• Aucun résultat trouvé

A Note on the Descriptional Complexity of Semi-Conditional Grammars

N/A
N/A
Protected

Academic year: 2022

Partager "A Note on the Descriptional Complexity of Semi-Conditional Grammars"

Copied!
6
0
0

Texte intégral

(1)

Semi-Conditional Grammars

Tom65 Nlasopust

Dept. of Information Systems, Faculty of Information Technology, Brno University of Technology, BoZet6chova 2,

612 66 Brno, Czech Republic nasopust@fit . vutbr. cz

Abstract. The descriptional complexity of semi-conditional grammars is studied. A proof that every recursively enumerable language is gener-

lji";'I":ff ffi ru*::?H:"ffi Xl,?li".ffi f iJ,,-,T?,ll,J"'"'n'"

Keywords: formal languages, descriptional complexity, semi-conditio- nal grammars

1 fntroduction

This paper studies the descriptional complexity of semi-conditional grammars (see [4, 7-9] for more details) with respect to the number of conditional produc- tions and nonterminals.

Semi-conditional grammars are modified context-free grammars, where a per- mitting and a forbidding context is associated with each production. This means that a production is applicable if its permitting context is contained in the cur- rent sentential form and its forbidding context is not. As a special case of semi- conditional grammars, we obtain simple semi-conditional grammars introduced in [3], where one of the contexts is required to be a special symbol 0, i.e., either a permitting or a forbidding context is associated with each production.

Whereas the descriptional complexity of simple semi-conditionai grammars has been studied carefully (see [5, 7, B, 10]), the descriptional complexity of semi- conditional grammars has not been studied at all, and all results concerning the descriptional complexity of semi-conditional grammars are consequences of results concerning the descriptional complexity of simple semi-conditional gram- mars. Specifically, in [8], u proof that every recursively enumerable language is generated by a (simple) semi-conditional grammar of degree (2,1) with no more than twelve conditionai productions and thirteen nonterminals was given. Later, in Ii0], this result was improved and a proof that every recursively enumerable Ianguage is generated by a (simple) semi-conditional grammar of d.egree (2,I) with no more than ten conditional productions and twelve nonterminals was given. Finally, the result from [10] was improved in [5], where a proof that ev- ery recursively enumerable language is generated by a (simple) semi-conditional

(2)

grammar of d.egree (2, 1) with no more than nine conditional productions and ten nonterminals was given. However, a better result can be achieved for semi- cond.itional grammars than for simple semi-conditional grammars. In this pa- per, a proof that every recursively enumerable language is generated by a semi- conditional grammar of degree (2, 1) with no more than seven conditional pro- ductions and eight nonterminais is given.

2 Preliminaries and Definitions

This paper assumes that the reader is familiar with the theory of formal lan- guages (see 11,6]). For an aiphabetV, V* represents the free monoid generated by v. The unit off v* is d.enoted by e. set v+ :v* -

{r}. set sub(u) : {u : u is a substring of tr.r).

In [2], it was shown that every recursively enumerable language is generated by a grammar

G : ( { S , A , B , C ) , 7 , P U { A B C - + s } , S ) in the Geffert

form

S -, uSa, S -. uSu, S '-+ uu,

nor.mal form, where P contains context-free productions of the

w h e r e u e { A , A B } . , a e T , w h e r e u € { A , A B } - , u € { B C , C } * , w h e r e u € { A , A B } . , , u e { B C , C } - .

In addition, any terminal derivation is of the form ,S =+* w1u2w

b y p r o d u c t i o n s fr o m P , w h e r e w t e {A, B}-, u2 € {B,C}.,, w e ?*, and

W 1 U 2 W 1 " 1 1 1

by ABC -- e.

A semi,-condi,tional grammar,G, is a quadruple G : (l/, T, P, S),

where

- ,Af is a nonterminal alphabet,

- ? is a terntinal alphabet such that I/ )T :0, - S e l/ is the start sYmbol, and

- P is a fi.nite set of productions of tlr.e form ( X * a , u , u )

w i t h X € l / , a € ( l / U ? ) - , a n d u , t ) e ( l / u T ) * u { 0 } , r v h e r e 0 / I / u T i s a special symbol.

(3)

It u l0 or u f 0, then the production (X - a,u,u) € P is said to be condi,t'ional.

G h a s d e g r e e ( i , j ) i f f o r a l i p r o d u c t i o n s (X , etu,u) € P,u* 0 implies l"l <i and u l 0 implies lul < j. For n € (.4/UZ)+ and y € (liU?)., r directly deriues y a c c o r d i n g t o t h e p r o d u c t i o n ( X * a , u , u ) € P , d e n o t e d b y

n + a

i f . r: r 1 X r 2 , U : r y a r , 2 , f o r s o m e fr r t r z € ( l f U T ) * , a n d u , l0 i m p l i e s th a t u e sub(r) and u + 0 implies that u / sub(r). As usual, =+ is extended to =)', for i ) 0, 9*, and +*. The language generated by a semi-conditional grammar, G, is defined as

g ( G ) :

{ w e T " : S = + * u } .

L e t G : ( l / , 7 , P , . 9 ) b e a s e m i - c o n d i t i o n a l g r a m m a r . I f ( X > e..tt",u) € P implies that 0 e {u,u}, then G is said to be a si,mpLe semi,-condi,ti,onal gro,n'Lmar.

Main Result

This section presents the main result concerning the descriptionai complexity of semi-condit ic.rnal grammars.

Theorem 1. Euery recursiuelA enumerable language i,s generated by a semi- conditional grammar of degree (2, 1) w'ith no more than 7 condit'ional product,ions and B nonterminals.

Proof i,dea.

The main idea of the proof is to simulate a terminal derivation of a grammar, G, in the Geffert normal form.

To do this, we first appiy all context-free productions as applied in the G's derivation, and then we simulate the production ABC --+ E so that we mark with ' only one occurrence of A, one of B, and one of C and check that these marked symbols form a substring A'B'C' of the current sentential form. If so, the marked symbols can be removed, which completes the simulation of the production ABC --+ e in G; otherwise, the derivation must be blocked.

The formal proof follows.

Proof. Let L be a recursively enumerable language. There is a grammar G : ( { S , A , B , C } , T , P U { A B C - + s i , ^ 9 )

in thc Geffcrt normal form such that L: 9(G). Construct the grammar

where

G' : ({^9, A, B, C, A' , B' , C', $} ,7, P' U P" ,, S),

P ' : { ( X * a , 0 , 0 ) : X - - - a e P } ,

and P" contains foiiowing seven conditional productions:

(4)

1 . ( A * $ , 4 ' , 0 , $ ) , 2 . ( B - B ' , A ' , B ' ) , 3 . ( C - C ' $ , A ' B ' , C ' ) , 4 . ( B ' - e , B ' C ' , 0 ) , , 5 . ( C ' - e , A ' C ' ,0 ) , 6 . ( A ' - e , A ' $ , 0 ) , 7 . ( $ - e , 0 , A ' ) .

To prove thar I (G) e

g(G'), consider a derivation ,S =+* wABCw'y ) 111111ty

in G by productions from P with only one application of the productio n ABC -- e , w h e r e w , w ' e {A, B,C}* and u €. T*. Then,

S + * w A B C w ' u

in G' by productions from P'. iVloreover, try productions r, 2, z, 4, b, 6, 7, 7, we

C'pr.

b - "

uABCw'u + w$A'BCw'u + w$A' B'Cw'u + w$A' B'C'$w'u + w$A'C'$w'u + w$A'$w'u

=+ tl$$tr'u + w $ w ' u

+ lt;lt)'U.

The inclusion follows by induction.

To prove that 9(G) )_ -?(G'), consider a terminal derivation. Let X € {A,B,C} be in a sentential form of this d.erivation. To eliminate X, there are followirig three possibilities:

1. If X - A, then there must be derivation;

2. If X - B, then there must be derivation;

3. If X - C, then there must be derivation.

C and A (bV productions 6 and 3) in the C and A (by productions 4 and 3) in the A and B (bV productions 5 and 3) in the

In all above ca^ses, there are A, B, and C in the derivation. By productions 1,2, 3, and 7, there cannot be more than one A', B', and C, in any sentential form of this terminal derivation. Nloreover, by productions 3 and ,1, A' B'C' is a substring of a sentential form of tltis terminal derivation, and. there is no terminal syrnbol between any fwo nonterminals; otherrvise, there will be a situation in which (zrt

(5)

Ieast) one of productions 3 and 4 will not be applicable. Thus, any first part of a terminal derivation in G' is of the fornr

, S + * u l A B C w 2 w + 3 w t $ A ' B ' C ' $ w 2 w ( 1 ) by productions from P' and productions 1, 2, and 3, where u.,1 € {A, B}*, w z e {B,C}*, and w e T*. Next, only production 4 is applicable. T h u s ,

I wt,g.A' C'$wzw .

Besides a possible application of production 2, oniy production 5 is applicable.

Thus,

+ + r \ 9 A ' $ w ' r w

w h e r e t r ' , € {A,B,B'}*,.L € {8, B',,C}*. Besides a possible application of production 2, only production 6 is applicable. Thus,

+ + u ' 1 $ $ w ' J w

w h e r e w ' i e {A,B,B'}",wU € {B,B',C}*. Finally, only production 7 is appli- c a b l e . i. e . .

+2 w'lw'/w .

Thus, by productions I, 2, 3, or 1, 3, if production 2 has already been applied, we get

*" "tt'utD .

H e r e '

u u t u € . {ur'A' B' c'$u2w : u1 e. {A, B}* , uz € {B , c}- } o r u u : € .

Thus, the substring ABC and only this substring was eliminated during the previous derivation. By induction (see (1)), the inclusion hoids. This derivation can be performed in G with an application of the production ABC ---+ €, too. D

Thzs work has been supported by the Grant Agency of the Czech Republi,c within the project I'{o. 102/05/H050, FRVS grant I{o. FR762/2007/G1, and the Czech Mtni,stry of Education under the Research PIan No. MSM 0021630528.

References

1. J. Dassow and Gh. PXun. Regulated Rewriti,ng i,n Formal Language Theory.

Springer-Verlag, Berlin, 1989.

2. V. Geffert. Context-free-like fornrs for the phrase-stnrcture grammars. Tn X,I. Chytil, L. Janiga, and V. Koubek, editors, MFCS, volume 324 of Lecture Notes i,n Computer Sc'ience, pages 309-317. Springer, 1988.

(6)

, f .

,1

o .

6 . 7 .

8 . 9 .

A. Gopalaratnam and A. Meduna. On semi-conditional grammars with productions having either forbidding or permitting conditions. .4cfa Cybernetica, LL(4):307-324, 1 9 9 4 .

J. Kelemen. Conditional grammars: Motivations, definitions, and some properties.

In Proc. Conf. Automata, Languages and Mathemati,cal Sciences, pages 110-123, Salg6tarjein, 1984.

T. IVlasopust. An improvement of the descriptional complexity of grammars reg- ulated by context conditions. In Second Doctoral Workshop on Mathematical and Engineering Method,s in Computer Sc'ience (MEMICS 2006), pages 105-1i2, NIikulov, 2006.

A. N{eduna. Automata and Languages: Theory and Applicat'ions. Springer-Verlag, London, 2000.

A. Nleduna and M. Svec. Reduction of simple semi-conditional grammars with respect to the number of conditional productions. Act,a Cybent,et'ica, 15:353-360, 2002.

A. l\,leduna and M. Svec. Gramrnars wi,th Contert Conditions and Thei,r Appli,ca- tions. John Wiley & Sons, New York, 2005.

Gh. Piun. A variant of random context grammars: Semi-conditional grammars.

Theoret'ical Computer Sc'ience, 47:7-17, 1985.

Gy. Vaszil. On the descriptional complexity of some rewriting mechanisms regu- lated by context conditions. Theoret'ical Computer Sci,ence,330:361-373, 2005.

1 0 .

Références

Documents relatifs

We propose here new conditional extreme value index estimators based on a class of functions satisfying some mild conditions.. We only assume that the conditional distribution of

Once this constraint is lifted by the deployment of IPv6, and in the absence of a scalable routing strategy, the rapid DFZ RIB size growth problem today can potentially

Potential action items for the W3C included investigating the formation of a privacy interest group and formulating guidance about fingerprinting, referrer headers,

The first was that the Internet Architecture Board should review and consider that document in the context of evaluating Birds of a Feather (BoF) session proposals at the

Radiation protection instrumentation – Recommended climatic, electromagnetic and mechanical performance requirements and methods of tests.. Instrumentation pour la radioprotection

This part of IEC 61757 defines detail specifications for fibre optic sensors using one or more fibre Bragg gratings FBG as the sensitive element for strain measurements..

Typically, users access the on-line hosts via a phone number which links into a a public data network using packet switching (there's more on these networks in chapter 7).. The

Different types of reliability testing, such as fixed duration, sequential test-to-failure, success test, reliability demonstration, or reliability growth/improvement tests can