Scribd
Upload
165 views

Assign 3 Ans

This document contains the solutions to 5 problems regarding theory of computation. Problem 1 discusses how a 2-PDA can recognize the language {0n1n2n | n ≥ 1} which is not context-free. Problem 2 shows that a language is decidable if and only if its strings can be enumerated in a certain ordering. Problem 3 constructs a decider to check if a DFA accepts a language and its reverse. Problem 4 builds a decider to determine if a DFA accepts any palindromes. Problem 5 designs a decider that checks if a CFG generates exactly k strings or infinitely many.

Uploaded by

Kamal Walia
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
165 views

Assign 3 Ans

This document contains the solutions to 5 problems regarding theory of computation. Problem 1 discusses how a 2-PDA can recognize the language {0n1n2n | n ≥ 1} which is not context-free. Problem 2 shows that a language is decidable if and only if its strings can be enumerated in a certain ordering. Problem 3 constructs a decider to check if a DFA accepts a language and its reverse. Problem 4 builds a decider to determine if a DFA accepts any palindromes. Problem 5 designs a decider that checks if a CFG generates exactly k strings or infinitely many.

Uploaded by

Kamal Walia
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

CS5371 Theory of Computation

Homework 3 (Suggested Solution)

1. Ans. The language {0n 1n 2n | n ≥ 1} is not context-free, so that it cannot be recognized


by a 1-PDA. However, we can easily design a 2-PDA to recognize this language as follows
(Let S1 and S2 be the two stacks in the 2-PDA):
1. For each 0 it reads, push a 0 in S1 and push a 0 in S2
2. For each 1 it reads, pop a 0 from S1
3. For each 2 it reads, pop a 0 from S2
4. If at any step, we discover the input string is not in correct order (e.g., a 0 is read
after 1 is read), we reject the input string
5. If S1 and S2 become just empty at the end, we accept the input string
Thus, we have found a language that can be recognized by some 2-PDA but not by any 1-
PDA. On the other hand, if a language can be recognized by a 1-PDA, it must be recognized
by some 2-PDA. Therefore, 2-PDAs are more powerful than 1-PDAs.
2. Ans. If a language L is decidable, there exists a decider D that decides L. Then, we
can construct an enumerator E that enumerates the strings of L in the desired ordering
(shorter string first, then lexicographical order) as follows:
E = “On any input,
1. Ignore the input
2. For k = 1, 2, . . .
i. Run D on the kth string in Σ∗ , according to the desired ordering
ii. If D accepts, print the string”
Conversely, if some enumerator E enumerates the strings of L in the desired ordering, then
either L is a finite set so that it is decidable, or if L is an infinite set, we can construct a
TM D based on E as follows:
D = “On input w,
1. Run E
i. For every string s printed by E, if s = w, accept w
ii. Else if s < w in the desired ordering, continue
iii. Else if s > w, reject w”
Since at most a finite number of strings of L are smaller than w in the desired ordering,
so after a finite number of strings are printed by E, we can decide if w is in L or not. So,
D runs in finite steps and is thus a decider.
3. Let S = {hM i | M is a DFA that accepts w whenever it accepts the reverse of w}.
(a) Ans. An example of a DFA in S: A DFA that accepts all strings.
(b) Ans. To show S is decidable, we construct a decider D for S as follows (Let C be a
TM that decides EQDF A ):

1
D = “On input hM i,
1. Construct an NFA M 0 such that L(M 0 ) = {wR | w ∈ L(M )}
2. Convert M 0 into an equivalent DFA M 00
3. Use C to compare L(M 00 ) and L(M )
4. If L(M 00 ) = L(M ), accept. Else, reject.
In the above TM, Step 1 can be done by converting M into M 0 in finite steps. The
idea is to (i) reverse the directions of all transition arrows in M , (ii) create a new
state q 0 in M 0 , and connects q 0 to each original final states of M with ε-transitions,
and (iii) make the original start state of M a final state of M 0 . It is easy to check
that L(M 0 ) = {wR | w ∈ L(M )}.
Also, both Step 2 and Step 3 can be done in finite steps, as we learnt from the lectures
(See Notes 4 pages 13–15, and Notes 12 pages 16–17). So, D runs in finite steps and
is thus a decider.

4. Ans. Let P ALDF A = {hM i | M is a DFA that accepts some palindrome}. To show
P ALDF A is decidable, we construct a decider D for P ALDF A as follows (Let K be a
TM that decides ECF G ):

D = “On input hM i,
1. Construct a PDA P such that L(P ) = {w | w is a palindrome}
2. Construct a PDA P 0 such that L(P 0 ) = L(P ) ∩ L(M )
3. Convert P 0 into an equivalent CFG G
4. Use K to check if L(G) is empty.
5. If L(G) is empty, reject. Else, accept.

In the above TM, Step 1 can be done in finite steps. Step 2 is based on Prob 2.18 and can
be done in finite steps. Step 3 is the conversion of PDA into an equivalent CFG, which can
be done in finite steps (See Notes 8, pages 28–34). Step 4 is done in finite steps, because
the decider K can check whether the language of a CFG is empty (For the existence of K,
see Notes 12, pages 20–21). In summary, D runs in finite steps for any input, and is thus
a decider.

5. Ans. Let C be the language

C CF G = {hG, ki | G is a CFG and L(G) contains exactly k strings where k ≥ 0 or k = ∞}

In this problem, we are given a decider D that decides if the language of a CFG is infinite.
Then, we can show that C is decidable, by finding a corresponding decider F as follows:

F = “On input hG, ki,


1. Use D to check if L(G) is an infinite set.
2. There are four cases:
i. If yes, and k = ∞, accept.
ii. If yes, but k 6= ∞, reject.
iii. If no, but k = ∞, reject.
iv. If no, and k 6= ∞, continue.
3. Compute the pumping length p for the grammar G.

2
4. Set count to be 0.
5. For x = 1, 2, . . . , p
i. For all string s with length = x,
Check if s can be generated by G; if so, increment count by 1.
6. If count = k, accept. Else, reject.

In the above TM F , Steps 1–2 correctly answer the case where L(G) is an infinite set, or
k = ∞. So, after Step 2, we only deal with a grammar G whose language is a finite set,
and our task is to check whether the language size is exactly k. To do so, the loop in Step
5 counts all string that can be generated by G, whose length is at most the pumping length
p. Because we know that L(G) is finite, we are sure that no strings of L(G) can be longer
than p. In other words, the value count correctly computes the exact number of strings in
L(G). So, Step 6 can check correctly answer the case when L(G) is finite, and k is finite.
Finally, it is easy to check that each step runs in finite number of steps. Thus, F is a
decider, so that C is a decidable language.

You might also like