Name Asu Id: CSE 355 Theory of Computing ID: - Spring 2012 - Colbourn Midterm # 1 Page 1 of 7 16 February 2012

CSE 355
Spring 2012 - Colbourn
Theory of Computing
Midterm # 1
ID: ___________
Page 1 of 7
16 February 2012
IMA SAMPLE
NAME
ASU ID
You have one hour and 15 minutes to complete the exam.

Do not open the exam until instructed to do so.
No notes, texts, computers, calculators, or communication devices are permitted.
Write all answers on the examination paper itself.
BUDGET YOUR TIME WELL!
SHOW ALL WORK!
Question 1
[13]
Question 2
[10]
Question 3
[8]
Question 4
[9]
Question 5
[10]
Total
[50]
Bonus question [1 mark]: What is the language *? It is {} (Note that * = .)
Question 1. [13 marks total]

Let G be the grammar with start variable S, terminals {0,1}, and productions:
S 0A | 0M
A 0A | 1B |
B 0C
C 0A
M 1N
N 1P
P 1P | 0M |
Call its language L.
(a) [3 marks] Using grammar G and the algorithms developed in class, give an NFA for L.
(Describe briefly the steps that you follow.)
CSE 355
Theory of Computing
Midterm # 1
ID: ___________
Page 2 of 7
I make every variable a state, and make the start state be the start variable and the final states be
every variable X for which X . Every production X aY makes a transition from X to Y
labeled a.
So I get an NFA with states Q={S,A,B,C,M,N,P}, start state S, final states {A,P}, = {0,1}, and
transition function defined by (S,0) = {A,M}, (A,0) = {A}, (A,1) = {B}, (B,0) = {C}, (C,0)
= {A}, (M,1) = {N}, (N,1) = {P}, (P,0) = {M}, (P,1) = {P}. When (X,a) is not defined
here, it is .
(It is probably easier to draw a transition graph, and most people did.)
(b) [5 marks] Using your answer to (a) and the algorithms developed in class, give a DFA for
L. (Describe briefly the steps that you follow.)
My DFA has states corresponding to sets of states of the NFA from the (a) part. For each set of
states that I encounter and each symbol in , I make a transition to the set of all states that the
NFA can get to by reading the specified symbol. I start with the state {S}. I determine whether a
state is final according to whether its set contains a final state of the NFA.
State of DFA
Transition on 0 to state
Transition on 1 to state

{S}
{A,M}

{A,M} final!
{A} final!
{B,N}
{B}
{C}
{P} final!
{M}
{N}
{A}
{A}
{C}
{C}
{A}
{M}
{B,N}
{B}
{P}

{P}
{N}
{P}
(c) [5 marks] Using your answer to (a) and the algorithms developed in class, give a regular
expression for L. (Describe briefly the steps that you follow.)
First I modify the NFA from (a) by giving it a single final state F. To do this, I add transitions
(A,) = {F} and (P,) = {F}; make F a final state; and make A and P non-final. Now I
repeatedly eliminate a non-start, non-final state by figuring out all the ways to enter and leave it
and making a transition labeled with the corresponding regular expression.
To start, eliminate B and the two transitions involving it, and add the transition (A,10) = {C}.
Then eliminate C, and replace the transition (A,0) = {A} by (A,0+100) = {A}.
Then eliminate N and the two transitions involving it, and add the transition (M,11) = {P}.
Then eliminate M and the three transitions involving it, add (S,011) ={P}, and replace (P,1) =
{P} by (P,1+011) = {P}.
Finally eliminate A and P to get a two-state NFA with (S,0(0+100)* + 011(1+011)*) = {F}.
Now read out the regular expression, 0(0+100)* + 011(1+011)*.
There are many reduction orders, and the answer can look different.
CSE 355
Theory of Computing
Midterm # 1
ID: ___________
Page 3 of 7
Question 2. [10 marks]

(a) [3 marks] Show that if L is a regular language, then the reverse of L, LR, is also a regular
language.
Many solutions are possible. Here is one. Because L is regular, L has an NFA. By adding transitions, we can make an NFA that has exactly one final state. Now reverse all of the
transitions, make the start state final and the final state the start state. This new NFA accepts the
reverse of L, so it is regular.
(b) [5 marks] Suppose that for a language L, prefix(L) = {w : wy L}. Show that if L is
regular, then prefix(L) is also regular.
Many solutions are possible. One way uses DFAs, one way uses regular expressions. I will show
the second. Because L is regular, it has a regular expression r.
If r = , prefix(r) = . If r = , prefix(r) = . If r = a, prefix(r) = a+.
If r = r1+r2, prefix(r) = prefix(r1) + prefix(r2).
If r = r1r2, prefix(r) = r1prefix(r2) + prefix(r1).
If r = (r1)*, prefix(r) = (r1)*prefix(r1).
Because prefix(r) has a regular expression, it represents a regular language.
Note: this does NOT follow from closure under concatenation.
(c) [2 marks] Use your answers to the (a) and (b) parts to show that if L if regular then
suffix(L) = {w : yw L}is also regular.
suffix(L) = {wR : w prefix(LR)} = (prefix(LR))R.
Use closure under reversal and prefix to get closure under suffix.
Question 3 [8 marks]
(a) [2 marks] Define each of the following types of grammar: left linear; right linear; regular;
linear.
left-linear: all productions of the form A Bw or A w with A,B V, w T*.
right-linear: all productions of the form A wB or A w with A,B V, w T*.
regular: the grammar is either right-linear or left-linear.
linear: all productions of the form A wB, A Bw, or A w with A,B V, w T*.
(b) [2 marks] If L has a regular grammar, is L (always) a regular language? Explain.
Yes. If the grammar is right-linear we can convert it to an NFA (as in Question 1(a)).
If it is left-linear, reversing each right hand side gives a right-linear grammar for the reverse of
the language, and we have closure under reversal, so it is again regular.
(c) [4 marks] If L has a linear grammar, is L (always) a regular language? Explain.
No. Here is a linear grammar: S aB | , B Sb. Its language is {anbn : n 0}, which we
know is not regular.
CSE 355
Theory of Computing
Midterm # 1
ID: ___________
Page 4 of 7
It is not enough to just say that one cannot make an NFA or DFA.
(a) [2 marks] Build a DFA to accept the language of strings over {a,b} that start and end
with an a.
I allow the case that there is only one letter, and then the first and the last letters are the
same.
State
Transition on input a to
Transition on input b to
q0 start state
q1
qdead
qdead
qdead
qdead
q1 final state
q1
q2
q2
q1
q2
(b) [7 marks] Let L be the language of strings over {a,b} that start and end with an a and
contain at least as many as as bs. Is L regular? Answer yes or no, and show that your
answer is correct.
No! The easiest is to use pumping, but let me use a different argument. Suppose to the contrary
that there is a DFA for L with start state q0 and transition function . Lets define an infinite
sequence of states by qi+1 = (qi,a) for all i 0. By the pigeonhole principle, there must be an i
and j with i < j for which qi = qj.
Because ajb j+1a is in L, it follows that *(qj, b j+1a) is a final state.
But qi = qj so *(qi, b j+1a) is a final state, and hence aib j+1a is in L.
This is a contradiction because i < j.
(Arguments that just say not enough memory, while true, are not convincing at all. Arguments
that presume that the numbers of as and bs are the same are not correct.)
CSE 355
Theory of Computing
Midterm # 1
ID: ___________
Page 5 of 7

(a) [5 marks] Outline a general method for converting a regular expression with language L
into a regular expression whose language is the complement of L.
1. Convert the regular expression to an NFA.
2. Convert the NFA to a DFA.
3. Swap final and nonfinal states to get a DFA M for the complement of L. (Note: you need
this step you cannot do this on an NFA!)
4. Change the DFA so that it has only one final state this may make it an NFA M.
5. Eliminate non-start non-final states one at a time in M (as in 1(c)) to recover an r.e. for
the complement.
(b) [5 marks] Apply your method to the regular expression (ab + ba)* to produce a regular
expression for its complement.
First I make an NFA and simplify it and convert to a DFA to get
State
Transition on input a to
Transition on input b to
q0 start state and final state
q1
q2
q1
q3
q0
q2
q0
q3
q3
q3
q3
Then I swap final and nonfinal states and add a new single final state to get the NFA:
State
Transition on a to
Transition on b to
Transition on to
q0 start state
q1
q2
q1
q3
q0
qf
q2
q0
q3
qf
q3
q3
q3
qf
qf final state
And now I eliminate the three states that are non-start, non-final, one at a time, to get:
(ab+ba)*(a + b + (aa+bb)(a+b)*)
Again, the specific NFA that you start with, and the reduction sequence, can make the answer
look different.
Check for yourself: If you do not convert the NFA to a DFA before swapping final and non-final
states, you may not get the complement of the language!

Name Asu Id: CSE 355 Theory of Computing ID: - Spring 2012 - Colbourn Midterm # 1 Page 1 of 7 16 February 2012

Uploaded by

Copyright:

Available Formats

Name Asu Id: CSE 355 Theory of Computing ID: - Spring 2012 - Colbourn Midterm # 1 Page 1 of 7 16 February 2012

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

Name Asu Id: CSE 355 Theory of Computing ID: - Spring 2012 - Colbourn Midterm # 1 Page 1 of 7 16 February 2012

Uploaded by

Copyright:

Available Formats

CSE 355

Spring 2012 - Colbourn

You have one hour and 15 minutes to complete the exam.

Question 1. [13 marks total]

Question 2. [10 marks]

Question 5 [10 marks]

You might also like