7 Higher level methods for pcp-groups

returns the polycyclic series of U defined by an igs of U.

returns a normal series of U with elementary or free abelian factors.

returns an efa series of U such that every factor in the series is semisimple as a module for U over a finite field or over the rationals.

the derived series of U.

the derived series of U refined to an efa series such that in each abelian factor of the derived series the free abelian factor is at the top.

the derived series of U refined to an efa series such that in each abelian factor of the derived series the free abelian factor is at the bottom.

the lower central series of U. If U does not have a largest nilpotent quotient group, then this function may not terminate.

the upper central series of U. This function always terminates, but it may terminate at a proper subgroup of U.

returns an efa series of U such that all torsion-free factors are at the top and all finite factors are at the bottom. Such a series might not exist for U and in this case the function returns fail.

gap> G := ExamplesOfSomePcpGroups(5);
Pcp-group with orders [ 2, 0, 0, 0 ]
gap> Igs(G);
[ g1, g2, g3, g4 ]

gap> PcpSeries(G);
[ Pcp-group with orders [ 2, 0, 0, 0 ],
  Pcp-group with orders [ 0, 0, 0 ],
  Pcp-group with orders [ 0, 0 ],
  Pcp-group with orders [ 0 ],
  Pcp-group with orders [  ] ]

gap> List( PcpSeries(G), Igs );
[ [ g1, g2, g3, g4 ], [ g2, g3, g4 ], [ g3, g4 ], [ g4 ], [  ] ]

Algorithms for pcp-groups often use an efa series of G and work down over the factors of this series. Usually, pcp's of the factors are more useful than the actual factors. Hence we provide the following.

returns a list of pcp's corresponding to the factors of the series. If the second argument is present, then each pcp corresponds to a decomposition of the abelian groups into direct factors.

returns a list of pcps corresponding to an efa series of U.

gap> G := ExamplesOfSomePcpGroups(5);
Pcp-group with orders [ 2, 0, 0, 0 ]

gap> PcpsBySeries( DerivedSeries(G));
[ Pcp [ g1, g2, g3, g4 ] with orders [ 2, 2, 2, 2 ],
  Pcp [ g2^-2, g3^-2, g4^2 ] with orders [ 0, 0, 4 ],
  Pcp [ g4^8 ] with orders [ 0 ] ]
gap> PcpsBySeries( RefinedDerivedSeries(G));
[ Pcp [ g1, g2, g3 ] with orders [ 2, 2, 2 ],
  Pcp [ g4 ] with orders [ 2 ],
  Pcp [ g2^2, g3^2 ] with orders [ 0, 0 ],
  Pcp [ g4^2 ] with orders [ 2 ],
  Pcp [ g4^4 ] with orders [ 2 ],
  Pcp [ g4^8 ] with orders [ 0 ] ]

gap> PcpsBySeries( DerivedSeries(G), "snf" );
[ Pcp [ g2, g3, g1 ] with orders [ 2, 2, 4 ], 
  Pcp [ g4^2, g3^-2, g2^2*g4^2 ] with orders [ 4, 0, 0 ], 
  Pcp [ g4^8 ] with orders [ 0 ] ]
gap> G.1^4 in DerivedSubgroup( G );
true
gap> G.1^2 = G.4;                          
true

gap>  PcpsOfEfaSeries( G );
[ Pcp [ g1 ] with orders [ 2 ],
  Pcp [ g2 ] with orders [ 0 ],
  Pcp [ g3 ] with orders [ 0 ],
  Pcp [ g4 ] with orders [ 0 ] ]

7.2 Orbit stabilizer methods for pcp-groups

Let U be a pcp-group which acts on a set Omega. One of the fundamental problems in algorithmic group theory is the determination of orbits and stabilizers of points in Omega under the action of U. We distinguish two cases: the case that all considered orbits are finite and the case that there are infinite orbits. In the latter case, an orbit cannot be listed and a description of the orbit and its corresponding stabilizer is much harder to obtain.

If the considered orbits are finite, then the following two functions can be applied to compute the considered orbits and their corresponding stabilizers.

The input gens can be an igs or a pcp of a pcp-group U. The elements in the list gens act as the elements in the list acts via the function oper on the given points; that is, oper( point, acts[i] ) applies the ith generator to a given point. Thus the group defined by acts must be a homomorphic image of the group defined by gens. The first function returns a record containing the orbit as component 'orbit' and and igs for the stabilizer as component 'stab'. The second function returns a list of records, each record contains 'repr' and 'stab'. Both of these functions run forever on infinite orbits.

gap> G := DihedralPcpGroup( 0 );
Pcp-group with orders [ 2, 0 ]
gap> mats := [ [[-1,0],[0,1]], [[1,1],[0,1]] ];;
gap> pcp := Pcp(G);
Pcp [ g1, g2 ] with orders [ 2, 0 ]
gap> PcpOrbitStabilizer( [0,1], pcp, mats, OnRight );
rec( orbit := [ [ 0, 1 ] ],
     stab := [ g1, g2 ],
     word := [ [ [ 1, 1 ] ], [ [ 2, 1 ] ] ] )

If the considered orbits are infinite, then it may not always be possible to determine a description of the orbits and their stabilizers. However, as shown in EOs01 and Eic02, it is possible to determine stabilizers and check if two elements are contained in the same orbit if the given action of the polycyclic group is a unimodular linear action on a vector space. The following functions are available for this case.

The first function computes the stabilizer in U of the vector v where the pcp group U acts via mats on an integral space and v and w are elements in this integral space. The second function checks whether v and w are in the same orbit and the function returns either false or a record containing an element in U mapping v to w and the stabilizer of v.

The first function computes the normalizer in U of the lattice with the basis B, where the pcp group U acts via mats on an integral space and B is a subspace of this integral space. The second functions checks whether the two lattices with the bases B and C are contained in the same orbit under U. The function returns either false or a record with an element in U mapping B to C and the stabilizer of B.

# get a pcp group and a free abelian normal subgroup
gap> G := ExamplesOfSomePcpGroups(8);
Pcp-group with orders [ 0, 0, 0, 0, 0 ]
gap> efa := EfaSeries(G);
[ Pcp-group with orders [ 0, 0, 0, 0, 0 ],
  Pcp-group with orders [ 0, 0, 0, 0 ], 
  Pcp-group with orders [ 0, 0, 0 ],
  Pcp-group with orders [  ] ]
gap> N := efa[3];
Pcp-group with orders [ 0, 0, 0 ]
gap> IsFreeAbelian(N);
true

# create conjugation action on N
gap> mats := LinearActionOnPcp(Igs(G), Pcp(N));
[ [ [ 1, 0, 0 ], [ 0, 1, 0 ], [ 0, 0, 1 ] ],
  [ [ 0, 0, 1 ], [ 1, -1, 1 ], [ 0, 1, 0 ] ],
  [ [ 1, 0, 0 ], [ 0, 1, 0 ], [ 0, 0, 1 ] ],
  [ [ 1, 0, 0 ], [ 0, 1, 0 ], [ 0, 0, 1 ] ],
  [ [ 1, 0, 0 ], [ 0, 1, 0 ], [ 0, 0, 1 ] ] ]

# take an arbitrary vector and compute its stabilizer
gap> StabilizerIntegralAction(G,mats, [2,3,4]);
Pcp-group with orders [ 0, 0, 0, 0 ]
gap> Igs(last);
[ g1, g3, g4, g5 ]

# check orbits with some other vectors
gap> OrbitIntegralAction(G,mats, [2,3,4],[3,1,5]);
rec( stab := Pcp-group with orders [ 0, 0, 0, 0 ], prei := g2 )

gap> OrbitIntegralAction(G,mats, [2,3,4], [4,6,8]);
false

# compute the orbit of a subgroup of Z^3 under the action of G
gap> NormalizerIntegralAction(G, mats, [[1,0,0],[0,1,0]]);
Pcp-group with orders [ 0, 0, 0, 0, 0 ]
gap> Igs(last);
[ g1, g2^2, g3, g4, g5 ]

7.3 Centralizers, Normalizers and Intersections

In this section we list a number of operations for which there are methods installed to compute the corresponding features in polycyclic groups.

These functions solve the conjugacy problem for elements in pcp-groups and they can be used to compute centralizers. The first method returns a subgroup of the given group U, the second method either returns a conjugating element or false if no such element exists.

The methods are based on the orbit stabilizer algorithms described in EOs01. For nilpotent groups, an algorithm to solve the conjugacy problem for elements is described in Sims94.

These three functions solve the conjugacy problem for subgroups and compute centralizers and normalizers of subgroups. The first two functions return subgroups of the input group U, the third function returns a conjugating element or false if no such element exists.

The methods are based on the orbit stabilizer algorithms described in Eic02. For nilpotent groups, an algorithm to solve the conjugacy problems for subgroups is described in Lo98.

A general method to compute intersections of subgroups of a pcp-group is described in Eic01b, but it is not yet implemented here. However, intersections of subgroups U, N leqG can be computed if N is normalising U. See Sims94 for an outline of the algorithm.

7.4 Finite subgroups

There are various finite subgroups of interest in polycyclic groups. See Eic00 for a description of the algorithms underlying the functions in this section.

If the set of elements of finite order forms a subgroup, then we call it the torsion subgroup. This function determines the torsion subgroup of U, if it exists, and returns fail otherwise. Note that a torsion subgroup does always exist if U is nilpotent.

Each polycyclic groups has a unique largest finite normal subgroup. This function computes it for U.

This function checks if U is torsion free. It returns true or false.

There exist only finitely many conjugacy classes of finite subgroups in a polycyclic group U and this function can be used to compute them. The algorithm underlying this function proceeds by working down a normal series of U with elementary or free abelian factors. The following function can be used to give the algorithm a specific series.

gap> G := ExamplesOfSomePcpGroups(15);
Pcp-group with orders [ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 5, 4, 0 ]
gap> TorsionSubgroup(G);
Pcp-group with orders [ 5, 2 ]
gap> NormalTorsionSubgroup(G);
Pcp-group with orders [ 5, 2 ]
gap> IsTorsionFree(G);
false
gap> FiniteSubgroupClasses(G);
[ Pcp-group with orders [ 5, 2 ]^G,
  Pcp-group with orders [ 2 ]^G,
  Pcp-group with orders [ 5 ]^G,
  Pcp-group with orders [  ]^G ]

gap> G := DihedralPcpGroup( 0 );
Pcp-group with orders [ 2, 0 ]
gap> TorsionSubgroup(G);
fail
gap> NormalTorsionSubgroup(G);
Pcp-group with orders [  ]
gap> IsTorsionFree(G);
false
gap> FiniteSubgroupClasses(G);
[ Pcp-group with orders [ 2 ]^G,
  Pcp-group with orders [ 2 ]^G,
  Pcp-group with orders [  ]^G ]

7.5 Subgroups of finite index and maximal subgroups

Here we outline functions to determine various types of subgroups of finite index in polycyclic groups. Again, see Eic00 for a description of the algorithms underlying the functions in this section. Also, we refer to Lo99 for an alternative appraoch.

Each maximal subgroup of a polycyclic group U has p-power index for some prime p. This function can be used to determine the conjugacy classes of all maximal subgroups of p-power index for a given prime p.

There are only finitely many subgroups of a given index in a polycyclic group U. This function computes conjugacy classes of all subgroups of index n in U.

This function computes the normal subgroups of index n in U.

This function returns a normal subgroup N of finite index in U such that N is nilpotent-by-abelian. Such a subgroup exists in every polycyclic group and this function computes such a subgroup using LowIndexNormal. However, we note that this function is not very efficient and the function NilpotentByAbelianByFiniteSeries may well be more efficient on this task.

gap> G := ExamplesOfSomePcpGroups(2);
Pcp-group with orders [ 0, 0, 0, 0, 0, 0 ]

gap> MaximalSubgroupClassesByIndex( G, 61 );;
gap> max := List( last, Representative );;
gap> List( max, x -> Index( G, x ) );
[ 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61,
  61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61,
  61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61,
  61, 61, 61, 61, 61, 61, 226981 ]

gap> LowIndexSubgroupClasses( G, 61 );;
gap> low := List( last, Representative );;
gap> List( low, x -> Index( G, x ) );
[ 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61,
  61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61,
  61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61, 61,
  61, 61, 61, 61, 61, 61 ]

7.6 Further attributes for pcp-groups based on the Fitting subgroup

In this section we provide a variety of other attributes for pcp-groups. Most of the methods below are based or related to the Fitting subgroup of the given group. We refer to Eic01 for a description of the underlying methods.

returns the Fitting subgroup of U; that is, the largest nilpotent normal subgroup of U.

checks whether the Fitting subgroup of U has finite index.

returns the centre of U.

returns the FC-centre of U; that is, the subgroup containing all elements having a finite conjugacy class in U.

returns a normal subgroup N of U such that N has a polycyclic series with infinite factors only.

returns a normal series 1 leqF leqA leqU such that F is nilpotent, A/F is abelian and U/A is finite. This series is computed using the Fitting subgroup and the centre of the Fitting factor.

7.7 Functions for nilpotent groups

There are (very few) functions which are available for nilpotent groups only. First, there are the different central series. These are available for all groups, but for nilpotent groups they terminate and provide series though the full group. Secondly, the determination of a minimal generating set is available for nilpotent groups only.

gap> G := ExamplesOfSomePcpGroups(14);
Pcp-group with orders [ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 5, 4, 0, 5, 5, 4, 0, 6,
  5, 5, 4, 0, 10, 6 ]
gap> IsNilpotent(G);
true

gap> PcpsBySeries( LowerCentralSeries(G));
[ Pcp [ g1, g2 ] with orders [ 0, 0 ],
  Pcp [ g3 ] with orders [ 0 ],
  Pcp [ g4 ] with orders [ 0 ],
  Pcp [ g5 ] with orders [ 0 ],
  Pcp [ g6, g7 ] with orders [ 0, 0 ],
  Pcp [ g8 ] with orders [ 0 ],
  Pcp [ g9, g10 ] with orders [ 0, 0 ],
  Pcp [ g11, g12, g13 ] with orders [ 5, 4, 0 ],
  Pcp [ g14, g15, g16, g17, g18 ] with orders [ 5, 5, 4, 0, 6 ],
  Pcp [ g19, g20, g21, g22, g23, g24 ] with orders [ 5, 5, 4, 0, 10, 6 ] ]

gap> PcpsBySeries( UpperCentralSeries(G));
[ Pcp [ g1, g2 ] with orders [ 0, 0 ],
  Pcp [ g3 ] with orders [ 0 ],
  Pcp [ g4 ] with orders [ 0 ],
  Pcp [ g5 ] with orders [ 0 ],
  Pcp [ g6, g7 ] with orders [ 0, 0 ],
  Pcp [ g8 ] with orders [ 0 ],
  Pcp [ g9, g10 ] with orders [ 0, 0 ],
  Pcp [ g11, g12, g13 ] with orders [ 5, 4, 0 ],
  Pcp [ g14, g15, g16, g17, g18 ] with orders [ 5, 5, 4, 0, 6 ],
  Pcp [ g19, g20, g21, g22, g23, g24 ] with orders [ 5, 5, 4, 0, 10, 6 ] ]

gap> MinimalGeneratingSet(G);
[ g1, g2 ]

7.8 Random methods for pcp-groups

Below we introduce a function which computes orbit and stabilizer using a random method. This function tries to approximate the orbit and the stabilizer, but the returned orbit or stabilizer may be incomplete. This function is used in the random methods to compute normalizers and centralizers. Note that determinstic methods for these purposes are also available.

gap> G := DihedralPcpGroup(0);
Pcp-group with orders [ 2, 0 ]
gap> mats := [[[-1, 0],[0,1]], [[1,1],[0,1]]];
[ [ [ -1, 0 ], [ 0, 1 ] ], [ [ 1, 1 ], [ 0, 1 ] ] ]
gap> pcp := Pcp(G);
Pcp [ g1, g2 ] with orders [ 2, 0 ]

gap> RandomPcpOrbitStabilizer( [1,0], pcp, mats, OnRight ).stab;
#I  Orbit longer than limit: exiting.
[  ]

gap> g := Igs(G)[1];
g1
gap> RandomCentralizerPcpGroup( G, g );
#I  Stabilizer not increasing: exiting.
Pcp-group with orders [ 2 ]
gap> Igs(last);
[ g1 ]

7.9 Non-abelian tensor product and Schur extensions

Let G be a polycyclic group with a polycyclic generating sequence consisting of n elements. This function computes the largest central extension H of G such that H is generated by n elements. If F/R is the underlying polycyclic presentation for G, then H is isomorphic to F/[R,F].

gap> G := DihedralPcpGroup( 0 );
Pcp-group with orders [ 2, 0 ]
gap> Centre( G );
Pcp-group with orders [  ]
gap> H := SchurExtension( G );
Pcp-group with orders [ 2, 0, 0, 0 ]
gap> Centre( H );
Pcp-group with orders [ 0, 0 ]
gap> H/Centre(H);
Pcp-group with orders [ 2, 0 ]
gap> Subgroup( H, [H.1,H.2] ) = H;
true

returns the projection from the Schur extension G^* of G onto G. See the function SchurExtension. The kernel of this epimorphism is the direct product of the Schur multiplicator of G and a direct product of n copies of Z where n is the number of generators in the polycyclic presentation for G. The Schur multiplicator is the intersection of the kernel and the derived group of the source. See also the function SchurCovering.

gap> gl23 := Range( IsomorphismPcpGroup( GL(2,3) ) );
Pcp-group with orders [ 2, 3, 2, 2, 2 ]
gap> SchurExtensionEpimorphism( gl23 );
[ g1, g2, g3, g4, g5, g6, g7, g8, g9, g10 ] -> [ g1, g2, g3, g4, g5,
id, id, id, id, id ]
gap> Kernel( last );
Pcp-group with orders [ 0, 0, 0, 0, 0 ]
gap> SchurMultiplicator( gl23 );
[  ]
gap> Intersection( Kernel(epi), DerivedSubgroup( Source(epi) ) );
[  ]

There is a crossed pairing from G into (G^*)' which can be defined via this epimorphism:

gap> G := DihedralPcpGroup(0);
Pcp-group with orders [ 2, 0 ]
gap> epi := SchurExtensionEpimorphism( G );
[ g1, g2, g3, g4 ] -> [ g1, g2, id, id ]
gap> PreImagesRepresentative( epi, G.1 );
g1
gap> PreImagesRepresentative( epi, G.2 );
g2
gap> Comm( last, last2 );
g2^-2*g4

computes a Schur covering group of the polycyclic group G. A Schur covering is a largest central extension H of G such that the kernel M of the projection of H onto G is contained in the commutator subgroup of H.

If G is given by a presentation F/R, then M is isomorphic to the subgroup R cap[F,F] / [R,F]. Let C be a complement to R cap[F,F] / [R,F] in R/[R,F]. Then F/C is isomorphic to H and R/C is isomorphic to M.

gap> G := AbelianPcpGroup( 3,[] );
Pcp-group with orders [ 0, 0, 0 ]
gap> ext := SchurCovering( G );
Pcp-group with orders [ 0, 0, 0, 0, 0, 0 ]
gap> Centre( ext );
Pcp-group with orders [ 0, 0, 0 ]
gap> IsSubgroup( DerivedSubgroup( ext ), last );
true

computes the isomorphism type of the Schur multiplicator of G and returns a list of pairs describing the isomorphism type as a direct product of cyclic groups. The first component of a pair is the order of the cyclic group. The second component specifies how often that cyclic group occurs as a direct factor. The isomorphism type of the Schur multiplicator is the direct product of the groups specified by the list of pairs.

If a cyclic factor is infinite, then the first component of the corresponding pair is 0.

Note that the Schur multiplicator of a polycyclic group is a fintiely generated abelian group.

gap> G := DihedralPcpGroup( 0 );
Pcp-group with orders [ 2, 0 ]
gap> DirectProduct( G, AbelianPcpGroup( 2, [] ) );
Pcp-group with orders [ 0, 0, 2, 0 ]
gap> SchurMultiplicator( last );
[ [ 2, 4 ], [ 0, 1 ] ]

returns the epimorphism of the non-abelian exterior square of a polycylic group G onto the derived group of G. The non-abelian exterior square can be defined as the derived subgroup of a Schur cover of G. The isomorphism type of the non-abelian exterior square is unique despite the fact that the isomorphism type of a Schur cover of a polycyclic groups need not be unique. The derived group of a Schur cover has a natural projection onto the derived group of G which is what the function returns.

The kernel of the epimorphism is isomorphic to the Schur multiplicator of G.

gap> G := ExamplesOfSomePcpGroups( 3 );
Pcp-group with orders [ 0, 0 ]
gap> G := DirectProduct( G,G );
Pcp-group with orders [ 0, 0, 0, 0 ]
gap> SchurMultiplicator( G );
[ [ 0, 1 ], [ 2, 3 ] ]
gap> epi := NonAbelianExteriorSquareEpimorphism( G );
[ g2^-2*g5, g4^-2*g10, g6, g7, g8, g9 ] -> [ g2^-2, g4^-2, id, id, id, id ]
gap> Kernel( epi );
Pcp-group with orders [ 0, 2, 2, 2 ]
gap> Collected( AbelianInvariants( last ) );
[ [ 0, 1 ], [ 2, 3 ] ]

computes the non-abelian exterior square of a polycylic group G. See the explanation for NonAbelianExteriorSquareEpimorphism. The natural projection of the non-abelian exterior square onto the derived group of G is stored in the component !.epimorphism.

There is a crossed pairing from G into GwedgeG. See the function SchurExtensionEpimorphism for details. The crossed pairing is stored in the component !.crossedPairing. This is the crossed pairing lambda in EickNickel07.

gap> G := DihedralPcpGroup(0);
Pcp-group with orders [ 2, 0 ]
gap> GwG := NonAbelianExteriorSquare( G );
Pcp-group with orders [ 0 ]
gap> lambda := GwG!.crossedPairing;
function( g, h ) ... end
gap> lambda( G.1, G.2 );
g2^2*g4^-1

returns for a polycyclic group G the projection of the non-abelian tensor square GotimesG onto the non-abelian exterior square GwedgeG. The range of that epimorphism has the component !.epimorphism set to the projection of the non-abelian exterior square onto the derived group of G. See also the function NonAbelianExteriorSquare.

With the result of this function one can compute the groups in the commutative diagram at the beginning of the paper EickNickel07. The kernel of the returned epimorphism is the group nabla(G). The kernel of the composition of this epimorphism and the above mention projection onto G' is the group J(G).

gap> G := DihedralPcpGroup(0);
Pcp-group with orders [ 2, 0 ]
gap> G := DirectProduct(G,G);
Pcp-group with orders [ 2, 0, 2, 0 ]
gap> alpha := NonAbelianTensorSquareEpimorphism( G );
[ g9*g25^-1, g10*g26^-1, g11*g27, g12*g28, g13*g29, g14*g30, g15, g16,
g17,
  g18, g19, g20, g21, g22, g23, g24 ] -> [ g2^-2*g6, g4^-2*g12, g8,
  g9, g10,
  g11, id, id, id, id, id, id, id, id, id, id ]
gap> gamma := Range( alpha )!.epimorphism;
[ g2^-2*g6, g4^-2*g12, g8, g9, g10, g11 ] -> [ g2^-2, g4^-2, id, id,
id, id ]
gap> JG := Kernel( alpha * gamma );
Pcp-group with orders [ 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2 ]
gap> Image( alpha, JG );
Pcp-group with orders [ 2, 2, 2, 2 ]
gap> SchurMultiplicator( G );
[ [ 2, 4 ] ]

\>NonAbelianTensorSquare( <G> )

computes  for a  polycyclic group  <G> the  non-abelian  tensor square
$G\otimes G$.

\beginexample
gap> G := AlternatingGroup( IsPcGroup, 4 );
<pc group of size 12 with 3 generators>
gap> PcGroupToPcpGroup( G );
Pcp-group with orders [ 3, 2, 2 ]
gap> NonAbelianTensorSquare( last );
Pcp-group with orders [ 2, 2, 2, 3 ]
gap> PcpGroupToPcGroup( last );
<pc group of size 24 with 4 generators>
gap> DirectFactorsOfGroup( last );
[ Group([ f1, f2, f3 ]), Group([ f4 ]) ]
gap> List( last, Size );
[ 8, 3 ]
gap> IdGroup( last2[1] );
[ 8, 4 ]       # the quaternion group of Order 8

gap> G := DihedralPcpGroup( 0 );
Pcp-group with orders [ 2, 0 ]
gap> ten := NonAbelianTensorSquare( G );
Pcp-group with orders [ 0, 2, 2, 2 ]
gap> IsAbelian( ten );
true

returns an embedding from the non-abelian exterior square GwedgeG into an extensions of GwedgeG by GtimesG. For the significance of the group see the paper EickNickel07. The range of the epimorphism is the group tau(G) in that paper.

returns the group tau(G) in EickNickel07.

returns an epimorphisms of nu(G) onto tau(G). The group nu(G) is an extension of the non-abelian tensor square GotimesG of G by GtimesG. The group tau(G) is an extension of the non-abelian exterior square GwedgeG by GtimesG. For details see EickNickel07.

returns the group nu(G) in EickNickel07.

returns Whitehead's universal quadratic functor of G, see EickNickel07 for a description.

7.10 Schur covers and Schur towers

A finite p-group G is a Schur tower, if G/gamma_i+1(G) is a Schur cover of G/gamma_i(G) for every i, where gamma_i(G) is the i-th term of the lower central series of G. This section contains a function to determine the Schur covers of a finite p-group up to isomorphism and it gives access to two libraries of Schur tower p-groups.

Let G be a finite p-group defined as a pcp group. This function returns a complete and irredundant set of isomorphism types of Schur covers of G. The algorithm implements a method of Nickel's Phd Thesis.

[Up] [Previous] [Next] [Index]