<Chapter Label="Finding cliques and independent sets">
<Heading>Finding cliques and independent sets</Heading>
In &Digraphs;, a <E>clique</E> of a digraph is a set of mutually adjacent
vertices of the digraph, and an <E>independent set</E> is a set of mutually
non-adjacent vertices of the digraph. A <E>maximal clique</E> is a clique
which is not properly contained in another clique, and a <E>maximal
independent set</E> is an independent set which is not properly contained
in another independent set. Using this definition in &Digraphs;, cliques
and independent sets are both permitted, but not required, to contain vertices
at which there is a loop. Another name for a clique is a <E>complete
subgraph</E>.
<P/>
&Digraphs; provides extensive functionality for computing cliques and
independent sets of a digraph, whether maximal or not. The fundamental
algorithm used in most of the methods in &Digraphs; to calculate cliques
and independent sets is a version of the Bron-Kerbosch algorithm.
Calculating the cliques and independent sets of a digraph is a well-known
and hard problem, and searching for cliques or independent sets in a
digraph can be very lengthy, even for a digraph with a small number of
vertices. &Digraphs; uses several strategies to increase the performance
of these calculations.
<P/>
From the definition of cliques and independent sets, it follows that the
presence of loops and multiple edges in a digraph is irrelevant to the
existence of cliques and independent sets in the digraph. See <Ref
Prop="DigraphHasLoops"/> and <Ref Prop="IsMultiDigraph"/> for more
information about these properties. Therefore given a digraph
<A>digraph</A>, the cliques and independent sets of <A>digraph</A> are
equal to the cliques and independent sets of the digraph:
See <Ref Oper="DigraphRemoveLoops"/> and <Ref
Oper="DigraphRemoveAllMultipleEdges"/> for more information about these
attributes. Furthermore, the cliques of this digraph are equal to the
cliques of the digraph formed by removing any edge <C>[u,v]</C> for which
the corresponding reverse edge <C>[v,u]</C> is not present. Therefore,
overall, the cliques of <A>digraph</A> are equal to the cliques of the
symmetric digraph:
See <Ref Oper="MaximalSymmetricSubdigraphWithoutLoops"/> for more
information about this. The <Ref Attr="AutomorphismGroup"
Label="for a digraph" /> of this symmetric digraph contains the
automorphism group of <A>digraph</A> as a subgroup. By performing the
search for maximal cliques with the help of this larger automorphism group
to reduce the search space, the computation time may be reduced. The
functions and attributes which return representatives of cliques of
<A>digraph</A> will return orbit representatives of cliques under the
action of the automorphism group of the <E>maximal symmetric subdigraph
without loops</E> on sets of vertices.<P/>
The independent sets of a digraph are equal to the independent sets of the
<Ref Oper="DigraphSymmetricClosure"/>. Therefore, overall, the independent
sets of <A>digraph</A> are equal to the independent sets of the symmetric
digraph:
Again, the automorphism group of this symmetric digraph contains the
automorphism group of <A>digraph</A>. Since a search for independent sets
is equal to a search for cliques in the <Ref Oper="DigraphDual"/>, the
methods used in &Digraphs; usually transform a search for independent sets
into a search for cliques, as described above. The functions and
attributes which return representatives of independent sets of
<A>digraph</A> will return orbit representatives of independent sets under
the action of the automorphism group of the <E>symmetric closure</E> of the
digraph formed by removing loops and multiple edges.<P/>
Please note that in &Digraphs;, cliques and independent sets are not
required to be maximal. Some authors use the word clique to mean
<E>maximal</E> clique, and some authors use the phrase independent set to
mean <E>maximal</E> independent set, but please note that &Digraphs; does
not use this definition.
<P/>
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.