Krotov D.S.
Potapov V.N.
Sobolev Institute of Mathematics, Novosibirsk
Sobolev Institute of Mathematics, Novosibirsk
Let be an arbitrary natural number, be a finite 2 set of cardinality , and . Let be a mapping from to . If holds for every that differ in exactly one coordinate then the triple is called partial -quasigroup (of order ). If in addition then the pair is called -quasigroup (of order ).
It is naturally to think of a -quasigroup as a value array of . In such -dimensional array of size every row (column) of every direction contains all elements of (see, for example, Figure 2).
Two -quasigroups and are called equivalent if there exist permutations and such that .
The -quasigroup is called extension of the partial -quasigroup if . Note that a partial -quasigroup may have more than one extension, as well as it may have no extensions at all.
There are exactly two -quasigroups of order . All -quasigroups of order are equivalent, and their number is . The order is the first nontrivial from the view-point of variability case. Let and be the number of different -quasigroups . It was shown in [1] that
The upper estimations (2)-(6) on are based on the number of extensions of special-purpose partial -quasigroups.
The mapping is uniquely defined by two subfunctions
and
.
The pairs and are -quasigroups.
Therefore Lemma 2 implies
For -ary mapping we call -ary minorant the mapping obtained by the substituting zeros for some arguments of . Lemma 3 implies that -ary minorant of can be reconstructed from -ary minorants by at most four ways, provided is a -quasigroup and .
Let
, where
(see, f.i., Fig.1d).
Let be -quasigroup.
The partial -quasigroup ,
where
,
specifies all the -ary
minorants of . The number of minorants is , of which
are - and -ary. Therefore, assuming is specified,
the number of ways for to be reconstructed is at most .
The number of ways to specify is , therefore
Let the mapping
be defined by the value table
Call the -quasigroup bicubical if it is equivalent to the -quasigroup with .
Call -quasigroup cubical if it is equivalent to some -quasigroup such that . Figure 2 gives examples of bicubical, cubical, and non-cubical -quasigroups.
Lemma 4 allows to strengthen the previous bound:
Let be the set of all decomposed -quasigroups , and be the set of -quasigroups which are not cubical.
By the inequality (5) there is no difficulty in proving that
The folowing statement is established by the total computer enumeration of -quasigroups or order .
Let , and are defined as in Section 2.
Therefore,
.
Moreover, since
for every
, it is true that
, and
if . Taking Proposition 3 as a base, it's now easy to prove
by induction that
provided .
Since
by Proposition 1, it holds
Comments |
[Home] [Conference] |
©2001, Siberian Branch of Russian Academy of Science, Novosibirsk
©2001, United Institute of Computer Science SB RAS, Novosibirsk
©2001, Institute of Computational Techologies SB RAS, Novosibirsk
©2001, A.P. Ershov Institute of Informatics Systems SB RAS, Novosibirsk
©2001, Institute of Mathematics SB RAS, Novosibirsk
©2001, Institute of Cytology and Genetics SB RAS, Novosibirsk
©2001, Institute of Computational Mathematics and Mathematical Geophysics SB RAS, Novosibirsk
©2001, Novosibirsk State University
Last modified Saturday, 06-Oct-2001 18:20:30 NOVST