Quelle morphisms.xml Sprache: XML

<Chapter Label="morphisms">
<Heading>Geometry Morphisms</Heading>

Here we describe what is meant by a <E>geometry morphism</E>
in <Package>FinInG</Package> and the various operations and tools available
to the user.

When using groups in &GAP;, we often use homomorphisms to pass from
one situation to another, even though mathematically it may appear to
be unnecessary, there can be ambiguities if the functionality
is too flexible. This also applies to finite geometry. Take for example
the usual exercise of thinking of a hyperplane in a projective
space as another projective space. To conform with similar situations in
&GAP;, the right thing to do is to embed one projective space into another,
rather than having one projective space automatically as a substructure
of another. The reason for this is that there are many ways one can
do this embedding, even though we may dispense with this choice when we
are working mathematically. So to avoid ambiguity, we stipulate that
one should construct the embedding explicitly. How this is done will
be described this chapter.

Suppose that <M>S</M> and <M>S' are two incidence geometries.
A <E>geometry morphism</E> from <M>S</M> to <M>S' is defined to be a map
from the elements of <M>S</M> to the elements of <M>S' which preserves
incidence and induces a function from the type set of <M>S</M> to the
type set of <M>S'. For instance, a correlation and a collineation are
examples of geometry morphisms, but they have been dealt with in more specific
ways in <Package>FinInG</Package>. We will mainly be concerned with geometry
morphisms where the source and range are different. Hence, the natural embedding
of a projective space in a larger projective space, the mapping induced by field
reduction, and e.g. the Klein correspondence are examples of such geometry morphisms.

As a geometry morphism from <M>S</M> to <M>S' preserves incidence, it also preserves
the symmetry, and hence it induces also a map from the collineation group
of <M>S</M> into the collineation group of <M>S'. Such a map will be called an Intertwiner,
and <Package>FinInG</Package> can provide these maps for some of the geometry morphisms.

Note that quite some technicalities are needed in the implementation of some geometry morphisms.
This chapters deals only with the user interface. Some low level functions for geometry morphisms
are described in Appendix <Ref Chap="morphisms_app"/>.

<Section>
<Heading>Geometry morphisms in FinInG</Heading>

<ManSection>
<Fam Name = "IsGeometryMorphism"/>
<Description>
The category <C>IsGeometryMorphism</C> represents a special object
in <Package>FinInG</Package> which carries attributes and the given element
map. The element map is given as a <C>IsGeneralMapping</C>, and so has a source
and range.
<Example>
<#Include SYSTEM "../examples/include/morphisms_filters.include">
</Example>
The usual operations of <C>ImageElm</C>, and <C>PreImageElm</C>,
have methods installed for geometry
morphisms, as well as the overload operator <C>\^</C>.
</Description>
</ManSection>

<ManSection>
<Attr Name = "Intertwiner" Arg = "f"/>
<Returns>a group homomorphism</Returns>
<Description>
The argument <A>f</A> is a geometry morphism. If <A>f</A> comes equipped with
a natural intertwiner from an automorphism group of the source of <A>f</A> to
the automorphism group to the image of <A>f</A>, then the user may is able
to obtain the intertwiner by calling this operation (see the individual geometry
morphism constructions).
For most geometry morphisms, there is also an accompanying intertwiner for
the automorphism groups of the source and range. Given a geometry morphism
<M>f</M> from <M>S</M> to <M>S', an intertwiner \phi
<Alt Only="HTML">φ</Alt> is a map
from the automorphism group of <M>S</M> to the automorphism group of
<M>S', such that for every element p of S and every
automorphism <M>g</M> of <M>S</M>, we have
<Alt Only="Text"><M>f(p^g) = f(p)^{\phi(g)}.</M></Alt>
<Alt Only="LaTeX"><Display> f(p^g) = f(p)^{\phi(g)}.</Display></Alt>
<Alt Only="HTML MathJax"><Display> f(p^g) = f(p)^{\phi(g)}.</Display></Alt>
<Alt Only="HTML noMathJax"><Display>f(pg)=f(p)φ(g).</Display></Alt>
There is no method to compute an intertwiner for a given
geometry morphism, the attribute is or is not set during the construction of the geometry
morphism, depending whether the Source and Range of the morphism have the appropriate
automorphism group known as an attribute. When this condition is not satisfied, the
user is expected to call the appropriate automorphism groups, so that they are computed,
and to recompute the geometry morphism (which will not cost a lot of computation time then). This will make
the attribute <F>Intertwiner</F> available. Here is a simple example of the
intertwiner for the isomorphism of two polar spaces (see <Ref Oper="IsomorphismPolarSpaces" Style="Text"/>).
The source of the homomorphism is dependent on the geometry.
<Example>
<#Include SYSTEM "../examples/include/morphisms_intertwiners.include">
</Example>
</Description>
</ManSection>
</Section>

<Section>
<Heading>Type preserving bijective geometry morphisms</Heading>

An important class of geometry morphisms in <Package>FinInG</Package> are the isomorphisms between
polar spaces of the same kind that are induced by coordinate transformations.

<ManSection>
<Oper Name="IsomorphismPolarSpaces" Arg="ps1, ps2"/>
<Oper Name="IsomorphismPolarSpaces" Arg="ps1, ps2, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The arguments <A>ps1</A> and <A>ps2</A> are <E>equivalent</E> polar spaces, i.e. up to
coordinate transformation, the underlying sesquilinear or quadratic form determines the
same polar space, or, <A>ps1</A> is a parabolic quadric over a finite field <M>f</M> of even characteristic in
dimension <M>2n</M> and <A>ps2</A> is a symplectic space over <M>f</M> in dimension <M>2n-1</M>, then
this operation returns a geometry isomorphism between them.
The optional third argument <A>boolean</A> can take either <C>true</C> or
<C>false</C> as input, and then the operation will or will not compute the
intertwiner accordingly. The user may wish that the intertwiner is not computed
when working with large polar spaces. The default (when calling the operation
with two arguments) is set to <C>true</C>, and in this case,
if at least one of <A>ps1</A> or <A>ps2</A> has
a collineation group installed as an attribute, then an intertwining
homomorphism is installed as an attribute of the resulting geometry morphism.
Hence we also obtain a natural group isomorphism from the collineation group of
<A>ps1</A> onto the collineation group of <A>ps2</A> (see also <Ref Attr="Intertwiner" Style="Text"/>).
<Example>
<#Include SYSTEM "../examples/include/morphisms_isopolar.include">
</Example>
</Description>
</ManSection>

</Section>

<Section Label="sec:klein">
<Heading>Klein correspondence and derived dualities</Heading>

The Klein correspondence is a well known geometry morphism from the lines
of &pg3q; to the points of a hyperbolic quadric in &pg5q;.
This morphism and some derived morphisms are provided in <Package>FinInG</Package>.

The bare essential of the Klein correspondence is the so-called Plücker map.

<ManSection>
<Oper Name="PluckerCoordinates" Arg="line"/>
<Description>
This operation takes a line of &pg3q; as argument. It returns the plucker coordinates of the argument as list of finite
field elements. The returned list can be used in operations as <F>vector spaceToElement</F>, and represents a point of the hyperbolic
quadric in &pg5q; with equation <Alt Only="HTML MathJax"><M>X_0X_5+X_1X_4+X_2X_3 = 0</M></Alt>
<Alt Not="HTML"><M>X_0X_5+X_1X_4+X_2X_3 = 0</M></Alt>
<Alt Only="HTML noMathJax">X0X5+X1X4+X2X3 = 0</Alt>
<Example>
<#Include SYSTEM "../examples/include/morphisms_plucker.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="KleinCorrespondence" Arg="f"/>
<Oper Name="KleinCorrespondence" Arg="f, boolean"/>
<Oper Name="KleinCorrespondence" Arg="q"/>
<Oper Name="KleinCorrespondence" Arg="q, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The argument <A>f</A> is a finite field, the argument <A>q</A> is a prime power.
The first an the third version use <K>true</K> as value for <A>boolean</A>.
When using <K>true</K> as value for the boolean, the intertwiner is computed.
This variant of the operation <F>KleinCorrespondence</F> has always as ambient geometry of
its range the hyperbolic quadric &qplus5q;
with equation <Alt Not="HTML"><M>X_0X_5+X_1X_4+X_2X_3 = 0</M></Alt>
<Alt Only="HTML noMathJax">X0X5+X1X4+X2X3 = 0</Alt>
The returned geometry morphism has the lines of &pg3q; as source and the points of &qplus5q; as range.
<Example>
<#Include SYSTEM "../examples/include/morphisms_klein.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="KleinCorrespondence" Arg="quadric"/>
<Oper Name="KleinCorrespondence" Arg="quadric, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The argument <A>quadric</A> is a hyperbolic quadric in a 5 dimensional projective space.
If <A>boolean</A> is <K>true</K> or not given, this operation returns the geometry morphism
equipped with an intertwiner. The returned geometry morphism has the lines of &pg3q; as source
and the points of &qplus5q; as range.
<Example>
<#Include SYSTEM "../examples/include/morphisms_klein2.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="KleinCorrespondenceExtended" Arg="quadric"/>
<Oper Name="KleinCorrespondenceExtended" Arg="quadric, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The argument <A>quadric</A> is a hyperbolic quadric in a 5 dimensional projective space.
If <A>boolean</A> is <K>true</K> or not given, this operation returns the geometry morphism
equipped with an intertwiner. The returned geometry morphism has all the elements of &pg3q;
as source (not just the lines) and the elements of &qplus5q; as range, hence this operation
is a kind of extension of <F>KleinCorrespondence</F>.
<Example>
<#Include SYSTEM "../examples/include/morphisms_klein3.include">
</Example>
</Description>
</ManSection>

It is well known that the classical generalised quadrangles &w3q; and &q4q;
are dual incidence structures, the same holds for the classical generalised quadrangles
&qmin5q; and &h3q2;.
Essentially, these dual dualities are based on the Klein correspondence, and
are implemented through the operation <F>NaturalDuality</F>,
this operation will return a geometry morphism with <C>ElementsOfIncidenceStructure(gq1)</C> as
source and <C>ElementsOfIncidenceStructure(gq2)</C> as range, in other words, it is a geometry
morphism from all the elements of <A>gq1</A> onto all the elements of <A>gq2</A>, preserving the incidence,
and swapping the types.

<ManSection>
<Oper Name="NaturalDuality" Arg="gq1, gq2"/>
<Oper Name="NaturalDuality" Arg="gq1, gq2, boolean"/>
<Oper Name="NaturalDuality" Arg="gq"/>
<Oper Name="NaturalDuality" Arg="gq, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The operation allows the construction of the duality between &w3q; and &q4q;,
respectively &qmin5q;, in two directions. It is checked if the arguments are appropriate,
i.e. the right type of generalised quadrangle(s).

The first version requires two arguments: either the symplectic or parabolic
quadrangle, in any order, and defined by any suitable bilinear/quadratic and bilinear form; or the
elliptic or hermitian quadrangle (in dimension <M>3</M>), in any order, and defined by any suitable
bilinear/quadratic and hermitian form. In all cases the generalised quadrangles may be the standard
one provided by the package <Package>FinInG</Package>. 

The third version requires only one argument, either &w3q;, &q4q;,
&qmin5q;, or &h3q2;,
standard or user specified using an appropriate bilinear, quadratic or hermitian
form. The range of the returned geometry morphism will be the set of all elements of a suitable generalised
quadrangle, in standard form.

The first and third version without a boolean as argument will, if possible return a geometry morphism
equipped with an intertwiner. Using the boolean argument <A>false</A> will return a geometry morphism
that is not equipped with an intertwiner.
<Example>
<#Include SYSTEM "../examples/include/morphisms_duality.include">
</Example>
</Description>
</ManSection>

The combination of the isomorphism of the GQs &w3q;, &q4q; when <M>q</M> is even
and the duality between the same GQs, yields a duality from each of these GQs itself. The operation
<F>SelfDuality</F> implements this combination.

<ManSection>
<Oper Name="SelfDuality" Arg="gq"/>
<Oper Name="SelfDuality" Arg="gq, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
It is checked whether the base field of <A>gq</A> is a field of characteristic 2 and whether
<A>gq</A> is a symplectic generalised quadrangle in 3-dimensional projective space or a parabolic
quadric in 4-dimensional projective space. The first version will return, when possible,
a geometry morphism equipped with an intertwiner. Using the boolean argument <A>false</A> will
return a geometry morphism that is not equipped with an intertwiner. The example shows the use
of the boolean argument.
<Example>
<#Include SYSTEM "../examples/include/morphisms_selfduality.include">
</Example>
</Description>
</ManSection>
</Section>

<Section>
<Heading>Embeddings of projective spaces</Heading>

The most natural of geometry morphisms include, for example, the embedding of a
projective space into another via a subspace, the embedding of a projective
space over a field into a projective space of the same dimension over an extended
field, or the embedding of a projective space over a field into a projective
space of higher dimension over a subfield through so-called field reduction.

<ManSection>
<Oper Name="NaturalEmbeddingBySubspace" Arg="geom1, geom2, v"/>
<Returns>a geometry morphism</Returns>
<Description>
The arguments <A>geom1</A> and <A>geom2</A> are both projective spaces, and <A>v</A> is an element of a geom2.
This function returns a geometry morphism representing the natural embedding of <A>geom1</A>
into <A>geom2</A> as the subspace <A>v</A>. Hence <A>geom1</A> and <A>v</A> must
be equivalent as geometries. An Intertwiner is not implemented for this geometry morphism.
<Example>
<#Include SYSTEM "../examples/include/morphisms_embedding1.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="NaturalEmbeddingBySubField" Arg="geom1, geom2"/>
<Oper Name="NaturalEmbeddingBySubField" Arg="geom1, geom2, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The arguments <A>geom1</A> and <A>geom2</A> are projective spaces of
the same dimension. This function returns a geometry morphism representing the natural embedding
of <A>geom1</A> into <A>geom2</A> as a subfield geometry.
The geometry morphism also comes equipped with an intertwiner (see
<Ref Attr="Intertwiner" Style="Text"/>). The optional third argument <A>boolean</A>
can take either <C>true</C> or
<C>false</C> as input, and then our operation will or will not compute the
intertwiner accordingly. The default (when calling the operation
with two arguments) is set to <C>true</C>. Note that the source of the intertwiner is the
projectivity group of <A>geom1</A> and its range is a subgroup of the projectivity group
of <A>geom2</A>. Here is a simple example where the geometry morphism embeds
<Alt Only="HTML MathJax"><M>\mathrm{PG}(2,3)</M></Alt><Alt Only="HTML noMathJax">PG(2,3)</Alt>
<Alt Only="LaTeX"><M>\mathrm{PG}(2,3)</M></Alt><Alt Only="Text">PG(2,3)</Alt> into
<Alt Only="HTML MathJax"><M>\mathrm{PG}(2,9)</M></Alt><Alt Only="HTML noMathJax">PG(2,9)</Alt>
<Alt Only="LaTeX"><M>\mathrm{PG}(2,9)</M></Alt><Alt Only="Text">PG(2,9)</Alt>.
<Example>
<#Include SYSTEM "../examples/include/morphisms_subfield1.include">
</Example>
</Description>
</ManSection>

<Subsection Label="proj_red">
<Heading> Embedding of projective spaces by field reduction</Heading>
We briefly describe the mathematics behind field reduction. For more details we
refer to <Cite Key="LaVa"/>. Consider the fields &gfq; and &gfqt;. The field &gfqt; is a
<M>t</M>-dimensional vector space over &gfq;. Hence, with respect to a chosen basis
<M>B</M> for &gfqt; as a &gfq;-vector space, the bijection between the vector spaces
&vnqt; and <M>V(tn,q)</M> can be implemented.

Consider the projective space &pgnminus1qt;. The elements are represented by subspaces
of &vnqt;. Clearly, a <M>k</M> dimensional subspace of
&vnqt; is also a <M>kn</M>-dimensional subspace of <M>V(nt,q)</M>.
This induces an embedding from &pgnminus1qt; into &pgntminus1q;.
The embedding will be determined by the chosen basis of &gfqt; as a vector space over &gfq;
</Subsection>

<ManSection>
<Oper Name="BlownUpSubspaceOfProjectiveSpace" Arg="B, subspace"/>
<Returns>a subspace of a projective space</Returns>
<Description>
Let <A>B</A> be a basis for the field &gfqt; as &gfq; vector space, and
let <A>subspace</A> be a <M>k-1</M>-dimensional subspace of
&pgnminus1qt; represented by a <M>k</M>-dimensional subspace <M>S</M> of &vnqt;.
This operation returns the <M>kt-1</M>-dimensional subspace of &pgntminus1q;
represented by blowing up <M>S</M> with respect to the base <A>B</A>. This operation
relies on the &GAP; operation <F>BlownUpMat</F>. In the example, the effect
of choosing a different basis is shown.
<Example>
<#Include SYSTEM "../examples/include/morphisms_blownupsubspace.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="geom1, f2, B"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="geom1, f2"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="geom1, geom2"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="geom1, geom2, B"/>
<Returns>a geometry morphism</Returns>
<Description>
This operation comes in four flavours. For the first flavour, the argument <A>geom1</A> is a projective
space over a field &gfqt;. The argument <A>f2</A> is a subfield &gfq; of &gfqt;. The argument <M>B</M>
is a basis for &gfqt; as a &gfq;-vector space. When this argument is not given, a basis for &gfqt; over &gfq;
is computed using <F>Basis(Asvector space(GF(q),GF(q^t)))</F>. It is checked whether <A>f2</A> is a
subfield of the base field of <A>geom1</A>. The third and fourth flavour are comparable,
where now &gfq; is found as the base field of <A>geom2</A>. In fact the arguments <A>geom1</A> and
<A>geom2</A> are the projective spaces &pgnminus1qt; and &pgntminus1q; respectively. As in the previous flavours, the
argument <A>B</A> is optional. An intertwiner is always available for this geometry morphism, and has source the
homography group of <A>geom1</A> and as range a subgroup of the homography group of <A>geom2</A> (or the projective space
of the appropriate dimension over <A>f2</A>.
Notice in the example below the difference of a factor 2 in the orders of the group,
which comes of course from restricting the homomorphism to the homography group, which
differs a factor 2 from the collineation group of the projective line, that has an extra
automorphism of order two, corresponding with the Frobenius automorphism.
<Example>
<#Include SYSTEM "../examples/include/morphisms_fieldreducproj.include">
</Example>
</Description>
</ManSection>

</Section>

<Section>
<Heading>Embeddings of polar spaces</Heading>

<ManSection>
<Oper Name="NaturalEmbeddingBySubspace" Arg="geom1, geom2, v"/>
<Returns>a geometry morphism</Returns>
<Description>
The arguments <A>geom1</A> and <A>geom2</A> both polar spaces, and <A>v</A> is an element of a projective space.
This function returns a geometry morphism representing the natural embedding of <A>geom1</A>
into the intersection of <A>geom2</A> and <A>v</A>. Hence the intersection of <A>geom2</A> and <A>v</A>
must induce a polar space of the same type as <A>geom1</A> in <A>v</A>. This operation performs all necessary checks.
An Intertwiner is not implemented for this geometry morphism.
<Example>
<#Include SYSTEM "../examples/include/morphisms_embedding2.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="NaturalEmbeddingBySubField" Arg="geom1, geom2"/>
<Oper Name="NaturalEmbeddingBySubField" Arg="geom1, geom2, boolean"/>
<Returns>a geometry morphism</Returns>
<Description>
The arguments <A>geom1</A> and <A>geom2</A> are projective or polar spaces with an underlying vector
space of the same dimension and the base field <M>L</M> of <A>geom2</A> is an extension of the base field
<M>K</M> of <A>geom1</A>. The form <M>f</M> determining <A>geom1</A> also defines a form over <M>L</M>,
and determines a polar space. By considering the underlying vector spaces determining the elements of <A>geom1</A>
over the extension field <M>L</M>, there is an obvious embedding of <A>geom1</A> in the polar space over the
extension field. Considering <M>f</M> over a field extension might change its type. The possible embeddings,
where the polar spaces may be chosen up to equivalent form,
are listed in the table below (see <Cite Key="KleidmanLiebeck"/>):

<Table Align="l|l|l">
<Caption>Subfield embeddings of polar spaces</Caption>
<Row>
<Item>Polar Space 1</Item><Item>Polar Space 2</Item><Item>Conditions</Item>
</Row>
<HorLine/>
<HorLine/>
<Row>
<Item><Alt Only="HTML MathJax"><M>\mathrm{W}(2n-1,q)</M></Alt><Alt Only="HTML noMathJax"><M>W(2n-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{W}(2n-1,q)</M></Alt><Alt Only="Text"><M>W(2n-1,q)</M></Alt></Item>
<Item><Alt Only="HTML MathJax"><M>\mathrm{W}(2n-1,q^a)</M></Alt><Alt Only="HTML noMathJax">W(2n-1,qa)</Alt><Alt Only="LaTeX"><M>\mathrm{W}(2n-1,q^a)</M></Alt><Alt Only="Text"><M>W(2n-1,q^a)</M></Alt></Item>
<Item>--</Item>
</Row>
<Row>
<Item><Alt Only="HTML MathJax"><M>\mathrm{W}(2n-1,q)</M></Alt><Alt Only="HTML noMathJax"><M>W(2n-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{W}(2n-1,q)</M></Alt><Alt Only="Text"><M>W(2n-1,q)</M></Alt></Item>
<Item><Alt Only="HTML MathJax"><M>\mathrm{H}(2n-1,q^a)</M></Alt><Alt Only="HTML noMathJax">H(2n-1,qa)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(2n-1,q^a)</M></Alt><Alt Only="Text"><M>H(2n-1,q^a)</M></Alt></Item>
<Item>--</Item>
</Row>
<Row>
<Item><Alt Only="HTML MathJax"><M>\mathrm{H}(d,q^2)</M></Alt><Alt Only="HTML noMathJax">H(d,q2)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(d,q^2)</M></Alt><Alt Only="Text"><M>H(d,q^2)</M></Alt></Item>
<Item><Alt Only="HTML MathJax"><M>\mathrm{H}(d,q^{2r})</M></Alt><Alt Only="HTML noMathJax">H(d,q2r)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(d,q^{2r})</M></Alt><Alt Only="Text"><M>H(d,q^(2r))</M></Alt></Item>
<Item><M>r</M> odd</Item>
</Row>
<Row>
<Item><Alt Only="HTML MathJax"><M>\mathrm{O}^{\epsilon}(d,q)</M></Alt><Alt Only="HTML noMathJax">Oε(d,q)</Alt><Alt Only="LaTeX"><M>\mathrm{O}^{\epsilon}(d,q)</M></Alt><Alt Only="Text"><M>O^\epsilon(d,q)</M></Alt></Item>
<Item><Alt Only="HTML MathJax"><M>\mathrm{H}(d,q^2)</M></Alt><Alt Only="HTML noMathJax">H(d,q2)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(d,q^2)</M></Alt><Alt Only="Text"><M>H(d,q^2)</M></Alt></Item>
<Item><M>q</M> odd</Item>
</Row>
<Row>
<Item><Alt Only="HTML MathJax"><M>\mathrm{O}^{\epsilon}(d,q)</M></Alt><Alt Only="HTML noMathJax">Oε(d,q)</Alt><Alt Only="LaTeX"><M>\mathrm{O}^{\epsilon}(d,q)</M></Alt><Alt Only="Text"><M>O^\epsilon(d,q)</M></Alt></Item>
<Item><Alt Only="HTML MathJax"><M>\mathrm{O}^{\epsilon'}(d,q^r)Oε'(d,qr)</Alt><Alt Only="LaTeX"><M>\mathrm{O}^{\epsilon'}(d,q^r)O^\epsilon'(d,q^r)</M></Alt></Item>
<Item><Alt Not="HTML"><M>\epsilon=(\epsilon')^r\epsilon=(\epsilon')^r</M></Alt><Alt Only="HTML noMathJax">ε=(ε')r
</Row>
</Table>

The geometry morphism also comes equipped with an intertwiner (see
<Ref Attr="Intertwiner" Style="Text"/>). The optional third argument <A>boolean</A>
can take either <C>true</C> or
<C>false</C> as input, and then our operation will or will not compute the
intertwiner accordingly. When set <C>true</C>, the intertwiner
will be computed if <F>HasCollineationGroup(geom1)</F> is true.
The user may wish that the intertwiner is not computed when embedding large polar spaces.
The default (when calling the operation with two arguments) is set to <C>true</C>.
<Example>
<#Include SYSTEM "../examples/include/morphisms_subfield2.include">
</Example>
</Description>
</ManSection>

<Subsection Label="polar_red">
<Heading>Embedding of polar spaces by field reduction</Heading>
Field reduction for polar spaces is somewhat more involved than for
projective spaces, and we give a brief description. Let <M>L</M> be the
field &gfqt; and let <M>K</M> be the field &gfq; Let <M>P</M> be a polar
space over a field <M>L</M>. Let <M>f</M> be the form on the <M>r</M> dimensional
vector space <M>V</M> over <M>L</M> determining <M>P</M>. Consider the trace map

<Alt Only="HTML MathJax"><Display>T: L \to K: x \mapsto x^{q^t}+ x^{q^{t-1}} + \ldots + x.</Display></Alt>
<Alt Only="HTML noMathJax"><M>T: L \to K: x \mapsto </M> xqt + xqt-1 + … + x.</Alt> 
<Alt Only="LaTeX"><Display>T: L \to K: x \mapsto x^{q^t}+ x^{q^{t-1}} + \ldots + x.</Display></Alt>
<Alt Only="Text"><M>T: L \to K: x \mapsto x^(q^t)+ x^(q^(t-1)) + \ldots + x.</M></Alt>

Define for any <M>\alpha \in L</M> the map

<Alt Only="HTML MathJax"><Display>T_\alpha: L \to K: x \mapsto T_\alpha(x) = T(\alpha x).</Display></Alt>
<Alt Only="HTML noMathJax">Tα: L: → K: x<M>\mapsto</M> Tα = T(α x).</Alt>
<Alt Only="LaTeX"><Display>T_\alpha: L \to K: x \mapsto T_\alpha(x) = T(\alpha x).</Display></Alt>
<Alt Only="Text"><M>T_\alpha: L \to K: x \mapsto T_\alpha(x) = T(\alpha x).</M></Alt>

Consider the <M>rt</M> dimensional vector space <M>W</M> over <M>K</M>. There is a bijective map Φ: <M>V \to W</M>
and <Alt Only="HTML noMathJax">Tα • f • Φ -1</Alt>
<Alt Only="HTML MathJax"><M>T_{\alpha} \circ f \circ \Phi^{-1}</M></Alt>
<Alt Not="HTML"><M>T_{\alpha} \circ f \circ \Phi^{-1}</M></Alt> defines a
quadratic or sesquilinear form (depending on <M>\alpha</M>, and <M>f</M> being quadratic
or sesquilinear) acting on <M>W</M>, and hence, if not singular or degenerate, inducing a
polar space <M>S</M> over the finite field <M>GF(q)</M>. An element of <M>P</M> can be
mapped onto an element of <M>S</M> by simply blowing up <M>P</M>
using field reduction for projective spaces. So the resulting polar space <M>S</M> is
dependent on the original form <M>f</M>, the parameter <M>\alpha</M> and the blowing up of
elements by field reduction, the latter being dependent on the basis of <M>L</M> as
a <M>K</M> vector space. <Package>FinInG</Package> provides two approaches.
The first approach starts from <M>P</M> and the parameters <M>K</M>, <M>\alpha</M>
and a basis for <M>L</M> as <M>K</M> vector space. Then the resulting form
<Alt Only="HTML noMathJax">Tα • f • Φ -1</Alt>
<Alt Only="HTML MathJax"><M>T_{\alpha} \circ f \circ \Phi^{-1}</M></Alt>
<Alt Not="HTML"><M>T_{\alpha} \circ f \circ \Phi^{-1}</M></Alt>
is determined, and the associated polar space <M>S</M> will be the range of the embedding.
Note that the resulting polar space will not necessarily be canonical. The second approach starts
from two given polar spaces <M>P</M> and <M>S</M>. Based on this input, it is determined whether an embedding
based on the above described principle is possible, and the necessary parameters are computed. The resulting
embedding is a geometry morphism from <M>P</M> to <M>S</M>. Note that the polar spaces used as an argument
may be freely chosen and are not required to be in the canonical form.
For more information on the embeddings by field reduction of polar spaces, including conditions on
the parameter <M>\alpha</M>, we refer to <Cite Key="Gill"/> and <Cite Key="LaVa"/>.
The possible embeddings are listed in the following table.

<Table Align="l|l|l">
<Caption>Field reduction of polar spaces</Caption>
 <Row>
 <Item>Polar Space 1</Item><Item>Polar Space 2</Item><Item>Conditions</Item>
 </Row>
<HorLine/>
<HorLine/>
 <Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{W}(2n-1,q^t)</M></Alt><Alt Only="HTML noMathJax">W(2n-1,qt)</Alt><Alt Only="LaTeX"><M>\mathrm{W}(2n-1,q^t)</M></Alt><Alt Only="Text"><M>W(2n-1,q^t)</M></Alt></Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{W}(2nt-1,q)</M></Alt><Alt Only="HTML noMathJax"><M>W(2nt-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{W}(2nt-1,q)</M></Alt><Alt Only="Text"><M>W(2nt-1,q)</M></Alt></Item><Item>--</Item>
 </Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^+(2n-1,q^t)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^+(2n-1,q^t)</M></Alt><Alt Only="HTML noMathJax">Q+(2n-1,qt)</Alt><Alt Only="Text"><M>Q^+(2n-1,q^t)</M></Alt></Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^+(2nt-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^+(2nt-1,q)</M></Alt><Alt Only="HTML noMathJax">Q+(2nt-1,q)</Alt><Alt Only="Text"><M>Q^+(2nt-1,q)</M></Alt></Item>
 <Item>--</Item>
 </Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^-(2n-1,q^t)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^-(2n-1,q^t)</M></Alt><Alt Only="HTML noMathJax">Q-(2n-1,qt)</Alt><Alt Only="Text"><M>Q^-(2n-1,q^t)</M></Alt></Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^-(2nt-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^-(2nt-1,q)</M></Alt><Alt Only="HTML noMathJax">Q-(2nt-1,q)</Alt><Alt Only="Text"><M>Q^-(2nt-1,q)</M></Alt></Item>
 <Item>--</Item>
 </Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}(2n,q^{2a+1})</M></Alt><Alt Only="HTML noMathJax">Q(2n,q2a+1)</Alt><Alt Only="LaTeX"><M>\mathrm{Q}(2n,q^{2a+1})</M></Alt><Alt Only="Text"><M>Q(2n,q^(2a+1))</M></Alt>
</Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}((2a+1)(2n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax"><M>Q((2a+1)(2n+1)-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}((2a+1)(2n+1)-1,q)</M></Alt><Alt Only="Text"><M>Q((2a+1)(2n+1)-1,q)</M></Alt></Item>
 <Item><M>q</M> odd</Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}(2n,q^{2a})</M></Alt><Alt Only="HTML noMathJax">Q(2n,q2a)</Alt><Alt Only="LaTeX"><M>\mathrm{Q}(2n,q^{2a})</M></Alt><Alt Only="Text"><M>Q(2n,q^(2a))</M></Alt>
</Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^-(2a(2n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax">Q-(2a(2n+1)-1,q)</Alt><Alt Only="LaTeX"><M>\mathrm{Q}^-(2a(2n+1)-1,q)</M></Alt><Alt Only="Text"><M>Q^-(2a(2n+1)-1,q)</M></Alt></Item>
 <Item><Alt Not="HTML"><M>q \equiv 1 \pmod{4}</M></Alt>
 <Alt Only="HTML MathJax"><M>q \equiv 1 \pmod{4}</M></Alt>
 <Alt Only="HTML noMathJax">q ≡ 1 mod 4</Alt></Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}(2n,q^{4a+2})</M></Alt><Alt Only="HTML noMathJax">Q(2n,q4a+2)</Alt><Alt Only="LaTeX"><M>\mathrm{Q}(2n,q^{4a+2})</M></Alt><Alt Only="Text"><M>Q(2n,q^(4a+2))</M></Alt></Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^+((4a+2)(2n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax">Q+((4a+2)(2n+1)-1,q)</Alt><Alt Only="LaTeX"><M>\mathrm{Q}^+((4a+2)(2n+1)-1,q)</M></Alt><Alt Only="Text"><M>Q^+((4a+2)(2n+1)-1,q)</M></Alt></Item>
 <Item><Alt Not="HTML"><M>q \equiv 3 \pmod{4}</M></Alt>
 <Alt Only="HTML MathJax"><M>q \equiv 3 \pmod{4}</M></Alt>
 <Alt Only="HTML noMathJax">q ≡ 3 mod 4</Alt></Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}(2n,q^{4a})</M></Alt><Alt Only="HTML noMathJax">Q(2n,q4a)</Alt><Alt Only="LaTeX"><M>\mathrm{Q}(2n,q^{4a})</M></Alt><Alt Only="Text"><M>Q(2n,q^(4a))</M></Alt>
</Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^-(4a(2n+1)-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^-(4a(2n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax">Q-(4a(2n+1)-1,q)</Alt><Alt Only="Text"><M>Q^-(4a(2n+1)-1,q)</M></Alt>
</Item>
 <Item><Alt Not="HTML"><M>q \equiv 3 \pmod{4}</M></Alt>
 <Alt Only="HTML MathJax"><M>q \equiv 3 \pmod{4}</M></Alt>
 <Alt Only="HTML noMathJax">q ≡ 3 mod 4</Alt></Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{H}(n,q^{2a+1})</M></Alt><Alt Only="HTML noMathJax">H(n,q2a+1)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(n,q^{2a+1})</M></Alt><Alt Only="Text"><M>H(n,q^(2a+1))</M></Alt>
</Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{H}((2a+1)(n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax"><M>H((2a+1)(n+1)-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{H}((2a+1)(n+1)-1,q)</M></Alt><Alt Only="Text"><M>H((2a+1)(n+1)-1,q)</M></Alt>
</Item>
 <Item><M>q</M> square</Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{H}(n,q^{2a})</M></Alt><Alt Only="HTML noMathJax">H(n,q2a)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(n,q^{2a})</M></Alt><Alt Only="Text"><M>H(n,q^(2a))</M></Alt>
</Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{W}(2a(n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax"><M>W(2a(n+1)-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{W}(2a(n+1)-1,q)</M></Alt><Alt Only="Text"><M>W(2a(n+1)-1,q)</M></Alt></Item>
 <Item><M>q</M> even</Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{H}(2n,q^{2a})</M></Alt><Alt Only="HTML noMathJax">H(2n,q2a)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(2n,q^{2a})</M></Alt><Alt Only="Text"><M>H(2n,q^(2a))</M></Alt></Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^-(2a(2n+1)-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^-(2a(2n+1)-1,q)</M></Alt><Alt Only="HTML noMathJax">Q-(2a(2n+1)-1,q)</Alt><Alt Only="Text"><M>Q^-(2a(2n+1)-1,q)</M></Alt></Item>
 <Item><M>q</M> odd</Item>
</Row>
<Row>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{H}(2n+1,q^{2a})</M></Alt><Alt Only="HTML noMathJax">H(2n+1,q2a)</Alt><Alt Only="LaTeX"><M>\mathrm{H}(2n+1,q^{2a})</M></Alt><Alt Only="Text"><M>H(2n+1,q^(2a))</M></Alt></Item>
 <Item><Alt Only="HTML MathJax"><M>\mathrm{Q}^+(2a(2n+2)-1,q)</M></Alt><Alt Only="LaTeX"><M>\mathrm{Q}^+(2a(2n+2)-1,q)</M></Alt><Alt Only="HTML noMathJax">Q+(2a(2n+2)-1,q)</Alt><Alt Only="Text"><M>Q^+(2a(2n+2)-1,q)</M></Alt></Item>
 <Item><M>q</M> odd</Item>
</Row>
</Table>
</Subsection>

<ManSection>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1,f2, alpha, basis, bool"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1,f2, alpha, basis"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1,f2, alpha, bool"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1,f2, alpha"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1,f2, bool"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1,f2"/>
<Returns>a geometry morphism</Returns>
<Description>
<A>ps1</A> is a polar space over a field extension <M>L</M> of <A>f2</A>, <A>basis</A> is a
basis for <M>L</M> over <A>f2</A>, <A>alpha</A> is a non-zero element of <M>L</M>.
The version of <F>NaturalEmbeddingByFieldReduction</F> implements the first approach as explained in
<Ref Sect="polar_red"/>. When no argument <A>basis</A> is given, a basis for <M>L</M> over <A>f2</A>
is computed using <F>Basis(Asvector space(K,L))</F>. When no argument <A>alpha</A> is given,
<F>One(f2)</F> is used as value for <A>alpha</A>. When <A>bool</A> is true or not given, an intertwiner
is computed, when <A>bool</A> is <A>false</A>, no intertwiner is computed. This intertwiner has as
its domain the isometry group of <A>ps1</A>. The user may wish that the intertwiner is not computed
when embedding large polar spaces. The default (when calling the operation
with two arguments) is set to <C>true</C>. In the first example, we construct a spread of maximal
subspaces (solids) in a 7 dimensional symplectic space. We compute a subgroup of its stabilizer
group using the intertwiner. In the second example, we construct a linear blocking set of the
symplectic generalised quadrangle over <Alt Only="HTML MathJax"><M>\mathrm{GF}(9)</M></Alt><Alt Only="HTML noMathJax">GF(9)</Alt>
<Alt Only="LaTeX"><M>\mathrm{GF}(9)</M></Alt><Alt Only="Text">GF(9)</Alt>.
<Example>
<#Include SYSTEM "../examples/include/morphisms_fieldreducpolar1.include">
<#Include SYSTEM "../examples/include/morphisms_fieldreducpolar2.include">
</Example>
</Description>
</ManSection>

<ManSection>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1, ps2, bool"/>
<Oper Name="NaturalEmbeddingByFieldReduction" Arg="ps1, ps2"/>
<Returns>a geometry morphism</Returns>
<Description>
If <A>ps1</A> and <A>ps2</A> are two polar spaces which are suitable for field
reduction as listed in the table with possible embeddings in Section
<Ref Sect="polar_red"/>, then this operation returns the corresponding embedding.
An intertwiner is computed if the third argument <A>bool</A> is true, or if there
is no third argument. This intertwiner has as its domain the isometry group of <A>ps1</A>.
The example shows two cases where a spread is computed, including a subgroup of its stabiliser group using
the intertwiner.
<Example>
<#Include SYSTEM "../examples/include/morphisms_fieldreducpolar3.include">
</Example>
</Description>
</ManSection>
</Section>

<Section>
<Heading>Projections</Heading>

<ManSection>
<Oper Name="NaturalProjectionBySubspace" Arg="ps, v"/>
<Returns>a geometry morphism</Returns>
<Description>
The argument <A>ps</A> is a projective or polar space, and <A>v</A> is a
subspace of <A>ps</A>. In the case that <A>ps</A> is a projective space,
the geometry of subspaces containing <A>v</A> is a projective space
of lower dimension over the same base field, and this operation returns the
corresponding geometry morphism. In the case that <A>ps</A> is a polar space,
the geometry of elements pf <A>ps</A> containing <A>v</A> is a polar space
of lower rank and of the same type over the same base field, and this operation
returns the corresponding geometry morphism.
It is checked whether <A>v</A> is a subspace of <A>ps</A>,
and whether the input of the function and preimage of the returned
geometry morphism is valid or not. There is a shorthand for
this operation which is basically an overload of the quotient operation. So,
for example, <C>ps / v</C> achieves the same thing
as <C>AmbientGeometry(Range(NaturalProjectionBySubspace(ps, v)))</C>.
An intertwiner is not available for this geometry morphism.
<Example>
<#Include SYSTEM "../examples/include/morphisms_projection.include">
</Example>
</Description>
</ManSection>

</Section>

<Section>
<Heading>Projective completion</Heading>

<ManSection>
<Oper Name="ProjectiveCompletion" Arg="as"/>
<Returns>a geometry morphism</Returns>
<Description>
The argument <A>as</A> is an affine space. This operation returns an embedding
of <A>as</A> into the projective space <A>ps</A> of the same dimension, and over the same field.
For example, the point <M>(x, y, z)</M> is mapped onto the projective point with
homogeneous coordinates <M>(1, x, y, z)</M>. An intertwiner is unnecessary,
<F>CollineationGroup(as)</F> is a subgroup of <F>CollineationGroup(ps)</F>.
<Example>
<#Include SYSTEM "../examples/include/morphisms_completion.include">
</Example>
</Description>
</ManSection>

</Section>

</Chapter>

quality93%

¤ Dauer der Verarbeitung: 0.40 Sekunden (vorverarbeitet) ¤

Wurzel

Suchen

Beweissystem der NASA

Beweissystem Isabelle

NIST Cobol Testsuite

Cephes Mathematical Library

Wiener Entwicklungsmethode

Haftungshinweis

Die Informationen auf dieser Webseite wurden nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit, noch Qualität der bereit gestellten Informationen zugesichert.

Bemerkung:

Die farbliche Syntaxdarstellung ist noch experimentell.