Question

Prove that a language A is regular if and only if there exists an NFA (nondeterministic finite automaton) which recognizes A.

Answer 1

The forward bearing is unimportant, since A normal methods there is a DFA

that remembers it, and a DFA can be viewed as a NFA rather quickly.

So we center around the regressive bearing. Expect that A will be a dialect that

is perceived by a NFA M = (Q, Σ, ∆, q0, F). Without loss of simplification,

we realize we can take the NFA to have no advances. To demonstrate An is standard,

we have to develop a DFA M0 = (R, Σ, δ, r0, G) that perceives A. (To

recognize conditions of M from conditions of M0

, we use r to extend over conditions of

M0

, and R to speak to the arrangement all things considered.)

The DFA M0 will mimic the NFA M, as in when following

images of a string in M0

, the way taken will some way or another catch all the

conceivable ways that can be taken in M.

Characterize M0 = (R, Σ, δ, r0, G) by taking:

R = {r | r ⊆ Q}

δ(r, a) = {q

0

| q

0 ∈ ∆(q, a) for some q ∈ r}

= ∪q∈r∆(q, a)

r0 = {q0}

G = {r | r ∩ F 6= ∅}

We currently need to confirm that L(M0

) = L(M), that will be, that M0 and M

perceive a similar dialect. At the end of the day, we have to demonstrate that for

each string w, M acknowledges w if and just if M0 acknowledges w.

We do this by acceptance on the length of w (since we have to appear

something valid for a vast number of things). In the first place, characterize some valuable

documentation. On the off chance that δ is the change connection of a DFA, δ

∗

(q, w) lets you know

which state you end up in on the off chance that you pursue every one of the images in w from state q,

in view of the change δ. Formally, δ

∗

is characterized inductively on the structure

of a string:

δ

∗

(q, ) = q

δ

∗

(q, w · a) = δ(δ

∗

(q, w), a),

where w · an is the link of string w and image a. It isn't difficult to

demonstrate that a DFA M = (Q, Σ, δ, q0, F) acknowledges w if and just if δ

∗

(q0, w) is in

F.

Correspondingly, if ∆ is the progress connection of a NFA without advances,

we can characterize ∆∗

that reveal to you which states you can finish up in the event that you pursue

1

every one of the images in w from state q of the NFA, in view of the change ∆. As

above:

∆∗

(q, ) = {q}

∆∗

(q, w · a) = ∪q

0∈∆∗(q,w)∆(q

0

, a)

Once more, it isn't difficult to demonstrate that a NFA M = (Q, Σ, ∆, q0, F) acknowledges w

on the off chance that and just ∆∗

(q0, w) has an express that shows up in F.

Presently, given our automata M and M0 as characterized above, we demonstrate that

for all strings w, δ

∗

(r0, w) = ∆∗

(q0, w). For the base case w = , since

r0 = {q0}, we have δ

∗

(r0, ) = δ

∗

({q0}, ) = {q0} = ∆∗

(q0, ), as required.

For the inductive case, accept the outcome is valid for a string w, we have to

show it is valid for a string w · a: By definition, δ

∗

(r0, w · a) = δ(δ

∗

(r0, w), a).

By the enlistment speculation, δ

∗

(r0, w) = ∆∗

(q0, w), and consequently δ

∗

(r0, w · a) =

δ(∆∗

(q0, w), a). By the meaning of δ, δ(∆∗

(q0, w), a) = ∪q∈∆∗(q0,w)∆(q, a),

which is simply ∆∗

(q0, w · an), as required. This demonstrates the announcement.

Presently, assume that M acknowledges w, that is, ∆∗

(q0, w) ∩ F 6= ∅. By the

above outcome, this is identical to δ

∗

(r0, w) ∩ F 6= ∅, that is, δ

∗

(r0, w) ∈ G,

what's more, this is equal to M0 tolerating w. This sets up that M and M0

acknowledge similar strings, that is, perceive a similar dialect.