Construct a recognizer, Theory of Computation

Let L1 and L2 be CGF. We show that L1 ∩ L2 is CFG too.

Let M1 be a decider for L1 and M2 be a decider for L2 .

Consider a 2-tape TM M:

"On input x:

1. copy x on the second tape

2. on the ?rst tape run M1 on x

M=

3. if M1 accepted then goto 4. else M rejects

4. on the second tape run M2 on x

5. if M2 accepted then M accepts else M rejects."

The machine M is a decider and it accepts a string x i? both M1 and M2 accept x.

Two-tape TM is as expressive as the single tape TM.

 

8.3 b)

Let L1 and L2 be recognizable languages with the corresponding recognizers M1 and M2 . We construct a recognizer M for L1 ∪ L2 .

Strategy I: run M1 and M2 in parallel on a 2-tape TM M

M = "On input x:

1. Copy x on the second tape.

2. Do one step of M1 on tape 1 and one step of M2 on tape 2.

3. If either M1 or M2 accepted, then M accepts, else goto 2."

Strategy II: nondeterministically choose to run M1 or M2

M = "On input x:

1. Nondeterministically choose i ∈ {1, 2}.

2. Run machine Mi on the input x.

3. If Mi accepted, then M accepts.

If Mi rejected, then M rejects."

Posted Date: 2/23/2013 12:52:16 AM | Location : United States







Related Discussions:- Construct a recognizer, Assignment Help, Ask Question on Construct a recognizer, Get Answer, Expert's Help, Construct a recognizer Discussions

Write discussion on Construct a recognizer
Your posts are moderated
Related Questions
Automaton (NFA) (with ε-transitions) is a 5-tuple: (Q,Σ, δ, q 0 , F i where Q, Σ, q 0 and F are as in a DFA and T ⊆ Q × Q × (Σ ∪ {ε}). We must also modify the de?nitions of th



Computer has a single FIFO queue of ?xed precision unsigned integers with the length of the queue unbounded. You can use access methods similar to those in the third model. In this

These assumptions hold for addition, for instance. Every instance of addition has a unique solution. Each instance is a pair of numbers and the possible solutions include any third

Define the following concept with an example: a.    Ambiguity in CFG b.    Push-Down Automata c.    Turing Machine

write grammer to produce all mathematical expressions in c.

Since the signi?cance of the states represented by the nodes of these transition graphs is arbitrary, we will allow ourselves to use any ?nite set (such as {A,B,C,D,E, F,G,H} or ev


The fact that regular languages are closed under Boolean operations simpli?es the process of establishing regularity of languages; in essence we can augment the regular operations