to d the orbit of the d'th Dade basis vector
tuples#d=(cart((orbits#d),{{}})); -- the orbit of the 0th DBV in this case
while d<(n-1) do(
basisvectorfound:=false;
while basisvectorfound==false do(
u:=((random(R^1,R^{-1}))_(0,0));
-- a candidate for the d+1'th DBV
l:=0;
while l<(#(tuples#d)) do(
-- will cycle through the possible tuples of orbit elements
k:=1; B:=matrix{{}};
for ww in((tuples#d)#l) do(
if k==1 then(
B=(coefficients(ww,Monomials=>(vars R)))#1; k=k+1
) else(
B=B|((coefficients(ww,Monomials=>(vars R)))#1);
k=k+1
);
); -- creates matrix of coefficients B
augB:=B|((coefficients(u,Monomials=>(vars R)))#1);
-- creates the augmented matrix of B
if (rank augB)==(rank B) then break() else l=l+1
-- rank test to see if u is linearly dependent on the linear
-- forms from the tuple being considered
);
if l==(#(tuples#d)) then(
basisvectorfound=true;
d=d+1;
dadebasis#d=u
);
);
-- if we successfully made it through all tuples, break loop,
-- increment d and take u to be our new DBV
orbits#d=groupAct(dadebasis#d,group G);
-- create the orbit of the new DBV
if #(keys dadebasis)<n then tuples#d=cart((orbits#d),tuples#(d-1));
-- create new set of tuples only if we still need DBVs
);
hsop:=sort apply(keys dadebasis,k->product orbits#k);
-- for each DBV, creates the product of all linear forms in the orbit
if (coefficientRing R)=!=QQ then return hsop;
SS:=apply(
hsop,g->(
lcm(
apply(
flatten entries transpose((coefficients(g))#1),
l->(denominator(sub(l,QQ))))
)*g
)
);
return apply(
SS,
g->(
(1/gcd(apply(flatten entries transpose((coefficients(g))#1),
l->sub(l,ZZ)))
)*g
)
)
-- apply pruning to coefficients if ground field of R is QQ
);
|