• Aucun résultat trouvé

Regular sets over extended tree structures

N/A
N/A
Protected

Academic year: 2021

Partager "Regular sets over extended tree structures"

Copied!
41
0
0

Texte intégral

(1)

HAL Id: hal-00379256

https://hal.archives-ouvertes.fr/hal-00379256

Preprint submitted on 28 Apr 2009

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires

Regular sets over extended tree structures

Severine Fratani

To cite this version:

Severine Fratani. Regular sets over extended tree structures. 2009. �hal-00379256�

(2)

SéverineFratani

Laboratoired'InformatiqueFondamentaledeMarseille(LIF)CNRS:UMR6166

UniversitédelaMéditerranée-Aix-MarseilleII

Universitéde Provene-Aix-MarseilleI

Abstrat

WeinvestigatenotionsofdeidabilityanddenabilityfortheMonadiSeond-

Order Logi of labeled tree strutures, and links with nite automata using

oralestotestinputprexes.

A generalframework is dened allowing to transfersome MSO-properties

from agraph-strutureto a labeledtree struture. Transferredpropertiesare

deidabilityofsentenesandexisteneofadenablemodelforeverysatisable

formula. Alassofniteautomatawithprex-oralesisalsodened,reogniz-

ingexatlylanguagesdened byMSO-formulasinanylabeledtree-struture.

Applying these results, the well-known equality between languages reog-

nized by nite automata, sets of verties MSO denable in a tree-struture

andsetsofpushdownontextsgeneratedbypushdown-automataisextendedto

k

-iteratedpushdownautomata.

Keywords: Labeledtreestrutures; MSOdenablesets; Automatawith

orale;Iteratedpushdownstrutures.

Introdution

InitiatedbytheworkofBühionwords,thestudyoflinksbetweenautomata

andlogihaspermittoidentifystrutureshavingadeidableMonadiSeond-

Ordertheory. Inpartiular,Rabinprovedin[28℄deidabilityoftheMSO-theory

ofinnitetreestruturesinwhihnumerouspropertiesaredenableandtheories

areinterpretable. Theseworkshavealsoledtoalogiharaterisationofregular

languages: languagesreognisedbyniteautomata areexatlysets denedby

MSO-formulasinatreestruture.

The goal of this paper is to extend these works to the study of labelled

tree strutures: identify labellings for whih tree struture have a deidable

MSO-theory, for whih every formula admits a denable model and give an

automata-haraterizationofthedenablesets.

Toahievethisgoal,weintroduenewinterestingobjetsandresults. First,

wedene alassof word/tree automatawith prex-orales(i.e.,sets ofwords

overthe inputalphabet) used to testthe alreadyproessed prexesofinputs.

Languagesandforestreognizedbyprex-oralesautomataenjoynieproperty,

(3)

ofMSO-formulasoverinnitetreesanbeextendedtotheselanguages: forests

reognized by automata with orales

O 1 , . . . , O m

are forestsMSO-denable in treestruturesextendedbyunaryrelations

O 1 , . . . , O m

. Remarkthat this ap-

proahhasalreadybeendevisedin [32℄to haraterisesomepropersublasses

ofregularlanguagesbyusingregularprex-oralesandtostudytheirdenabil-

ityinFirst-OrderLogioverextendedwordstrutures. However,thedenition

of automata with prex-oralesdoes not expliitlyappear in this paper sine

regularprex-oralesanbesimulatedbythediretprodutofniteautomata.

Seond,weestablishtransfertheorems, allowingfrom astruture,to onstrut

atreestruturehavingsomesimilarMSOproperties. Thisapproahisommon

for the transferof deidability (for example the transfer of deidability from

astruture to its tree-like struture,(see [31℄ or[35℄), or from a graphto it's

unfolding(see[10℄)),buthere,inadditiontodeidability,transferredproperties

alsoapplied to sets MSO denable in suhstrutures and lassesof automata

reognisingthem. Inaddition,ourtransferofdeidabilityallowstoobtainnew

deidabilityresultswhiharenotoverbytheonesitedabove. Propertiesare

transferred to a labelled tree struture from its image struture by any mor-

phism. If

µ : D → D

isamappingand

S

isarelationalstruture over

D

, the

imagestruture

µ( S )

of

S

has

D

asdomainanditsrelationsaretheimagesby

µ

oftherelationsof

S

.

Let

t

be a labelled tree, and

t

be the struture assoiated to

t

. For any

morphismofmonoid

µ

,andundersomesimplehypothesisonthelabellingof

t

,

weobtainthefollowingmainresults:

Transferofdeidability: (Theorem55)if

µ(t)

hasadeidableMSO-theory,

then

t

hasadeidableMSO-theory,

Transfer of the property of Denable Model: (Theorem57) under aon-

dition on

µ

, if

µ(t)

satisesthe propertyof Denable Model (DM),then

t

satises DM. This propertyensures for a struture

S

that any satis-

ableformulaadmitsatleastonemodelwhih isMSO-denablein

S

(see

Denition15),

Theorem of struture: (Theorem 58) under the same ondition on

µ

, if

µ(t)

satises DM, then anyset is MSO-denablein

t

i it is reognised

byanite automatonusing onlyorales oftheform

µ

−1

(D)

where

D

is

MSO-denablein

µ(t)

. (TheneahoraletestsapropertyMSO-denable in

µ(t)

, ontheimageby

µ

ofinputwordprexes).

Applying these results, weobtaintree strutures having adeidable MSO

theoryandlassesoflanguageshavingtwoequivalentharaterizations:aslan-

guagesreognizedbyautomatawithorales,andassetsMSO-denableinsome

labelled tree strutures. We thus extend the two haraterizations of regular

languagesmentionedabove.

But regularlanguagesadmit athird haraterization,assets of pushdown

ontextsgeneratedbyapushdownsystemoftransitions[21℄. Somereentworks

(4)

roughlyastakofstak... ofstak(seeforexamples[5,7,24,19℄),itisthennat-

uralto attemptto deneanotionofregular setsof

k

-pushdowns(i.e.,staks

with

k

level of embedded pushdowns) whih generalize the previousequality.

Wegiveequalities betweenlanguagesof

k

-pushdownsreognized byautomata

withp-orales,languagesMSO-denableinapartiulartreestrutureandsets

of

k

-pushdownontextsgenerated bya

k

-pushdownsystemof transitions. We iteratively use the three transfer theorems on a family of strutures

( P

k

) k≥1

havingaprexwordslanguage

P k

fordomain. Thelanguage

P k

denesanen-

odingoftheset

k

-pdsofall

k

-pushdowns. Thestruture

P

kisMSO-equivalent withthestruture PDS

k

whose domain is

k

-pdsand whoserelations arethose

induedby thelassial instrutionson

k

-pushdowns. This allows to dene a lassoflanguagesin

P k

thatanbeexpressedinfourequivalentways(Theorem

85):

aslanguagesreognisedbyniteautomatawithprex-orale,

aslanguagesdenedbyMSO-formulasinthetreestrutureofdomain

P k

,

asenodingsofsetsdened byMSO-formulasinPDS

k

,

asenodingofsets of

k

-pushdownsgenerated byastore-ontrolled

k

-pds

systemoftransitions.

Weshowin additionthatPDS

k

satisesthepropertyofDenableModel.

Thispaperisorganizedasfollows. Setion 1isdevoted tobasidenitions

onwords,logi,automataand

k

-pushdownsstrutures. Itisalsointroduedthe

notionofwordautomatawithorales. InSetion2,weextendtotreeautomata

the use of orales. The Rabin's orrespondene between regular forests and

modelsofMSO-formulasovertreesisadaptedtotheselanguages.InSetion3,

we developa game-theoretial approah to provethe three transfertheorems.

Wegivealsoasimpleappliation ofthetransfertheorems. Finally,wegivein

Setion4adenitionof

k

-regularsets ofpushdowns.

1. Preliminaries

1.1. Basi denitions

1.1.1. Somenotations andonventions

Givenaset

A

,wedenoteby

|A|

theardinalof

A

. If

s

isamapfromaset

A

, then

s(A) = {s(a) | a ∈ A}

. If

V ~ = (V 1 , . . . , V n )

isavetorofsubsets of

A

then

s( A) = (s(V ~ 1 ), . . . , s(V n ))

. Theharateristifuntion of

V ~

in

A

isamap

χ V A ~ : A → {0, 1} n

denedforall

x ∈ A

,by

χ V A ~ (x) = (b 1 , . . . , b n )

where

∀i

,

b i = 1

i

u ∈ S i

.

(5)

If

A

isaset,

A

denotesthesetofwords(nitesequenes)over

A

,and

ε

the

emptyword. For

u, v ∈ A

,thelengthof

u

isdenoted

|u|

andwewrite

v 4 u

if

v

isprexof

u

. Aset

P ⊆ A

isaprexlosedlanguageif

∀u ∈ P

,

∀v ∈ A

, if

v 4 u

then

v ∈ P

.

1.1.3. Freegroup

Givenanitealphabet

A

,letusassoiatetoeah

a ∈ A

theinversesymbol

¯

a

whihdoesnotbelongto

A

. Wedenote by

A

theset of inverselettersof

A

anddene

A b = A ∪ A

. Forevery

u = a 1 · · · a n ∈ A b

,theinverse wordof

u

is

u = b n · · · b 1

where

∀i ∈ [1, n]

:

if

a i ∈ A

then

b i = ¯ a i

,andif

a i = ¯ a ∈ A

then

b i = a

.

Letusthenonsidertheredutionsystem

S = {(a¯ a, ε), (¯ aa, ε)}

. Awordin

A b

issaidto bereduedifitis

S

-redued,i.e.,itdoesnotontainourrenesof

a¯ a

or

¯ aa

,for

a ∈ A

. Wedenote by

Irr(A)

the setof reduedwordsin

A b

. As

S

isonuent,eahword

w

is equivalent(mod

S

)to auniquereduedword

denoted

ρ(w)

.

Wedenethefreegroup

(Irr(A), ε, •)

, where

∀u, v ∈ Irr(A)

,

u • v = ρ(u · v)

.

1.1.4. Projetions

Foranyintegers

0 < i ≤ j ≤ n

, forany vetorofelements

(a 1 , . . . , a n )

, we

dene the projetion

π i (a 1 , . . . , a n ) = a i

and

π i,...,j (a 1 , . . . , a n ) = (a i , . . . , a j )

.

For any alphabets

B

and

A

with

B ⊆ A

, the projetion

π B : A

→ B

is a

morphismdened

∀a ∈ A

by

π B (a) = a

if

a ∈ B

and

π B (a) = ε

else.

1.1.5. Treesand forests

Given nite alphabets

Σ

and

A

and a prex losed language

P ⊆ A

, a

P

-tree

(Σ)

(treeofdomain

P

labelledby

Σ

)is atotalfuntion

t : P → Σ

. The

setofall

P

-tree

(Σ)

isdenoted

P

-Tree

(Σ)

. Inordertodealwithunlabelledtrees

in an uniform way, we introdue the speial symbol

. Unlabelled trees are

then funtions

t : P → {⊤}

. We will often onsider trees in

P

-Tree

({0, 1} n )

,

for

n ≥ 0

(withthe onvention that

{0, 1} 0 = {⊤}

), we will denote this lass

P

-Tree

n

. Remarkthatatreein

P

-Tree

n

analwaysbeseenastheharateristi funtion

χ S P ~

ofavetor

S ~ = (S 1 , . . . , S n )

,for

S i ⊆ P

.

Wewillusetwokindsofoperationsontreesandtree-languages:

Restrition: let

t ∈ A

-Tree

(Σ)

,

t

|P isthe

P

-tree

(Σ)

obtainedbyrestrit-

ingthedomainof

t

to

P

. If

F ⊆ A

-Tree

(Σ)

,then

F

|P

= {t

|P

, t ∈ F }

.

Produt: let

t 1

be a

P

-tree

(Σ 1 )

and

t 2

a

P

-tree

(Σ 2 )

, the produt of

t 1

and

t 2

is thetree

t 1 b t 2 ∈ P

-Tree

(Σ 1 × Σ 2 )

fullling

∀u ∈ P

,

t 1 b t 2 (u) = (t 1 (u), t 2 (u))

. Thisdenitionanbeextendedto treelanguages:

if

F 1 , F 2 ⊆ P

-Tree

(Σ)

,then

F 1 b F 2 = {t 1 b t 2 | t 1 ∈ F 1 , t 2 ∈ F 2 }

.

(6)

Finite automata with prex-orale (or p-orale) extend the lass of nite

automatabyallowingsomemembershiptests onprexof theinputword. An

automaton

A

with p-orales, on the input alphabet

A

is a nite automaton

assoiatedtoavetor

O ~ = (O 1 , . . . , O m )

ofsubsetsof

A

andwhosetransitions ontainabooleanvetorofsize

m

alledtest. Duringtheomputationby

A

of

aninputword,thealreadyproessedpart

u

oftheinputiskeptinmemoryand

atransition withtest

~o

anbeappliedif

~o

isequaltotheharateristivetor of

u

inside

O ~

(i.e.,if

~o = χ O A ~

(u)

).

Denition1(Finite automatonwith p-orales). Given

m ≥ 1

, an au-

tomaton with

m

p-orales is a tuple

A = (Q, A, ~ O, ∆, q 0 , F )

where

Q

is a

nite set of states,

A

is the input alphabet,

O ~ = (O 1 , . . . , O m )

,

O i ⊆ A

,

∆ ⊆ Q × A × {0, 1} m × Q

is theset oftransitions,

q 0 ∈ Q

isthe initialstate,

and

F ⊆ Q

isthesetofnalstates.

Aongurationof

A

is apair

(q, u

v)

where

uv ∈ A

and isasymbolwhih

doesnotbelongto

A

. Thebinaryrelationonongurationsis

Aandonsists

inallpairs

(q, u

av) →

A

(p, ua

v)

suhthat

(q, a, χ O A ~

(u), p) ∈ ∆

. Thelanguage

reognisedby

A

is

L(A) = {u ∈ A

| (q 0 ,

u) →

A

(q F , u

), q F ∈ F}

.

Wewill usethefollowingnotations: FA

O ~ (A)

isthefamilyof automataover

A

withp-orale

O ~

,thelassof

O ~

-regularlanguages(i.e.,reognisedbyautomata

inFA

O ~ (A)

) isREG

O ~ (A)

. Remark that anautomatonwithorale

is simplya

niteautomaton. WewritethenFA ratherthanFA

andREGforREG

.

Denition2. Anautomatonwith

m

p-orales

A = (Q, A, ~ O, ∆, q 0 , F )

issaid

to be deterministi if

∀p ∈ Q, a ∈ A, ~o ∈ {0, 1} m

, there is oneand only one

q ∈ Q

suhthat

(p, a, ~o, q) ∈ ∆

.

Example3. Thefollowing automatonisdeterministi and reognizethelan-

guage

{a n b n c n } n≥1

.

A = ({q 0 , q 1 , q 2 , q 3 , q F }, {a, b, c}, (O 1 , O 2 ), ∆, q 0 , {q F })

where

O 1 = {a n b n } n≥1

,

O 2 = {a m b n c n−1 } n≥1, m≥0

and

onsistsin

(q 0 , a, (0, 0), q 1 )

,

(q 1 , a, (0, 0), q 1 )

,

(q 1 , b, (0, 0), q 2 )

,

(q 2 , b, (0, 0), q 2 )

,

(q 2 , b, (0, 1), q 2 ) (q 2 , c, (1, 0), q 3 )

,

(q 2 , c, (1, 1), q F )

,

(q 3 , c, (0, 0), q 3 )

and

(q 3 , c, (0, 1), q F )

.

Weassoiatewith eah automatonwith

m

p-orales

A ∈

FA

O ~ (A)

, anite au-

tomaton

A ∈ e

FA

(A × {0, 1} m )

alled soureof

A

andonstruted bymoving

thetestofeahtransitionintotheinputletterofthetransition: eahtransition

(p, a, ~o, q)

istransformedin

(p, (a, ~o), q)

. Thelanguage

L(A)

anbeobtainfrom

thelanguage

L( A) e

andthe"harateristilanguageof

O ~

in

A

.

Denition4. For every

O ~ = (O 1 , . . . , O m )

,

O i ⊆ A

, the harateristi lan- guageof

O ~

isdenedby:

L O χ ~ = {(a 1 , ~o 1 ) . . . (a n , ~o n ) ∈ (A × {0, 1} m )

| ∀i ∈ [1, n], ~o i = χ O A ~

(a 1 . . . a i−1 )}

.

(7)

Observation 5. Forevery

O ~ = (O 1 , . . . , O m )

,

O i ⊆ A

:

REG

O ~ (A) = {π 1 (L ∩ L O χ ~ ) | L ∈

REG

(A × {0, 1} m )}.

UsingtheKleene'stheorem, weobtaineasily:

Theorem6. Let

A

analphabet, and

O ~

avetor ofsubsets of

A

,

1. REG

O ~ (A)

isthelass oflanguages reognizedbydeterministiautomatain

FA

O ~ (A)

,

2. REG

O ~ (A)

islosedunderbooleanoperations.

1.3. Iteratedpushdown stores

OriginallydenedbyGreibahin[22℄,iteratedpushdownstoresarestorage

struturesbuiltiteratively. Letusxaninnitesequene

A = A 1 , A 2 , . . . , A k , . . .

of disjoint andnite alphabets. Forall

k ≥ 1

, wedenote by

A k

the nite se- quene

A 1 , . . . , A k

and adoptthe onventionthat

A 0 = {⊥}

andthat

A 0 ∩ A i

isemptyofall

i ≥ 1

.

Denition7. Wedene indutivelythe set

k

-pds

( A k )

(or

k

-pdswhen alpha-

bets ofstoreareunderstood)of

k

-iteratedpushdown-storesover

A k

:

0

-pds

( A 0 ) = {⊥}

,

(k + 1)

-pds

( A k+1 ) = (A k+1 [k

-pds

( A k )])

⊥ [k

-pds

( A k )]

.

Theset forall

k

-pushdowns for

k ≥ 0

is denoted

it

-pds

( A )

. In therest of the

paper, any

1

-pds

a 1 [⊥]a 2 [⊥] · · · a n [⊥] ⊥ [⊥]

will bewritten simply

a 1 . . . a n ⊥

and

∀k ≥ 0

. We denoteby

⊥ k

the empty

k

-pds ontainingonly symbols

:

⊥ 0 =⊥

and

⊥ k+1 =⊥ [⊥ k ]

.

From the denition, every

ω

in

(k + 1)

-pds

( A k+1 )

,

k ≥ 0

, has a unique

deompositionas

ω = a[ω 1 ]ω

with

ω 1 ∈ k

-pds

(A k )

,

ω

∈ (k + 1)(A k+1 )

-pds

∪{ε}

and

a ∈ A k+1 ∪ {⊥}

. Furthermore,

a =⊥

i

ω

= ε

.

Example8. Let

A 1 = {a 1 , b 1 }

,

A 2 = {a 2 , b 2 }

,

A 3 = {a 3 }

bestoragealphabets,

ω ex = a 3 [b 2 [b 1 a 1 ⊥]a 2 [a 1 ⊥] ⊥ 2 ] ⊥ [a 2 [a 1 ⊥]a 2 [⊥] ⊥ 2 ] ∈ 3

-pds

( A 3 )

.

Its deomposition orresponds to

a = a 3

,

ω 1 = b 2 [b 1 a 1 ⊥]a 2 [a 1 ⊥] ⊥ 2

and

ω

=⊥ [a 2 [a 1 ⊥]a 2 [⊥] ⊥ 2 ]

.

Thetwofollowingmapswillbeuseful.

Projetion: themap assoiatingeah

k

-pdsto itstop

i

-pds,

1 ≤ i ≤ k

is

p k,i

:

k

-pds

(A 1 , . . . , A k ) → i

-pds

(A 1 , . . . , A i )

,where

∀ω = a[ω 1 ]ω

∈ k

-pds,

p k,k (ω) = ω

and

p k,i (ω) = p k−1,i (ω 1 )

if

1 ≤ i ≤ k − 1

.

The double subsript notation will be used to handle inverse funtions,

therestofthetime,wewillnote

p i

for

p k,i

.

Top symbols: themap assoiatingany

k

-pds,

k ≥ 1

to its

k

top-symbolsis

top : k

-pds

(A 1 , . . . , A k ) → (A k ∪ {⊥}) · · · (A 2 ∪ {⊥})(A 1 ∪ {⊥})

dened

∀ω = a[ω 1 ]ω

∈ k

-pdsby

(8)

top(ω) = a

,if

k = 1

,else

top(ω) = a · top(ω 1 )

.

For

i ∈ [1, k]

, and

ω ∈ k

-pds, we denote by

top i (ω)

the

i

-th letter of

top(ω)

.

Example9. Let

ω ex

bethe

3

-pdsgivenin Example8:

p 2 (ω ex ) = b 2 [b 1 a 1 ⊥]a 2 [a 1 ⊥] ⊥ 2

,

p 1 (ω ex ) = b 1 a 1 ⊥

,and

top(ω ex ) = a 3 b 2 b 1

,

top(p 2 (ω ex )) = b 2 b 1

,

top(p 1 (ω ex )) = b 1

.

An instrution on

it

-pds is a funtion from

it

-pds to

it

-pds whih does not

modifythelevelofthe

k

-pushdowns(i.e.,ifinstrisaninstrutionthenforany

k ≥ 0

and any

ω ∈ k

-pds,instr

(ω) ∈ k

-pds). An instrution of level

i

isan

instrutionwhihdoesnotmodifythelevelsgreaterthan

i

ofany

it

-pds. Hene,

giveninstraninstrutionoflevel

i

if

ω = a[ω 1 ]ω

∈ k

-pds,

k > i

,theninstr

(ω) = a[

instr

(ω 1 )]ω

if

ω ∈ k

-pds,

k < i

,theninstr

(w) = w

.

Therefore,to deneaninstrutionof level

i

, thereis onlyneedto deneitfor

any

ω ∈ i

-pds.

Threeinstrutionsoflevel

k

aregenerallyappliableto

it

-pushdowns.

Denition10. Classial instrutions of level

i ≥ 1

over

A

are dened for every

ω = b[ω 1 ]ω

∈ i

-pds

( A i )

by:

pop i (ω) = ω

if

b 6=⊥

,else

pop i (ω)

isundened,

push i,a (ω) = a[ω 1 ]ω

,

change i,a (ω) = a[ω 1 ]ω

,if

b 6=⊥

else

change i,a (ω)

isundened.

For

k ≥ 1

,

I k ( A k ) = {pop i | i ∈ [1, k]} ∪ {push i,a , change i,a | a ∈ A i , i ∈ [1, k]}

.

istheset ofinstrutionsover

A k

.

Thus, given

ω ∈ k

-pds and

i ≤ k

,

pop i (ω)

erases

p i (ω)

on the top of the

store,

push i,a

i

(ω)

onsists in add

a i [p i−1 (ω)]

on the top of the top

i

-pds and

change i,a

i

(ω)

onsistsin replae

top i (ω)

by

a i

.

Example11. Let

ω = b 3 [b 2 [b 1 ⊥] ⊥ 2 ] ⊥ 3

bea

3

-pds,

pop 3 (ω) =⊥ 3

,

pop 2 (ω) = b 3 [⊥ 2 ] ⊥ 3

,

pop 1 (ω) = b 3 [b 2 [⊥] ⊥ 2 ] ⊥ 3

,

push 2,a

2

(ω) = b 3 [a 2 [b 1 ⊥]b 2 [b 1 ⊥] ⊥ 2 ] ⊥ 3

,

push 1,a

1

(ω) = b 3 [b 2 [a 1 b 1 ⊥] ⊥ 2 ] ⊥ 3

,

change 3,a

3

(ω) = a 3 [b 2 [b 1 ⊥] ⊥ 2 ] ⊥ 3

,

change 1,a

1

(ω) = b 3 [b 2 [a 1 ⊥] ⊥ 2 ] ⊥ 3

.

Wealso dene theinverse instrutionof

push i,a

whih will beused to enode

the

k

-pushdownsaswords.

Denition12. Forany

i ≥ 1

and

a ∈ A i

, theinstrutionof level

i push i,a

is

denedforany

ω ∈ i

-pds

( A i )

by

push i,a (ω) = ω

ifthereexists

ω

∈ i

-pdssuhthat

ω = push i,a

) push i,a (ω)

isundenedelse.

Inotherwords,

∀ω ∈ k

-pds,

push k,a (ω) = ω

i

ω = a[ω 1 ]b[ω 1 ]ω

′′and

ω

= b[ω 1 ]ω

′′

.

Références

Documents relatifs

Tree automata are used to represent sets of reachable states and functional programs are modeled using term rewriting systems.. From a tree automaton representing the initial state,

Towards Deciding Second-order Unification Problems Using Regular Tree Automata..

We present a formal verification technique based on Tree Automata Completion (TAC) [20], capable of checking a class of properties, called regular properties, of higher-order

Nous avons réalisé une étude monocentrique au sein du service des urgences ophtalmologique de l’HSR. L’intérêt était d’évaluer le profil épidémiologique des

They are also the key for the decision procedures applied in order to check the conditions of constraint unsatisfiability (in rules for rewriting and Inductive Deletion,

De toute évidence, c’est dans cette perspective qui joint et l’information et la relation que réside l’enjeu dicté par l’objectif didactique initial, celui de réussir, dans

We establish a new relationship between tree automata and the logic SDom on basis of the old techniques for WS2S: We show that tree regular constraints R are equal in expressiveness

We introduce the notion of high-order deterministic top-down tree transducers (HODT) whose