
108
 Chapter
 3.
 Krohn-Rhodes
 Theory
 and
 Complete
 Classes
the
 cascade product, then
 it
 should
 be a
 minimal homomorphically complete class under
the
 cascade product too. Thus,
 it is
 enough
 to
 prove that every automaton
 A = (A, X,
 <5)
can be
 represented
 homomorphically
 by a
 cascade product
 of
 factors
 from
 (JC
 U
 /Ci).
Consider
 for an
 arbitrary automaton
 A the
 automaton
 Ai =
 (Ai,
 Xi, ;) in F
 having
an
 isomorphism onto
 A.
 Using (1),
 we
 have that
 Ai can be
 represented homomorphically
by
 a
 cascade product
 of
 factors
 from
 K U K
1
.
 Therefore
 A
 also
 has
 this property. This ends
the
 proof.
 
We
 obtained
 the
 following direct consequence
 of
 this result.
Corollary 3.46.
 There
 exists
 a
 minimal
 homomorphically
 complete
 class
 under
 the
 cascade
product.
 
Finally,
 we
 note that,
 by
 Lemma 3.41, there exists
 no finite
 homomorphically complete
class
 of finite
 automata under
 the
 cascade product.
3.5
 Bibliographical
 Remarks
Section
 3.1. Theorem
 3.1
 issued
 from
 K. B.
 Krohn
 and J. L.
 Rhodes [1962
 and
 1965].
 A
more detailed description
 of
 this result
 was
 developed
 by K. B.
 Krohn,
 J. L.
 Rhodes,
 and
B.
 R.
 Tilson
 [1968].
 It has a new
 proof
 in
 Esik [1999].
 An
 extension
 of
 Theorem
 3.1 was
given
 by Z.
 Esik [1989a].
 An
 attractive presentation
 of the
 Krohn-Rhodes theory
 was
 given
by
 A.
 Ginzburg
 [1968].
 The
 proof
 of
 Theorem
 3.2 was
 described
 in
 Lallement [1971]
 and
Eilenberg
 [1976].
 Some aspects
 of the
 Krohn-Rhodes theory were studied
 by J. L.
 Rhodes
and
 B. R.
 Tilson [1989],
 B.
 Austin
 et al.
 [1995],
 C. L.
 Nehaniv [1996],
 and Z.
 Esik
 [2000].
Results
 in
 this section (with
 the
 notable exception
 of the
 holonomy decomposition theorem)
are
 mostly originally
 due to and
 derived
 from
 K. B.
 Krohn
 and J. L.
 Rhodes [1962
 and
1965]
 and
 appear
 in the
 book edited
 by
 Arbib
 [1968].
 They have been reformulated
 and
treated
 by
 many authors, e.g.,
 A.
 Ginzburg [1968]
 and S.
 Eilenberg [1976]. Lemmas 3.4,
3.5, 3.6,
 and 3.7 can be
 derived
 from
 the
 results
 of A.
 Ginzburg [1968]. Theorem
 3.8 is due
to H. P.
 Zeiger [1967]. Theorem 3.10 issued
 from
 K. B.
 Krohn
 and J. L.
 Rhodes
 [1962,
1965].
 The
 original idea
 of the
 holonomy decomposition theorem (Theorem 3.9)
 is due to
H. P.
 Zeiger
 [1967],
 with
 a
 correct proof
 for
 partial transformation semigroups given
 by
S.
 Eilenberg
 [1976].
 Our
 proof, which makes reference only
 to
 fully
 defined
 transformation
semigroups,
 is
 inspired
 by
 Eilenberg's.
Section
 3.2. Proposition 3.11
 and
 Corollary 3.12 were discovered
 by Z.
 Esik [1991b].
Theorem 3.13
 can be
 derived
 from
 Z.
 Esik
 [1989a].
 Proposition
 3.14
 was
 found
 by
H. P.
 Zeiger [1967]. Lemma 3.15 appears
 in
 Domosi [1984]. Lemma 3.16
 is an
 extended
form
 of the
 Lemma
 1.3 of F.
 Gecseg's book [1986,
 p.
 25]. Lemma 3.16
 was
 also proved
 by
P.
 Domosi [1984]. Theorem 3.17
 was
 elaborated
 by Z.
 Esik
 and P.
 Domosi [1986]. Theorem
3.18
 was
 shown
 by P.
 Domosi
 and Z.
 Esik [1988a]. Proposition 3.19
 was
 found
 by Z.
 Esik
and
 J.
 Vkagh
 [1986].
 Proposition 3.20
 was
 developed
 by Z.
 Esik
 and P.
 Domosi
 [1986].
Lemmas 3.22, 3.23, 3.24, 3.25,
 and
 3.26 were proved
 by P.
 Domosi
 and Z.
 Esik [1988a].
Lemma
 3.27
 is a new
 result. Theorem 3.28
 is a
 strengthening
 of
 Theorem 3.17
 and can
be
 derived
 from
 Theorem 3.17 using results
 in
 Esik
 and
 Domosi [1986] with results
 from