• Aucun résultat trouvé

The recognition of the class of indecomposable digraphs under low hemimorphy

N/A
N/A
Protected

Academic year: 2021

Partager "The recognition of the class of indecomposable digraphs under low hemimorphy"

Copied!
4
0
0

Texte intégral

(1)

Discrete Mathematics 309 (2009) 3404–3407

Contents lists available atScienceDirect

Discrete Mathematics

journal homepage:www.elsevier.com/locate/disc

Note

The recognition of the class of indecomposable digraphs under low hemimorphy

A. Boussaïri

a

, P. Ille

b,

aFaculté des Sciences Aïn Chock, Département de Mathématiques et Informatique, Km 8 route d’El Jadida, BP 5366 Maarif, Casablanca, Maroc

bInstitut de Mathématiques de Luminy, CNRS – UMR 6206, 163 avenue de Luminy, Case 907, 13288 Marseille Cedex 09, France

a r t i c l e i n f o

Article history:

Received 3 February 2006

Received in revised form 17 April 2008 Accepted 29 August 2008

Available online 24 September 2008 Keywords:

Interval

Indecomposable digraph Hemimorphy

a b s t r a c t

Given a digraphG=(V,A), the subdigraph ofGinduced by a subsetXofVis denoted by G[X]. With each digraphG=(V,A)is associated its dualG?=(V,A?)defined as follows:

for anyx,yV,(x,y)∈A?if(y,x)∈A. Two digraphsGandHare hemimorphic ifGis isomorphic toHor toH?. Givenk >0, the digraphsG = (V,A)andH = (V,B)arek- hemimorphic if for everyXV, with|X| ≤k,G[X]andH[X]are hemimorphic. A classC of digraphs isk-recognizable if every digraphk-hemimorphic to a digraph ofCbelongs to C. In another vein, given a digraphG=(V,A), a subsetXofVis an interval ofGprovided that fora,bXandxVX,(a,x)∈Aif and only if(b,x)∈A, and similarly for(x,a) and(x,b). For example,∅,{x}, wherexV, andVare intervals called trivial. A digraph is indecomposable if all its intervals are trivial. We characterize the indecomposable digraphs which are 3-hemimorphic to a non-indecomposable digraph. It follows that the class of indecomposable digraphs is 4-recognizable.

©2008 Elsevier B.V. All rights reserved.

1. Introduction

Adirected graphor simplydigraph Gconsists of a finite and nonempty setVofverticestogether with a prescribed collection Aof ordered pairs of distinct vertices, called the set of thearcsofG. Such a digraph is denoted by

(

V

,

A

)

. For example, given a setV,

(

V

, ∅ )

is theemptydigraph onVwhereas

(

V

, (

V

×

V

) − { (

x

,

x

) ;

x

V

} )

is thecompletedigraph onV. Given a digraph G

= (

V

,

A

)

, with each nonempty subsetXofVassociate thesubdigraph

(

X

,

A

∩ (

X

×

X

))

ofGinduced byXdenoted byG

[

X

]

. In another respect, given digraphsG

= (

V

,

A

)

andG0

= (

V0

,

A0

)

, a bijectionf fromVontoV0is anisomorphismfromGonto G0provided that for anyx

,

y

V,

(

x

,

y

) ∈

Aif and only if

(

f

(

x

),

f

(

y

)) ∈

A0. Two digraphs are thenisomorphicif there exists an isomorphism from one onto the other. Finally, a digraphH embedsinto a digraphGifHis isomorphic to a subdigraph of G.

With each digraphG

= (

V

,

A

)

associate itsdual G?

= (

V

,

A?

)

and itscomplement G

= (

V

,

A

)

defined as follows. Given x

6=

y

V,

(

x

,

y

) ∈

A?if

(

y

,

x

) ∈

A, and

(

x

,

y

) ∈

Aif

(

x

,

y

) 6∈

A. The digraph

− →

G

= (

V

, − →

A

)

is then defined by

− →

A

=

A

A?. Given digraphsGandH,GandHarehemimorphicifGis isomorphic toHorH?. Given an integerk

>

0, consider digraphs G

= (

V

,

A

)

andH

= (

V

,

B

)

. The digraphsGandHarek-hemimorphicif for every subsetXofV, with

|

X

| ≤

k, the subdigraphs G

[

X

]

andH

[

X

]

are hemimorphic. A digraphGisk-forced(up to duality) ifGandG?are the only digraphsk-hemimorphic to G.

We need some notations. LetG

= (

V

,

A

)

be a digraph. Forx

6=

y

V,x

−→

Gyory

←−

Gxmeans

(

x

,

y

) ∈

Aand

(

y

,

x

) 6∈

A, x

←→

Gymeans

(

x

,

y

), (

y

,

x

) ∈

Aandx

· · ·

Gymeans

(

x

,

y

), (

y

,

x

) 6∈

A. Forx

VandY

V,x

−→

GYsignifies that for every

Corresponding author.

E-mail addresses:aboussairi@hotmail.com(A. Boussaïri),ille@iml.univ-mrs.fr(P. Ille).

0012-365X/$ – see front matter©2008 Elsevier B.V. All rights reserved.

doi:10.1016/j.disc.2008.08.023

(2)

A. Boussaïri, P. Ille / Discrete Mathematics 309 (2009) 3404–3407 3405

y

Y,x

−→

Gy. ForX

,

Y

V,X

−→

GY signifies that for everyx

X,x

−→

GY. Forx

V and forX

,

Y

V,x

←−

GY, x

←→

GY,x

· · ·

GY,X

←→

GYandX

· · ·

GYare defined in the same way. Furthermore, an equivalence relation, denoted by

G, between the ordered pairs of distinct vertices of a digraphG

= (

V

,

A

)

is defined in the following way. Forx

6=

y

V and foru

6= v ∈

V,

(

x

,

y

) ≡

G

(

u

, v)

if the function, which attributesutoxand

v

toy, is an isomorphism fromG

[{

x

,

y

}]

ontoG

[{

u

, v }]

. Equivalently,

(

x

,

y

) ≡

G

(

u

, v)

ifx

−→

Gyandu

−→

G

v

orx

←−

Gyandu

←−

G

v

orx

←→

Gyandu

←→

G

v

or x

· · ·

Gyandu

· · ·

G

v

. The negation is denoted by

(

x

,

y

) 6≡

G

(

u

, v)

. Given a subsetXofV, an elementxofV

Xis aseparator ofXif there existu

, v ∈

Xsuch that

(

x

,

u

) 6≡

G

(

x

, v)

. The set of the separators ofXis denoted bySG

(

X

)

.

A digraphG

= (

V

,

A

)

is aposetprovided that forx

,

y

,

z

V, ifx

−→

Gyandy

−→

Gz, thenx

−→

Gz. With each poset Q

= (

V

,

A

)

associate itscomparability digraph C

(

Q

) = (

V

,

A

A?

)

. Given a digraphG

= (

V

,

A

)

, distinct verticesxandyof Gform adirectedpair if eitherx

−→

Gyory

−→

Gx. A digraph is atournament if all its pairs are directed. For example,

( {

0

,

1

,

2

} , { (

0

,

1

), (

1

,

2

), (

2

,

0

) } )

is a tournament, called 3-cycleand denoted byC3. Atotal order is both a poset and a tournament. Given a total orderO

= (

V

,

A

)

,x

<

ymeansx

−→

Oyforx

,

y

V.

Given a digraphG

= (

V

,

A

)

, a subsetXofVis aninterval[3,7] (or anautonomousset [5,8,9] or aclan[4] or ahomogeneous set [2,6] or amodule[11]) ofGifSG

(

X

) = ∅

. For instance,

,Vand

{

x

}

, wherex

V, are intervals ofGcalledtrivialintervals.

A digraph isindecomposable[3,7,10] (orprime[2] orprimitive[4]) if all its intervals are trivial; otherwise, it isdecomposable.

The indecomposability bears a certain rigidity. The next result illustrates this fact in the case of the posets.

Theorem 1 ([5,9]).Let Q

= (

V

,

A

)

be an indecomposable poset. For every poset Q0

= (

V

,

A0

)

, if C

(

Q0

) =

C

(

Q

)

, then Q0

=

Q or Q0

=

Q?.

Given a posetQ, any digraphG, 3-hemimorphic toQ, is a poset such thatC

(

G

) =

C

(

Q

)

. Therefore, every indecomposable poset is 3-forced. To obtain an analogue ofTheorem 1for the tournaments, the comparability digraph is replaced by theC3- structure. Given a tournamentT

= (

V

,

A

)

, the family of the subsetsXofV, such thatT

[

X

]

is isomorphic toC3, is called the C3-structureofTand denoted byC3

(

T

)

.

Theorem 2 ([1]). Let T

= (

V

,

A

)

be an indecomposable tournament. For every tournament T0

= (

V

,

A0

)

, if C3

(

T0

) =

C3

(

T

)

, then T0

=

T or T0

=

T?.

In other words, every indecomposable tournament is 3-forced. To generalize the two theorems above, we have to disallow the embedding of the following digraph and its dual. The digraph

( {

0

,

1

,

2

} , { (

0

,

2

), (

2

,

0

), (

0

,

1

) } )

is denoted byF. The digraphsFandF?are calledflags. A digraphGis then said to be without flags whenFandF?do not embed intoG.

Theorem 3 ([1]).An indecomposable digraph without flags is3-forced.

The flags are generalized in the following way. Given an integern

4, consider a permutation

σ

of

{

0

, . . . ,

n

2

}

. The digraphFn

(σ )

is defined on

{

0

, . . . ,

n

1

}

in the following manner:

(1) Fn

(σ ) [{

0

, . . . ,

n

2

}]

is the total order

σ(

0

) < · · · < σ (

n

2

)

;

(2) givenm

∈ {

0

, . . . ,

n

2

}

, eithermis even and

(

m

,

n

1

), (

n

1

,

m

)

are arcs ofFn

(σ)

ormis odd and

(

m

,

n

1

), (

n

1

,

m

)

are not.

Givenn

4,Fn

(

Id{0,...,n2}

)

is simply denoted byFn(seeFig. 1). Fork

2, the digraphsF2kandF2k(resp.F2k+1and

(

F2k+1

)

?) are calledgeneralized flags. By definition,F3

(

Id{0,1}

) =

F. We may verify that for a permutation

σ

of

{

0

, . . . ,

n

2

}

, wheren

3,Fn

(σ )

is decomposable if and only if there isi

∈ {

0

, . . . ,

n

3

}

such that

σ (

i

)

and

σ (

i

+

1

)

share the same parity. Therefore, the generalized flags are indecomposable. Furthermore, given an indecomposable digraphG, ifIis an interval of

− →

G, then the digraph obtained fromG, by reversing all the arcs included inI, is 3-hemimorphic toG. Sometimes, intervals are created in this way so that the obtained digraph equals neitherGnorG?. For instance, givenn

4, consider the generalized flagFnand an integeri

>

0 such that 2i

n

2. Clearly,

{

1

, . . . ,

2i

}

is an interval of

− →

Fn. FromFn, we obtain by reversing the arcs contained in

{

1

, . . . ,

2i

}

the digraphFn

i

)

, where

σ

iis the permutation of

{

0

, . . . ,

n

2

}

which interchangesjand 2i

j

+

1 for 1

j

2i. The pair

{

0

,

2i

}

forms an interval ofFn

i

)

. Consequently, the generalized flags are not 3-forced sinceFnandFn

i

)

differ regarding the indecomposability. Incidently, the problem of the recognition of the class of indecomposable digraphs also occurs. Precisely, givenk

>

0, a classCof digraphs isk-recognizableif every digraphk-hemimorphic to a digraph ofCbelongs toCas well. As showing byFnandFn

i

)

, the class of indecomposable digraphs is not 3-recognizable. We reconsider these counter-examples with the following observation:

{

0

, . . . ,

2i

}

is an interval of

− →

Fn and for everyx

∈ {

0

, . . . ,

n

1

} − {

0

,

2i

}

, we have

(

x

,

0

) 6≡

Fn

(

x

,

2i

)

if and only if 0

<

x

<

2i. Generally, consider an indecomposable digraphG

= (

V

,

A

)

. Given vertices

α

and

β

ofGsuch that

α −→

G

β

, the pair

{ α, β }

isweakly separatedif

{ α, β } ∪

SG

( { α, β } )

is an interval of

− →

G and if

α −→

GSG

( { α, β } ) −→

G

β

. The main result consists of the following characterization.

Theorem 4.Let G be an indecomposable digraph. There exists a decomposable digraph3-hemimorphic to G if and only if G admits a weakly separated pair.

As an immediate consequence, we obtain:

Theorem 5.The class of indecomposable digraphs is4-recognizable.

(3)

3406 A. Boussaïri, P. Ille / Discrete Mathematics 309 (2009) 3404–3407

Fig. 1. The generalized flagF2k.

2. The Gallai decomposition theorem

We begin with a well-known property of the intervals. Given a digraphG

= (

V

,

A

)

, ifXandYare disjoint intervals ofG, then

(

x

,

y

) ≡

G

(

x0

,

y0

)

for anyx

,

x0

Xandy

,

y0

Y. This property leads to considerinterval partitionsofG, that is, partitions ofV, all the elements of which are intervals ofG. The elements of such a partitionPbecome the vertices of thequotient G

/

P

= (

P

,

A

/

P

)

ofGbyPdefined as follows: givenX

6=

Y

P,

(

X

,

Y

) ∈

A

/

Pif

(

x

,

y

) ∈

Aforx

Xandy

Y. To state the Gallai decomposition theorem below, we need the following strengthening of the notion of interval. Given a digraph G

= (

V

,

A

)

, a subsetXofVis astrong interval[5,9] ofGprovided thatXis an interval ofGand for each intervalYofG, we have: ifX

Y

6= ∅

, thenX

YorY

X. The family of the maximal strong intervals under inclusion which are distinct fromVis denoted byP

(

G

)

.

Theorem 6 ([5,9]).Given a digraph G

= (

V

,

A

)

, with

|

V

| ≥

2, the family P

(

G

)

constitutes an interval partition of G. Moreover, the corresponding quotient G

/

P

(

G

)

is a complete digraph or an empty digraph or a total order or an indecomposable digraph.

The next result follows fromTheorem 3.

Corollary 7 ([1]).Given digraphs G and H without flags, if G and H are3-hemimorphic, then P

(

G

) =

P

(

H

)

. 3. Proof ofTheorems 4and5

Lemma 8. Consider 3-hemimorphic digraphs G

= (

V

,

A

)

and H

= (

V

,

B

)

. Given an interval I of G such that

|

I

| ≥

2, if

G

[

I

] /

P

( − →

G

[

I

] )

is not a total order, then I is an interval of H.

Proof. Givenx

V

I, sinceIis an interval ofG, we have:x

←→

GIorx

· · ·

GIorx

−→

GIorx

←−

GI. In the first two instances, it follows from the 2-hemimorphy thatx

←→

HIorx

· · ·

HI. In the last two ones, since

− →

G

[

I

] /

P

( − →

G

[

I

] )

is not a total order,P

( − →

G

[

I

∪ {

x

}] ) = {

I

, {

x

}}

. As

− →

G and

− →

H are 3-hemimorphic digraphs without flags, it follows fromCorollary 7 thatP

( − →

H

[

I

∪ {

x

}] ) = {

I

, {

x

}}

. Consequently, eitherx

−→

HIorx

←−

HI.

Corollary 9. Consider 3-hemimorphic digraphs G and H. If G is indecomposable, then for every interval I of H, H

[

I

]

is a total order.

Proof. Consider an intervalIofH. By the previous lemma,

− →

H

[

I

] /

P

( − →

H

[

I

] )

is a total order. We denote the elements of P

( − →

H

[

I

] )

byX1

, . . . ,

Xqin such a way that

− →

H

[

I

] /

P

( − →

H

[

I

] )

is the total orderX1

< · · · <

Xq. For a contradiction, suppose that there isi

∈ {

1

, . . . ,

q

}

such that

|

Xi

| ≥

2. SinceIis an interval ofH,Xiis also. It follows from the preceding lemma that

H

[

Xi

] /

P

( − →

H

[

Xi

] )

is a total order as well. By interchangingHandH?, we can assume thati

<

q. By denoting byYthe largest element of

− →

H

[

Xi

] /

P

( − →

H

[

Xi

] )

, we obtain thatY

Xi+1would be an interval of

− →

H

[

I

]

, which contradicts the fact thatXiis a strong interval of

− →

H

[

I

]

. Consequently, for eachi

∈ {

1

, . . . ,

q

}

,

|

Xi

| =

1, that is,H

[

I

]

is a total order.

Theorem 10. Consider3-hemimorphic digraphs G

= (

V

,

A

)

and H

= (

V

,

B

)

. If G is indecomposable and if H is decomposable, then there exist

α 6= β ∈

V such that

{ α, β }

is an interval of H which is weakly separated in G.

Proof. Given a non-trivial intervalIofH, by the preceding corollary,H

[

I

]

is a total order. Denote by

α

and

β

the first two elements of this total order, with

α −→

G

β

. Clearly,

{ α, β }

is an interval ofH. Consider the smallest interval

− →

J of

− →

G containing

α

and

β

. We useTheorem 6. Firstly, suppose that

− →

G

[ − →

J

] /

P

( − →

G

[ − →

J

] )

is empty. Since

{ α, β }

is directed, there is an element ofP

( − →

G

[ − →

J

] )

containing

α

and

β

, which contradicts the minimality of

− →

J . Secondly, assume that

− →

G

[ − →

J

] /

P

( − →

G

[ − →

J

] )

is indecomposable. As

− →

G

[ − →

J

]

and

− →

H

[ − →

J

]

are 3-hemimorphic digraphs without flags, it follows from Corollary 7that P

( − →

G

[ − →

J

] ) =

P

( − →

H

[ − →

J

] )

. Since

{ α, β }

is an interval ofH,

{ α, β }

is an interval of

− →

H

[ − →

J

]

. We obtain the same contradiction

(4)

A. Boussaïri, P. Ille / Discrete Mathematics 309 (2009) 3404–3407 3407

because

− →

H

[ − →

J

] /

P

( − →

H

[ − →

J

] )

is indecomposable byTheorem 3. Therefore,

− →

G

[ − →

J

] /

P

( − →

G

[ − →

J

] )

is a total order. We denote the elements ofP

( − →

G

[ − →

J

] )

byX1

, . . . ,

Xqin such a way that the corresponding quotient isX1

< · · · <

Xq. By the minimality of

− →

J ,

α ∈

X1and

β ∈

Xq. As previously noticed,P

( − →

H

[ − →

J

] ) = {

X1

, . . . ,

Xq

}

and hence

− →

H

[ − →

J

] /

P

( − →

H

[ − →

J

] )

is a total order as well.

Since

{ α, β }

is an interval ofH,

{ α, β }

is an interval of

− →

H

[ − →

J

]

. AsX1andXqare strong intervals of

− →

H

[ − →

J

]

,

{ α, β } =

X1

Xq or, equivalently,X1

= { α }

andXq

= { β }

. To conclude, we verify that

{ α, β }

is weakly separated inG. It suffices to show that for everyx

V

− { α, β }

,

(

x

, α) 6≡

G

(

x

, β)

if and only ifx

∈ − →

J

− { α, β }

. Clearly, ifx

∈ − →

J

− { α, β }

, then

α −→

Gx

−→

G

β

and hence

(

x

, α) 6≡

G

(

x

, β)

. Conversely, consider an elementuofV

− − →

J . If

{

u

, α }

is directed, then

(

u

, α) ≡

G

(

u

, β)

because

J is an interval of

− →

G. Otherwise,

(

u

, α) ≡

G

(

u

, β)

because

{ α, β }

is an interval ofH.

The proof of the main result follows.

Proof of Theorem 4. Consider an indecomposable digraphG

= (

V

,

A

)

. If there is a decomposable digraph 3-hemimorphic toG, then, byTheorem 10,Gpossesses a weakly separated pair. Conversely, consider

α 6= β ∈

Vsuch that

{ α, β }

is a weakly separated pair ofG. Since

{ α, β } ∪

SG

( { α, β } )

is an interval of

− →

G,

{ β } ∪

SG

( { α, β } )

is also. Consequently, by reversing all the arcs contained in

{ β } ∪

SG

( { α, β } )

, we obtain a digraphHwhich is 3-hemimorphic toG. The pair

{ α, β }

is then an interval ofHand thusHis decomposable.

The next result follows fromTheorem 10.

Corollary 11. Consider 3-hemimorphic digraphs G

= (

V

,

A

)

and H

= (

V

,

B

)

such that G is indecomposable and H is decomposable. There exists a subset X of V , with

|

X

| =

4, such that G

[

X

]

is indecomposable and H

[

X

]

is decomposable. More precisely, G

[

X

]

is isomorphic to F4(resp. F4) and H

[

X

]

is isomorphic to F4

(σ )

(resp. F4

(σ)

), where

σ

is the permutation of

{

0

,

1

,

2

}

which interchanges either0and1or1and2.

Proof. ByTheorem 10, there are

α 6= β ∈

V such that

{ α, β }

is an interval ofHwhich is weakly separated inG. If

{ α, β } ∪

SG

( { α, β } ) =

V, then

{ β } ∪

SG

( { α, β } )

would be an interval ofG. Consequently,

{ α, β } ∪

SG

( { α, β } ) 6=

V and hence

{ α, β } ∪

SG

( { α, β } )

is an interval of

− →

G and not ofG. Therefore, there exists

SG

( { α, β } )

andu

6∈ { α, β } ∪

SG

( { α, β } )

, such that

{ α, β,

s

}

is an interval of

− →

G

[{ α, β,

s

,

u

}]

and not ofG

[{ α, β,

s

,

u

}]

. It follows that

{ α,

u

}

,

{

s

,

u

}

and

{ β,

u

}

are not directed. For example, assume that

α ←→

Gu. Sinceu

6∈

SG

( { α, β } )

,

β ←→

Guand, necessarily,s

· · ·

Gu. Furthermore, G

[{ α, β,

s

}]

is the total order

α <

s

< β

or

β <

s

< α

becauses

SG

( { α, β } )

. In both cases,G

[{ α, β,

s

,

u

}]

is isomorphic toF4. AsGandHare 3-hemimorphic, we have

α ←→

Hu,s

· · ·

Hu,

β ←→

HuandH

[{ α, β,

s

}]

is a total order. To end, it is sufficient to recall that

{ α, β }

is an intervalH

[{ α, β,

s

}]

.

Theorem 5is directly deduced. Finally,Corollary 11leads to the following.

Remark 12. To obtainTheorem 5, it is not necessary to assume that the considered digraphsG

= (

V

,

A

)

andH

= (

V

,

B

)

to be 4-hemimorphic. It suffices to require thatGandHare 2-hemimorphic and that for every subsetXofV, with

|

X

| =

3 or 4, the subdigraphsG

[

X

]

andH

[

X

]

are both indecomposable or not.

Acknowledgement

The authors were supported by the France-Morocco cooperation CNRS/CNRST 2005.

References

[1] A. Boussaïri, P. Ille, G. Lopez, S. Thomassé, TheC3-structure of the tournaments, Discrete Math. 277 (2004) 29–43.

[2] A. Cournier, M. Habib, An efficient algorithm to recognize prime undirected graphs, in: Graph-theoretic Concepts in Computer Science, Proceedings of the 18th Internat. Workshop, WG’92, Wiesbaden-Naurod, Germany, June 1992, in: E.W. Mayr (Ed.), Lecture Notes in Computer Science, vol. 657, Springer, Berlin, 1993, pp. 212–224.

[3] A. Cournier, P. Ille, Minimal indecomposable graphs, Discrete Math. 183 (1998) 61–80.

[4] A. Ehrenfeucht, G. Rozenberg, Primitivity is hereditary for 2-structures, fundamental study, Theoret. Comput. Sci. 3 (70) (1990) 343–358.

[5] T. Gallai, Transitiv orientierbare Graphen, Acta Math. Acad. Sci. Hungar. 18 (1967) 25–66.

[6] M. Habib, Substitution des structures combinatoires, théorie et algorithmes, Ph.D. Thesis, Université Pierre et Marie Curie, Paris VI, 1981.

[7] P. Ille, Indecomposable graphs, Discrete Math. 173 (1997) 71–78.

[8] D. Kelly, Comparability graphs, in: I. Rival (Ed.), Graphs and Orders, Reidel, Drodrecht, 1985, pp. 3–40.

[9] F. Maffray, M. Preissmann, A translation of Tibor Gallai’s paper: Transitiv orientierbare Graphen, in: J.L. Ramirez-Alfonsin, B.A. Reed (Eds.), Perfect Graphs, Wiley, New York, 2001, pp. 25–66.

[10] J.H. Schmerl, W.T. Trotter, Critically indecomposable partially ordered sets, graphs, tournaments and other binary relational structures, Discrete Math.

113 (1993) 191–205.

[11] J. Spinrad, P4-trees and substitution decomposition, Discrete Appl. Math. 39 (1992) 263–291.

Références

Documents relatifs

We deduce that every commutative ring R with a Hausdorff and totally disconnected maximum prime spectrum is local-global, and moreover, R is an elementary divisor ring if, in

Since X is defined over a finitely generated field, motivic cohomology commutes with filtering inverse limits of smooth schemes (with affine transition morphisms) and l-adic

if it is irreducible in k[t, x], theorem 1.2 has the following consequence which makes it easy to produce indecomposable polynomials in one variable..

In order to maintain the direction convention of upward arcs in this Hasse diagram, each new concept must be assigned to a layer which is separated from the infimum [supremum] by

Comme le montre le résultat suivant, le Théorème 1.2 fournit une minoration optimale du nombre des sommets d’un graphe 2-reconstructible indécomposable ayant r paires orientées

4.1. HILBERTS THEOREM 90 FOR RELATIVE K^. — We now follow the proof of Suslin in [S] to prove Hilbert's Theorem 90 for relative K^. Let S be a semi-local PIR containing the field ko.

des bases compact-ouvert d’espaces de Stone que lorsqu’on utilise la definition de Frink comme il l’avait fait dans un article precedent.. Il obtient aussi une

if it is irreducible in k[t, x], theorem 1.2 has the following consequence which makes it easy to produce indecomposable polynomials in one variable..