Namespaces
Variants
Actions

Difference between revisions of "Rees semi-group of matrix type"

From Encyclopedia of Mathematics
Jump to: navigation, search
(Importing text file)
 
m (tex encoded by computer)
 
Line 1: Line 1:
A semi-group theoretical construction defined as follows. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804701.png" /> be an arbitrary [[Semi-group|semi-group]], let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804702.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804703.png" /> be (index) sets and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804704.png" /> be a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804705.png" />-matrix over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804706.png" />, i.e. a mapping from the Cartesian product <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804707.png" /> into <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804708.png" />. The following formula defines an operation on the set <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r0804709.png" />:
+
<!--
 +
r0804701.png
 +
$#A+1 = 66 n = 0
 +
$#C+1 = 66 : ~/encyclopedia/old_files/data/R080/R.0800470 Rees semi\AAhgroup of matrix type
 +
Automatically converted into TeX, above some diagnostics.
 +
Please remove this comment and the {{TEX|auto}} line below,
 +
if TeX found to be correct.
 +
-->
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047010.png" /></td> </tr></table>
+
{{TEX|auto}}
 +
{{TEX|done}}
  
Then <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047011.png" /> is a semi-group, called a Rees semi-group of matrix type over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047012.png" /> and denoted by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047013.png" />; the matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047014.png" /> is called the sandwich matrix of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047015.png" />. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047016.png" /> is a semi-group with zero 0, then <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047017.png" /> is an ideal in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047018.png" /> and the Rees quotient semi-group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047019.png" /> (see [[Semi-group|Semi-group]]) is denoted by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047020.png" />; in the case when <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047021.png" /> is a group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047022.png" /> with an adjoined zero, instead of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047023.png" /> one writes <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047024.png" /> and calls it a Rees semi-group of matrix type over the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047025.png" /> with an adjoined zero. The group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047026.png" /> is called the structure group for the semi-groups <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047027.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047028.png" />.
+
A semi-group theoretical construction defined as follows. Let  $  S $
 +
be an arbitrary [[Semi-group|semi-group]], let  $  I $
 +
and  $  \Lambda $
 +
be (index) sets and let  $  P = ( p _ {\lambda i }  ) $
 +
be a $  ( \Lambda \times I) $-
 +
matrix over  $  S $,  
 +
i.e. a mapping from the Cartesian product  $  \Lambda \times I $
 +
into  $  S $.  
 +
The following formula defines an operation on the set  $  M = I \times S \times \Lambda $:
  
Another representation of the Rees semi-group of matrix type over a semi-group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047029.png" /> with zero and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047030.png" />-sandwich matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047031.png" /> is realized in the following way. An <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047032.png" />-matrix over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047033.png" /> is called a Rees matrix if it does not contain more than one non-zero element. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047034.png" /> be the Rees matrix over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047035.png" /> that has <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047036.png" /> in the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047037.png" />-th row and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047038.png" />-th column, and zeros in all other places. On the set of all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047039.png" />-Rees matrices over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047040.png" /> one can define an operation
+
$$
 +
( i, s, \lambda ) ( j, t, \mu )  = ( i, sp _ {\lambda j }  t, \mu ).
 +
$$
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047041.png" /></td> <td valign="top" style="width:5%;text-align:right;">(1)</td></tr></table>
+
Then  $  M $
 +
is a semi-group, called a Rees semi-group of matrix type over  $  S $
 +
and denoted by  $  {\mathcal M} ( S;  I, \Lambda ;  P) $;
 +
the matrix  $  P $
 +
is called the sandwich matrix of  $  {\mathcal M} ( S;  I, \Lambda ;  P) $.
 +
If  $  S $
 +
is a semi-group with zero 0, then  $  Z = \{ {( i, 0, \lambda ) } : {i \in I,  \lambda \in \Lambda } \} $
 +
is an ideal in  $  M = {\mathcal M}( S; I, \Lambda ; P) $
 +
and the Rees quotient semi-group  $  M/Z $(
 +
see [[Semi-group|Semi-group]]) is denoted by  $  {\mathcal M}  ^ {0} ( S;  I, \Lambda ;  P) $;  
 +
in the case when  $  S = G  ^ {0} $
 +
is a group  $  G  ^ {0} $
 +
with an adjoined zero, instead of  $  {\mathcal M}  ^ {0} ( G  ^ {0} ;  I, \Lambda ;  P) $
 +
one writes  $  {\mathcal M}  ^ {0} ( G;  I, \Lambda ;  P) $
 +
and calls it a Rees semi-group of matrix type over the group  $  G  ^ {0} $
 +
with an adjoined zero. The group  $  G $
 +
is called the structure group for the semi-groups  $  {\mathcal M} ( G; I, \Lambda ; P) $
 +
and  $  {\mathcal M}  ^ {0} ( G;  I, \Lambda ;  P) $.
  
where on the right-hand side is the  "ordinary"  matrix product. This set becomes a semi-group with respect to this operation. The mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047042.png" /> is an isomorphism between this semi-group and the semi-group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047043.png" />; the notation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047044.png" /> is used for both of these semi-groups. Formula (1) provides an explanation of the term  "sandwich matrix"  for <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047045.png" />. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047046.png" /> is a group, then the semi-group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047047.png" /> is regular if and only if each row and each column of the matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047048.png" /> contains a non-zero element; any semi-group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047049.png" /> is completely simple (cf. [[Completely-simple semi-group|Completely-simple semi-group]]), any [[Regular semi-group|regular semi-group]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047050.png" /> is completely <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047051.png" />-simple. The converse of the last two statements gives the main content of Rees's theorem [[#References|[1]]]: Any completely-simple (completely <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047052.png" />-simple) semi-group can be isomorphically represented as a Rees semi-group of matrix type over a group (as a regular Rees semi-group of matrix type over a group with an adjoined zero). If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047053.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047054.png" /> are isomorphic, then the groups <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047055.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047056.png" /> are isomorphic, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047057.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047058.png" /> have the same cardinality, and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047059.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047060.png" /> have the same cardinality. Necessary and sufficient conditions for isomorphy of the semi-groups <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047061.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047062.png" /> are known, and together with the just-mentioned conditions they include a quite definite relation between the sandwich matrices <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047063.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047064.png" /> (see [[#References|[1]]]–[[#References|[3]]]). In particular, any completely <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047065.png" />-simple semi-group can be isomorphically represented as a Rees semi-group of matrix type in whose sandwich matrix <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/r/r080/r080470/r08047066.png" /> each element in a given row and a given column is either 0 or the identity element of the structure group; such a sandwich matrix is called normalized. Similar properties are valid for completely-simple semi-groups.
+
Another representation of the Rees semi-group of matrix type over a semi-group  $  S $
 +
with zero and  $  ( \Lambda \times I) $-
 +
sandwich matrix  $  P $
 +
is realized in the following way. An  $  ( I \times \Lambda ) $-
 +
matrix over  $  S $
 +
is called a Rees matrix if it does not contain more than one non-zero element. Let  $  \| a \| _ {i \lambda }  $
 +
be the Rees matrix over  $  S $
 +
that has  $  a $
 +
in the  $  i $-
 +
th row and  $  \lambda $-
 +
th column, and zeros in all other places. On the set of all  $  ( I \times \Lambda ) $-
 +
Rees matrices over  $  S $
 +
one can define an operation
 +
 
 +
$$ \tag{1 }
 +
A \circ B  =  APB,
 +
$$
 +
 
 +
where on the right-hand side is the  "ordinary"  matrix product. This set becomes a semi-group with respect to this operation. The mapping $  \| a \| _ {i \lambda }  \mapsto ( i, a, \lambda ) $
 +
is an isomorphism between this semi-group and the semi-group $  {\mathcal M}  ^ {0} ( S;  I, \Lambda ;  P) $;  
 +
the notation $  {\mathcal M}  ^ {0} ( S;  I, \Lambda ;  P) $
 +
is used for both of these semi-groups. Formula (1) provides an explanation of the term  "sandwich matrix"  for $  P $.  
 +
If $  G $
 +
is a group, then the semi-group $  {\mathcal M}  ^ {0} ( G;  I, \Lambda ;  P) $
 +
is regular if and only if each row and each column of the matrix $  P $
 +
contains a non-zero element; any semi-group $  {\mathcal M} ( G;  I, \Lambda ;  P) $
 +
is completely simple (cf. [[Completely-simple semi-group|Completely-simple semi-group]]), any [[Regular semi-group|regular semi-group]] $  {\mathcal M}  ^ {0} ( G;  I, \Lambda ;  P) $
 +
is completely 0 $-
 +
simple. The converse of the last two statements gives the main content of Rees's theorem [[#References|[1]]]: Any completely-simple (completely 0 $-
 +
simple) semi-group can be isomorphically represented as a Rees semi-group of matrix type over a group (as a regular Rees semi-group of matrix type over a group with an adjoined zero). If $  {\mathcal M}  ^ {0} ( G;  I, \Lambda ;  P) $
 +
and $  {\mathcal M}  ^ {0} ( G  ^  \prime  ;  I  ^  \prime  , \Lambda  ^  \prime  ;  P  ^  \prime  ) $
 +
are isomorphic, then the groups $  G $
 +
and $  G  ^  \prime  $
 +
are isomorphic, $  I $
 +
and $  I  ^  \prime  $
 +
have the same cardinality, and $  \Lambda $
 +
and $  \Lambda  ^  \prime  $
 +
have the same cardinality. Necessary and sufficient conditions for isomorphy of the semi-groups $  {\mathcal M}  ^ {0} ( G;  I, \Lambda ;  P) $
 +
and $  {\mathcal M}  ^ {0} ( G  ^  \prime  ;  I  ^  \prime  , \Lambda  ^  \prime  ;  P  ^  \prime  ) $
 +
are known, and together with the just-mentioned conditions they include a quite definite relation between the sandwich matrices $  P $
 +
and $  P  ^  \prime  $(
 +
see [[#References|[1]]]–[[#References|[3]]]). In particular, any completely 0 $-
 +
simple semi-group can be isomorphically represented as a Rees semi-group of matrix type in whose sandwich matrix $  P $
 +
each element in a given row and a given column is either 0 or the identity element of the structure group; such a sandwich matrix is called normalized. Similar properties are valid for completely-simple semi-groups.
  
 
====References====
 
====References====
 
<table><TR><TD valign="top">[1]</TD> <TD valign="top">  D. Rees,  "On semi-groups"  ''Proc. Cambridge Philos. Soc.'' , '''36'''  (1940)  pp. 387–400</TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top">  A.H. Clifford,  G.B. Preston,  "Algebraic theory of semi-groups" , '''1–2''' , Amer. Math. Soc.  (1961–1967)</TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top">  E.S. Lyapin,  "Semigroups" , Amer. Math. Soc.  (1974)  (Translated from Russian)</TD></TR></table>
 
<table><TR><TD valign="top">[1]</TD> <TD valign="top">  D. Rees,  "On semi-groups"  ''Proc. Cambridge Philos. Soc.'' , '''36'''  (1940)  pp. 387–400</TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top">  A.H. Clifford,  G.B. Preston,  "Algebraic theory of semi-groups" , '''1–2''' , Amer. Math. Soc.  (1961–1967)</TD></TR><TR><TD valign="top">[3]</TD> <TD valign="top">  E.S. Lyapin,  "Semigroups" , Amer. Math. Soc.  (1974)  (Translated from Russian)</TD></TR></table>

Latest revision as of 08:10, 6 June 2020


A semi-group theoretical construction defined as follows. Let $ S $ be an arbitrary semi-group, let $ I $ and $ \Lambda $ be (index) sets and let $ P = ( p _ {\lambda i } ) $ be a $ ( \Lambda \times I) $- matrix over $ S $, i.e. a mapping from the Cartesian product $ \Lambda \times I $ into $ S $. The following formula defines an operation on the set $ M = I \times S \times \Lambda $:

$$ ( i, s, \lambda ) ( j, t, \mu ) = ( i, sp _ {\lambda j } t, \mu ). $$

Then $ M $ is a semi-group, called a Rees semi-group of matrix type over $ S $ and denoted by $ {\mathcal M} ( S; I, \Lambda ; P) $; the matrix $ P $ is called the sandwich matrix of $ {\mathcal M} ( S; I, \Lambda ; P) $. If $ S $ is a semi-group with zero 0, then $ Z = \{ {( i, 0, \lambda ) } : {i \in I, \lambda \in \Lambda } \} $ is an ideal in $ M = {\mathcal M}( S; I, \Lambda ; P) $ and the Rees quotient semi-group $ M/Z $( see Semi-group) is denoted by $ {\mathcal M} ^ {0} ( S; I, \Lambda ; P) $; in the case when $ S = G ^ {0} $ is a group $ G ^ {0} $ with an adjoined zero, instead of $ {\mathcal M} ^ {0} ( G ^ {0} ; I, \Lambda ; P) $ one writes $ {\mathcal M} ^ {0} ( G; I, \Lambda ; P) $ and calls it a Rees semi-group of matrix type over the group $ G ^ {0} $ with an adjoined zero. The group $ G $ is called the structure group for the semi-groups $ {\mathcal M} ( G; I, \Lambda ; P) $ and $ {\mathcal M} ^ {0} ( G; I, \Lambda ; P) $.

Another representation of the Rees semi-group of matrix type over a semi-group $ S $ with zero and $ ( \Lambda \times I) $- sandwich matrix $ P $ is realized in the following way. An $ ( I \times \Lambda ) $- matrix over $ S $ is called a Rees matrix if it does not contain more than one non-zero element. Let $ \| a \| _ {i \lambda } $ be the Rees matrix over $ S $ that has $ a $ in the $ i $- th row and $ \lambda $- th column, and zeros in all other places. On the set of all $ ( I \times \Lambda ) $- Rees matrices over $ S $ one can define an operation

$$ \tag{1 } A \circ B = APB, $$

where on the right-hand side is the "ordinary" matrix product. This set becomes a semi-group with respect to this operation. The mapping $ \| a \| _ {i \lambda } \mapsto ( i, a, \lambda ) $ is an isomorphism between this semi-group and the semi-group $ {\mathcal M} ^ {0} ( S; I, \Lambda ; P) $; the notation $ {\mathcal M} ^ {0} ( S; I, \Lambda ; P) $ is used for both of these semi-groups. Formula (1) provides an explanation of the term "sandwich matrix" for $ P $. If $ G $ is a group, then the semi-group $ {\mathcal M} ^ {0} ( G; I, \Lambda ; P) $ is regular if and only if each row and each column of the matrix $ P $ contains a non-zero element; any semi-group $ {\mathcal M} ( G; I, \Lambda ; P) $ is completely simple (cf. Completely-simple semi-group), any regular semi-group $ {\mathcal M} ^ {0} ( G; I, \Lambda ; P) $ is completely $ 0 $- simple. The converse of the last two statements gives the main content of Rees's theorem [1]: Any completely-simple (completely $ 0 $- simple) semi-group can be isomorphically represented as a Rees semi-group of matrix type over a group (as a regular Rees semi-group of matrix type over a group with an adjoined zero). If $ {\mathcal M} ^ {0} ( G; I, \Lambda ; P) $ and $ {\mathcal M} ^ {0} ( G ^ \prime ; I ^ \prime , \Lambda ^ \prime ; P ^ \prime ) $ are isomorphic, then the groups $ G $ and $ G ^ \prime $ are isomorphic, $ I $ and $ I ^ \prime $ have the same cardinality, and $ \Lambda $ and $ \Lambda ^ \prime $ have the same cardinality. Necessary and sufficient conditions for isomorphy of the semi-groups $ {\mathcal M} ^ {0} ( G; I, \Lambda ; P) $ and $ {\mathcal M} ^ {0} ( G ^ \prime ; I ^ \prime , \Lambda ^ \prime ; P ^ \prime ) $ are known, and together with the just-mentioned conditions they include a quite definite relation between the sandwich matrices $ P $ and $ P ^ \prime $( see [1][3]). In particular, any completely $ 0 $- simple semi-group can be isomorphically represented as a Rees semi-group of matrix type in whose sandwich matrix $ P $ each element in a given row and a given column is either 0 or the identity element of the structure group; such a sandwich matrix is called normalized. Similar properties are valid for completely-simple semi-groups.

References

[1] D. Rees, "On semi-groups" Proc. Cambridge Philos. Soc. , 36 (1940) pp. 387–400
[2] A.H. Clifford, G.B. Preston, "Algebraic theory of semi-groups" , 1–2 , Amer. Math. Soc. (1961–1967)
[3] E.S. Lyapin, "Semigroups" , Amer. Math. Soc. (1974) (Translated from Russian)
How to Cite This Entry:
Rees semi-group of matrix type. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Rees_semi-group_of_matrix_type&oldid=12298
This article was adapted from an original article by L.N. Shevrin (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article