Faculty of Information Technology

Spring 2024

Compiler Design
CS-352
Lecture (4)

1
Outline

• Deterministic Finite Automata

• From Regular Expression to Automata
• Construction of an NFA-ε from a Regular
Expression
• Conversion of an NFA-ε to DFA

2
Outline

• Deterministic Finite Automata

• From Regular Expression to Automata
• Construction of an NFA-ε from a Regular
Expression
• Conversion of an NFA-ε to DFA

3
Deterministic Finite Automat

• A deterministic finite automaton (DFA) is a special case of an

NFA where:-
1. There are no moves on input ε, and
2. For each state s and input symbol a, there is exactly one edge
out of s labelled a.

• If we are using a transition table to represent a DFA, then

each entry is a single state. We may therefore represent
this state without the curly braces that we use to form
sets.

• Every regular expression and every NFA can be converted

to a DFA accepting the same language, because it is the
DFA that we really implement or simulate when building
lexical analyzers.

4
Deterministic Finite Automat
(Cont.)

5
Outline

• Deterministic Finite Automata

• From Regular Expression to Automata
• Construction of an NFA-ε from a Regular
Expression
• Conversion of an NFA-ε to DFA

6
From Regular Expression to
Automata
• The regular expression is the notation of choice for
describing lexical analyzers and other pattern-processing
software. However, implementation of that software requires
the simulation of a DFA, or perhaps simulation of an NFA.

• Because an NFA often has a choice of move on an input

symbol or on ε, or even a choice of making a transition on ε
or on a real input symbol, its simulation is less
straightforward than for a DFA.

• Thus, often it is important to convert an NFA to a DFA that

accepts the same language.

7
From Regular Expression to
Automata (Cont.)

Lexical Analyzer
R.E.
NFA
F.A.
DFA
R.G.

8
From Regular Expression to
Automata (Cont.)

Lexical Analyzer
R.E.
NFA
F.A.
DFA
R.G.

9
Outline

• Deterministic Finite Automata

• From Regular Expression to Automata
• Construction of an NFA-ε from a Regular
Expression
• Conversion of an NFA-ε to DFA

10
Construction of an NFA-ε from a
Regular Expression
• We now give an algorithm for converting any regular
expression to an NFA-ε that defines the same language. For
each subexpression the algorithm constructs an NFA-ε with a
single accepting state.

11
Construction of an NFA-ε from a
Regular Expression (Cont.)
• Basis: For expression ε construct the NFA-ε.

• Here, i is a new state, the start state of this NFA-ε, and f is another
new state, the accepting state for the NFA-ε.

• For any subexpression a in Σ, construct the NFA-ε.

• where again i and f are new states, the start and accepting states,
respectively.

• Note that in both of the basis constructions, we construct a distinct

NFA-ε, with new states, for every occurrence of ε or some a as a
subexpression of r.
12
Construction of an NFA-ε from a
Regular Expression (Cont.)
• Suppose r = s | t

s
ε ε

ε ε
t

NFA-ε for the Union of two Regular Expressions

13
Construction of an NFA-ε from a
Regular Expression (Cont.)
• Suppose r = st

ε s ε t ε

s t

NFA-ε for the Concatenation of two Regular Expressions

14
Construction of an NFA-ε from a
Regular Expression (Cont.)
• Suppose r = s*
ε

ε s ε

NFA-ε for the Closure of a Regular Expression

15
Construction of an NFA-ε from a
Regular Expression (Cont.)
• Example: Let us use Algorithm 3.23 to construct an NFA-ε for r =
(a|b)*abb.

• The following Figure shows a parse tree for r that is analogous to

the parse trees constructed for arithmetic expressions.

Parse Tree for (a|b)*abb

16
Construction of an NFA-ε from a
Regular Expression (Cont.)
• For subexpression r1, the first a, we construct the NFA-ε:-

• State numbers have been chosen for consistency with what

follows. For r2 we construct:-

17
Construction of an NFA-ε from a
Regular Expression (Cont.)
• We can now combine N(r1) and N(r2), using the construction
of Union to obtain the NFA-ε for r3 = r1|r2; this NFA-ε is
shown as follows.

• The NFA-ε for r4 = (r3) is the same as that for r3.

18
Construction of an NFA-ε from a
Regular Expression (Cont.)
• The NFA-ε for r5 = (r3)* is then as shown as follows. We have
used the construction of Closure to build this NFA-ε from the
NFA-ε in the former Figure.

19
Construction of an NFA-ε from a
Regular Expression (Cont.)
• Now, consider subexpression r6, which is another a. We use
the basis construction for a again, but we must use new
states.

• It is not permissible to reuse the NFA-ε we constructed for r1,

even though r1 and r6 are the same expression.

• The NFA-ε for r6 is:-

20
Construction of an NFA-ε from a
Regular Expression (Cont.)
• To obtain the NFA-ε for r7 = r5r6, we apply the construction
of Concatenation. We merge states 7 and 7’, yielding the
NFA-ε of the following Figure. Continuing in this fashion with
new NFA's-ε for the two subexpressions b called r8 and r10,
we eventually construct the NFA for (a|b)*abb.

21
Construction of an NFA-ε from a
Regular Expression (Cont.)
• NFA-ε to recognize a regular expression letter (letter|digit)*.

22
Construction of an NFA-ε from a
Regular Expression (Cont.)
• NFA-ε to recognize a regular expression letter (letter|digit)*.

23
Outline

• Deterministic Finite Automata

• From Regular Expression to Automata
• Construction of an NFA-ε from a Regular
Expression
• Conversion of an NFA-ε to DFA

24
From Regular Expression to
Automata (Cont.)

Lexical Analyzer
R.E.
NFA
F.A.
DFA
R.G.

25
Conversion of an NFA-ε to DFA

• We shall first show how to convert NFA’s-ε to DFA's. Then,

we use this technique, known as "the subset construction," to
give a useful algorithm for simulating NFA’s-ε directly, in
situations (other than lexical analysis) where the NFA-to-DFA
conversion takes more time than the direct simulation.

• The general idea behind the subset construction is that each

state of the constructed DFA corresponds to a set of NFA-ε
states.

• After reading input a1a2 ... an, the DFA is in that state which
corresponds to the set of states that the NFA-ε can reach,
from its start state, following paths labelled a1a2 ... an.

26
Conversion of an NFA-ε to DFA
(Cont.)

27
Conversion of an NFA-ε to DFA
(Cont.)

Operations on NFA States

28
Conversion of an NFA-ε to DFA
(Cont.)
• Example: The following NFA-ε shows accepting (a|b)*abb; it
happens to be the one we shall construct directly from this
regular expression. Let us apply Algorithm 3.20 to this NFA-
ε.

29
Conversion of an NFA-ε to DFA
(Cont.)
• -closure({0}) = {0,1, 2, 4, 7}= A
• move(A, a) = {3, 8}
• -closure({3, 8}) = {3, 6, 7, 1, 2, 4, 8} = {1, 2, 3, 4, 6, 7, 8} = B
• Dtrans[A, a] = B
• move (A, b) = {5}
• -closure({5}) = {5, 6, 7, 1, 2, 4} = {1, 2, 4, 5, 6, 7} = C
• Dtrans[A, b] = C

30
Conversion of an NFA-ε to DFA
(Cont.)
• move(B, a) = {3, 8}
• -closure({3, 8}) = B
• Dtrans[B, a] = B
• move(B, b) = {5, 9}
• -closure({5, 9}) = {1, 2, 4, 5, 6, 7, 9} = D
• Dtrans[B, b] = D

31
Conversion of an NFA-ε to DFA
(Cont.)
• move(C, a) = {3, 8}
• -closure({3, 8}) = B
• Dtrans[C, a] = B
• move(C, b) = {5}
• -closure({5}) = C
• Dtrans[C, b] = C

32
Conversion of an NFA-ε to DFA
(Cont.)
• move(D, a) = {3, 8}
• -closure({3, 8}) = B
• Dtrans[D, a] = B
• move(D, b) = {5, 10}
• -closure({5, 10}) = {1, 2, 4, 5, 6, 7, 10} = E
• Dtrans[D, b] = E

33
Conversion of an NFA-ε to DFA
(Cont.)
• move(E, a) = {3, 8}
• -closure({3, 8}) = B
• Dtrans[E, a] = B
• move(E, b) = {5}
• -closure({5}) = C
• Dtrans[E, b] = C

34
Conversion of an NFA-ε to DFA
(Cont.)
Alphabet
NFA State DFA State
a b
{0,1, 2, 4, 7} A B C
{1, 2, 3, 4, 6, 7, 8} B B D
{1, 2, 4, 5, 6, 7} C B C
{1, 2, 4, 5, 6, 7, 9} D B E
{1, 2, 4, 5, 6, 7, 10} E B C

35
 Conversion of an NFA-ε to DFA
(Cont.)

Result of Applying the Subset

Construction

36