Teoría de Autómatas Y Lenguajes Formales: Francisco Vico

TEORÍA DE AUTÓMATAS Y LENGUAJES

FORMALES
Francisco Vico
departamento
Lenguajes y
Ciencias de la Computación
área de conocimiento
Ciencias de la Computación e
Inteligencia Artificial
ETSI Informática
Universidad de Málaga
[email protected]
geb.uma.es/fjv
10 de diciembre de 2015

OpenCourseWare
Teoría de autómatas y lenguajes formales Universidad de Malaga

Foreword
Writing just another textbook in theory of automata, formal languages or computability looks

like pointless: there are so many now, that we can hardly do better. However, every lecturer
knows what is to be taught, and the lectures benefit from this particular nomenclature in a
given subject, or that concrete demonstration. In the end, you miss that textbook that satisfies
your needs, that fits your communication skills or your expertise in the area. Such is the
motivation behind this compilation: to become the textbook in the subject of Theory of
automata and formal languages, in the traditional sense, but also, to be an everevolving
document, that integrates what others have expressed before, the examples and exercises that
were designed by other authors in distant parts of the planet, unsubscribing from a particular
terminology, and unifying different forms of representing the same concepts.
It is the sign of the XXI century: collaboration, contribution to common initiatives. In this
sense, this manuscript has benefitted from the efforts of many people. Part of its content was
extracted from previous documents, kindly offered to the open access community by their
authors [Ruohonen, 2009; Jian et al, 2002]. The resulting manuscript has the structure of
[Ramos and Morales, 2011], which is the reference book in Spanish for the subject Teoría de
autómatas y lenguajes formales (that is the reason because numbering in some epigraphs is
not always correlative). The selection and compilation of the manuscript has been performed
by students taking the subject during the course 201516: Iustina Andronic (Erasmus student)
and Esteban Delgado; and curated by Francisco Vico, professor of Computer Science and
Artificial Intelligence at the University of Malaga, responsible for the content of this subject
at the university’s OpenCourseWare programme.
A package of software that implements the main concepts in this manuscript is also under
development for the programming language Octave, and it has been made public under CC0
license at
https://bitbucket.org/fjvico/umafol
. This document itself, in its latest version can
also be accessed at http://j.mp/talf_ocw.
This is just version 1.0, coming years will see it develop, enrich and accommodate new
knowledge and experience.
Francisco Vico
Málaga, December 8, 2015

Vico, F. (2014) Teoría de autómatas y lenguajes formales.
OCWUniversidad de Málaga. http://ocw.uma.es.
Bajo Licencia Creative Commons
AttributionNonCommercial ShareAlike 3.0 Spain 2

OpenCourseWare

Notation

ℕ the set of nonnegative integers (or natural numbers), i.e., {0, 1, 2, … } (U+2115)
ℙ the set of positive integers (U+2119)
ℝ the set of real numbers (U+211D)
ℤ the set of integers (U+2124)
∅ the empty set (U+2205)
⊆ the (infix) subset relation between sets (U+2286)
⊂ the (infix) proper subset relation between sets (U+2282)
∪ the infix union operation on sets (U+222A)
∩ the infix intersection operation on sets (U+2229)
~ the prefix complementation operation on sets (U+007E)
− the infix set difference operation on sets (U+2212)
× the infix cartesian product of sets (U+00D7)
An the postfix n A
fold cartesian product of A
, i.e. A× … × n
( times)
A
2 the powerset of A


OpenCourseWare

1 Regular languages
Languages and grammars
Definition 2.1. An alphabet is a finite nonempty set of symbols. Symbols are assumed to be
indivisible.
Definition 2.2. A string over an alphabet Σ is a finite sequence of symbols of Σ.
Definition 2.3. A special string which contains no symbols at all is the empty string, and it is
represented by ε (sometimes λ or Λ).
Definition 2.4. The set of all strings over an alphabet Σ is denoted by Σ
∗
, and the set of all
+
nonempty strings over Σ is denoted by Σ . The empty set of strings is denoted by ∅ .
Definition 2.5. The number of symbols in a string x

x is called its length, denoted by ||. The
length of ε is 0.
Definition 2.9. Let a

x = 1 a
2 … a
n and b
y = 1 b
2 … bm be two strings. The concatenation of
x
y
and , denoted by xy , is the string a
1 a 2 … a

nb
1b 2 … b

m.

x
Then for any string x
x =
, ε x
ε = . For any string x and integer xn
n ≥ 0, we use to denote the
string formed by sequentially concatenating n copies of x
.
a
These are examples of word concatenation in the alphabet { b
, c
, }:
x aacbba
= , y caac
= , xy
=
aacbbacaac
x
= aacbba, y = ε, xy x
= =
aacbba
x
= ε, y caac
= , xy y
= =
caac
Concatenation is associative, i.e.,
xyz
( =
) xy
( )
z
As a consequence of this, repeated concatenations can be written without parentheses. On the

other hand, concatenation is usually not commutative, As a rule then
xy yx
≠ ,
but not always, and in the case of a unary alphabet concatenation is obviously commutative.


OpenCourseWare

Definition 2.14. The
nth w
(concatenation) power of the word is

x0
= ε
x
n
xxn
=
1
= xx… x
.

n
copies
Definition 2.9.
Nearly every characterization problem is algorithmically decidable for regular
languages. The most common ones are the following (where L or L 1 and L
2 are given regular
languages):
Emptiness Problem : Is the language L empty (i.e., does it equal ∅)?
L
It is fairly easy to check for a given finite automaton recognizing , whether or not
there is a state transition chain from an initial state to a terminal state.
Inclusion Problem L
: Is the language L
included in the language
1 ?
2
Clearly L L
1 ⊆ 2 if and only if
L L
1 − 2 = ∅
.

Equivalence Problem L
: Is L
1 = 2?

Clearly L =
1 L if and only if
2 L ⊆
1 L
2 and
L L
2 ⊆ 1.

Finiteness Problem L
: Is a finite language?
It is fairly easy to check for a given finite automaton recognizing L
, whether or not it
has arbitrarily long state transition chains from an initial state to a terminal state.

Membership Problem : Is the given word w in the language L or not?
Using a given finite automaton recognizing L it is easy to check whether or not it
accepts the given input word w.

Definition 2.20. For any alphabet Σ, a language over Σ is a set of strings over Σ. The
members of a language are also called the words of the language.
Definition 2.30. A grammar is a quadruple (Σ, V S
, P
, ), where:
1. Σ is a finite nonempty set called the terminal alphabet. The elements of Σ are called
the terminals.
2. V is a finite nonempty set disjoint from Σ. The elements of V are called the
nonterminals or variables.
3. S ∈ V
is a distinguished nonterminal called the start symbol.
4. P is a finite set of productions (or rules) of the form
α→β


OpenCourseWare

where α ∈ (Σ∪V
)∗V

(Σ∪ V
)∗
and β ∈ (Σ∪V
)∗
, i.e. α is a string of terminals and
nonterminals containing at least one nonterminal and β is a string of terminals and
nonterminals.
Definition 2.31 Let G = (Σ, V S

, P
, ) be a grammar, and let γ 2 be two sentential forms of
1 , γ G .
We say that γ
1 directly derives γ
2 , written γ
1 ⇒ γ
2 , if γ
1 = σατ , γ
2 = σβτ , and α → β is a
production in P .
Definition 2.35. Let γ 2 be two sentential forms of a grammar

1 and γ G. We say that γ1 derives
γ, written γ
2 1 ⇒
∗
γ, if there exists a sequence of (zero or more) sentential forms σ
2 , … , σ
1 n
such that
γ ⇒ σ
1 ⇒ … ⇒ σ
1 ⇒ γ
n .
2
The sequence γ ⇒ σ
1 ⇒ ... ⇒ σ
1 n ⇒ γ
is called a derivation of γ
2 from γ
2 .
1
Definition 2.37. Let V

G = (Σ, S
, P
, ) be a grammar. A sentential form of
G is any string of
terminals and nonterminals, i.e. a string over Σ ∪ V
.
Definition 2.39. Let V

G = (Σ, S
, P
, G
) be a grammar. The language generated by , denoted by
LG
(), is defined as
L G
( x
) = { x
|
∈ Σ S
,
∗
⇒
∗
x
}.
L
The words in G
( L
) are also called the sentences of G
().
G
Let V
= (Σ, S
, P
, ) be a grammar.
Chomsky’s Hierarchy In Chomsky’s hierarchy grammars are divided into four types:
Type 0: No restrictions.
Type 1: CS grammars.
Type 2: CF grammars.
X
Type 3: Linear grammars having productions of the form i→wX
X
j or j→w
X
where i
*
X
and are nonterminals and
j w ∈ ΣT , the socalled rightlinear grammars.
Grammars of Types 1 and 2 generate the socalled CSlanguages and CFlanguages,
respectively, the corresponding families of languages are denoted by CS and CF. Languages
generated by Type 0 grammars are called computably enumerable languages (CElanguages),
the corresponding family is denoted by CE.
G
Definition 2.58. is also called a Type0 grammar or an unrestricted grammar.


OpenCourseWare

Definition 2.59. G is a Type1 or contextsensitive grammar if each production α→β in P
satisfies |α| ≤ |β|. By “special dispensation,” we also allow a Type1 grammar to have the
production S→ε, provided S
does not appear on the righthand side of any production.
Definition 2.60. G is a Type2 or contextfree grammar if each production α→β in

P satisfies
|α| = 1; i.e., α is a single nonterminal.
Definition 2.61.
G is a Type3 or rightlinear or regular grammar if each production has one
of the following two forms:
A→ cB
A→ c
where A B
, are nonterminals (with B A
= allowed) and c
is a terminal.
Definition 2.78. Let L L

1 and L
2 be two languages over Σ. The concatenation of L
1 and ,
2
denoted by L L
1 , is the language {
2 xy x
| ∈ L L
, y ∈
1 }.
2
Definition 2.82. Let L

L be a language over Σ. Define Li
0 = {ε} and = for
LLi−1 i ≥ 1. The
Kleene closure of L
, denoted by L
∗
, is the language

i
L* = ∪
i≥0
L .
L
The positive closure of , denoted by
L+
, is the language

+ i
L = ∪
i≥1
L .
In other words, the Kleene closure of a language L consists of all strings that can be formed
by concatenating zero or more words from L
. For example, if
L = {0, 01}, then
LL = {00,
001, 010, 0101}, and L
∗
comprises all binary strings in which every 1 is preceded by a 0.
Note that concatenating zero words always gives the empty string, and that a string with no
1s in it still makes the condition on “every 1” true. L + has the meaning “concatenate one or
more words from L,” and satisfies the properties L∗ L+
= L+
∪ {ε} and =
LL∗ . Furthermore,
for any language L L
, ∗
always contains , and L + contains if and only if L does. Also note that
Σ is in fact the Kleene closure of the alphabet Σ when Σ is viewed as a language of words of
∗
+
length 1, and Σ is just the positive closure of Σ.
Closure of the types of languages
Closure properties are often useful in constructing new languages from existing languages,
and for proving many theoretical properties of languages and grammars. The closure
properties of the four types of languages in the Chomsky hierarchy are summarized below.
Proofs may be found in [Harrison, 1978], [Hopcroft and Ullman, 1979], or [Gurari, 1989];


OpenCourseWare

the closure of the CSLs under complementation is the famous ImmermanSzelepcsényi
theorem.
Theorem.
1. The class of Type0 languages is closed under union, intersection, concatenation, and

Kleene closure, but not under complementation.
2. The class of contextfree languages is closed under union, concatenation and Kleene
closure, but not under intersection or complementation.
3. The classes of contextsensitive and regular languages are closed under all of the five
operations.
For example, let L mnp

1 = {0 1 2 | n
m = L
}, mnp
2 = {012| p
n = }, and L 3 = {0
mnp
12 m =
| n or p
n = }.
Now L1 is the concatenation of the contextfree languages {0
nn
1 |
n ≥ 0} and 2
∗
L
, so 1
is
contextfree. Similarly L
2
is contextfree. Since L
3
L
= 1
∪ L
2
L
, 3
is contextfree. However,
intersecting L1
L
with 2
gives the language {{0 mnp
12 m
| = n p
= }, which is not contextfree.
Definition 2.11. A class of languages is said to be closed under a particular operation (such as

union, intersection, complementation, concatenation, or Kleene closure) if every application
of the operation on language(s) of the class yields a language of the class.


OpenCourseWare

Regular expressions
Definition 3.1. The regular expressions over an alphabet Σ and the languages they represent
are defined inductively as follows.
1. The symbol ∅ is a regular expression, and represents the empty language.
2. The symbol ε is a regular expression, and represents the language whose only member
is the empty string, namely {ε}.
3. For each c ∈ Σ,
c is a regular expression, and represents the language {c
}, whose
only member is the string consisting of the single character c.
4. If
r and
s are regular expressions representing the languages R and S r +
, then ( s rs
), ( )
and ( r
∗
) are regular expressions that represent the languages R ∪ S RS
, , and R
,
∗
respectively.


OpenCourseWare

Finitestate automata
Automata are used to recognize words of a language. An automaton then ”processes” a word

and, after finishing the processing, ”decides” whether or not the word is the language.
An automaton is finite if it has a finite memory, i.e., the automaton may be thought to be in
one of its (finitely many) (memory)states. A finite deterministic automaton is defined
formally by giving its states, input symbols (the alphabet), the initial state, rules for the state
transition, and the criteria for accepting the input word.
q
Definition 4.1. A finite (deterministic) automaton (DFA) is a quintuple M = (Q, Σ, 0 , δ, A)
where
Q = { q,
0 q,…,
1 q } is a finite set of states, the elements of which are called states;
m
Σ is the set input symbols (the alphabet of the language);

q0 is the initial state (
q0 ∈ Q);
δ is the (state) transition function which maps each pair ( qi
a
, qi
), where is a state, and
a
is an input symbol, to exactly one next state q q
: δ(
j a
i, q
) = j ;

A
is the socalled set of terminal states ( A ⊆ Q).
As its input the automaton M receives a word
w a
= a
…
1 n
which it starts to read from the left. In the beginning q

M is in its initial state 0 reading the first
symbol a 1 of w qj
. The next state is then determined by the transition function:

q q
j = δ( a
,
0 ).
1
In general, if q

M is in state j reading the symbol a , its next state is δ(
i q a
,
j ) and it moves on to
i
read the next input symbol a i , if any. If the final state of
+1 M after the last input symbol a n is
read is one of the terminal states ( A
a state in ), then M accepts w w
, otherwise it rejects . In
particular, M accepts the empty input ε if the initial state q is also a terminal state.
0
The language recognized by an automaton

M is the set of the words accepted by the
L
automaton, denoted by M
( ).
Any word a
w = a
1 … n , be it an input or not, determines a socalled state transition chain of
the automaton M from a state q q
0 to a state
j jn :

q
j q
0 , j q
1 , … , jn ,


OpenCourseWare

q
where always ji+1 q
= δ( a
,
ji i ).
+1
Any finite automaton can be represented graphically as a socalled state diagram. A state is
then represented by a circle enclosing the symbol of the state, and in particular a terminal
state is represented by a double circle:
q
A state transition δ(i,a
q
) = a
j is represented by an arrow labelled by , and in particular the
initial state is indicated by an incoming arrow:

Such a representation is in fact an edgelabelled directed graph.
A
Example. The automaton { B
, , 10}, {0, 1}, A, δ, {10} where δ is given by the state
transition table
is represented by the state transition diagram

The language recognized by the automaton is the regular language (0+1)
10.
∗
Definition 4.21. Defined formally a nondeterministic finite automaton (NFA) is a quintuple
M = (Q, Σ, S, δ, A) where:
● Q, Σ A
and are as for the deterministic finite automaton;
● S is the set of initial states;
q
● δ is the (state) transition function which maps each pair (i,
a), where qi is a state and
a is an input symbol, to exactly one subset T of the state set Q ( qi , a) = T
: δ .


OpenCourseWare

Definition 4.34. M accepts a word w if there is at least one terminal state in the set of states
δ (S, w)
*
. S
Λ is accepted if there is at least one terminal state in . The set of exactly all words
accepted by M is the language L(M) recognized by M .

Definition 4.36. The nondeterministic finite automaton may be thought of as a generalization
of the deterministic finite automaton, obtained by identifying in the latter each state qi by the
corresponding singleton set { qi}. It is however no more powerful in recognition ability.

Proof . Consider a language L recognized by the nondeterministic finite automaton M = (Q

, Σ
,
S δ
, A
, ). The equivalent deterministic finite automaton is then M Q
1 = ( Σ
1,
q
, δ
0, A
1, 1) where

Q =
1 2
Q
,
q =
0 S
,
δ =
1 δ̂
,
A
and 1 consists of exactly all sets of states having a nonempty intersection with A. The states
*
M
of M
1 are thus all sets of states of . We clearly have δ1 q
(
0, w ) = δ̂

(
∗
S w
, ), so M
M and 1
accept exactly the same words, and M1 recognizes the language
L.


OpenCourseWare

Regularity conditions
Definition 5.1. The language

L separates the words
w and v if there exists a word u such that
one of the words wu and
vu is in
L and the other one is not. If L does not separate the words w
v
and , then the words
wu and
vu are always either both in L
L or both in u
, depending on .
There is a connection between the separation power of a language recognized by a finite
automaton and the structure of the automaton:
Theorem. If the finite automaton Q

M = ( Σ
, δ
, q0 , A
, ) recognizes the language L
, and for the
w
words and v δ
it verifies *
q
( w
0 ,
δ
) = *
q
( v
0, ), then L does not separate w v
and .
Proof. As is easily seen, in general
δ*

q
( xy
i , δ*
) = δ
*
( q
( x
i, y
), ).
So
δ*

q
( wu
0, δ*
) = δ
*
( w
( q0 , u
), δ*
) = δ
*
( q
( v
0, u
), δ*
) =
q
( vu
0, )
Definition 5.6. If the language L can be recognized by a finite automaton with n states,

x ∈ L
and |x| ≥ n
, then
x can be written in the form x = uvw where |uv| ≤ n, v ≠ Λ and the ”pumped”
words uv w
m L
are all in .

Myhill–Nerode Theorem A language is regular if and only if it has a finite index.
If a regular language L is defined by a deterministic finite automaton M = (Q, Σ, q 0 , δ, A)
recognizing it, then the minimization naturally starts from M . The first step is to remove all
idle states of M , i.e., states that cannot be reached from the initial state. After this we may
assume that all states of M can expressed as δ∗(q0

, w) for some word
w . For the minimization
the states of M are partitioned into Mequivalence classes as follows. The states qi and qj are
not Mequivalent if there is a word u such that one of the states δ (q
∗
i , u) and δ (q
∗
j , u) is
terminal and the other one is not, denoted by qi ≡ M q j . If there is no such word u , then q i and
qj are Mequivalent, denoted by q
i ≡ M q We may obviously assume
j. qi ≡ M qi . Furthermore, if
qi ≡M q
j , then also qj ≡
M qi
, and if q
i ≡M q q
j and j ≡
M q
k it follows that q
i ≡M q
k. Each
equivalence class consists of mutually M equivalent states, and the classes are disjoint. (Cf.
the Lequivalence classes and the equivalence relation ≡ L
.) Let us denote the M equivalence
class represented by the state q i by { qi}.
Note that it does not matter which of the
M equivalent states is chosen as the representative of the class. Let us then denote the set of
all Mequivalence classes by Q .
M L
equivalence and equivalence are related since〈
(q

δ∗ w
0, )〉
= 〈
(q

δ∗ v
0, )〉if and only if


OpenCourseWare

w
[ v
] = []. Because now all states can be reached from the initial state, there are as many
Mequivalence classes as there are Lequivalence classes, i.e., the number given by the index
of L
. Moreover, Mequivalence classes and L
equivalence classes are in a onetoone
correspondence:

(q

δ∗
〈 , w
0 )〉⇌ w
[],
q
in particular 〈〉 ⇋
0 Λ
[].


Teoría de Autómatas Y Lenguajes Formales: Francisco Vico

Uploaded by

Copyright:

Available Formats

Teoría de Autómatas Y Lenguajes Formales: Francisco Vico

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Teoría de Autómatas Y Lenguajes Formales: Francisco Vico

Uploaded by

Copyright:

Available Formats

TEORÍA DE AUTÓMATAS Y LENGUAJES

Writing just another textbook in theory of automata, formal languages or computability looks

Definition 2.5. The number of symbols in a string ​ x​

Definition 2.9. Let ​ a​

As a consequence of this, repeated concatenations can be written without parentheses. On the

Definition 2.31 Let ​ G = (Σ, ​V​ S​

Definition 2.35. Let γ​ 2 be two sentential forms of a grammar ​

Definition 2.37. Let ​ V​

Definition 2.39. Let ​ V​

Definition 2.60. ​ G is a Type­2 or context­free grammar if each production α→β in ​

Definition 2.78. Let ​L​ L​

Definition 2.82. Let ​ L​

1. The class of Type­0 languages is closed under union, intersection, concatenation, and

For example, let ​L​ m​n​p​

Definition 2.11. A class of languages is said to be closed under a particular operation (such as

Automata are used to recognize words of a language. An automaton then ”processes” a word

which it starts to read from the left. In the beginning ​ q​

In general, if ​ q​

The language recognized by an automaton ​

Proof​ . Consider a language L recognized by the nondeterministic finite automaton ​ M = (​Q​

Definition 5.1. The language ​

Theorem. If the finite automaton ​ Q​

Definition 5.6. If the language ​ L can be recognized by a finite automaton with ​ n states, ​

You might also like

Definition 2.5. The number of symbols in a string x

Definition 2.9. Let a

Definition 2.31 Let G = (Σ, V S

Definition 2.35. Let γ 2 be two sentential forms of a grammar

Definition 2.37. Let V

Definition 2.39. Let V

Definition 2.60. G is a Type2 or contextfree grammar if each production α→β in

Definition 2.78. Let L L

Definition 2.82. Let L

1. The class of Type0 languages is closed under union, intersection, concatenation, and

For example, let L mnp

which it starts to read from the left. In the beginning q

In general, if q

The language recognized by an automaton

Proof . Consider a language L recognized by the nondeterministic finite automaton M = (Q

Definition 5.1. The language

Theorem. If the finite automaton Q

Definition 5.6. If the language L can be recognized by a finite automaton with n states,