Introduction to the Theory of Computation

Homework #1 Solutions

1. (Exercise 0.4) The Cartesian product of two sets A and B, written A × B, is the set of
ordered pairs (a, b) where a ∈ A and b ∈ B. How many elements does A × B have? Write
|A × B| in terms of |A| and |B|.
Solution: |A×B| = |A||B|, since we can choose any element of A and then independently
choose any element of B.
2. Given two sets A and B, let F be the set of functions f : A → B. In other words, each
f ∈ F assigns an element of B to each element of A. What is |F | in terms of |A| and |B|?
Solution: Each function from A to B is a string of length |A| over the alphabet B. Each
such string consists of |A| independent choices, each of which has |B| possibilities. The total
number of such strings is
|B| × |B| × · · · × |B| = |B||A| .
| {z }
|A| times

3. For a set A, the power set P(A) is the set of all subsets of A. Fill in the blank: we can
think P(A) as the set of functions from A into hblanki.
Solution: Specifying a subset S ⊆ A is equivalent to giving you a function F : A → {0, 1}
such that F (a) = 1 if a ∈ S and 0 if a ∈
/ S. Thus B = {0, 1}. Of course, any 2-element set
will do. (We can also see that |B| should be 2 given the answer to the previous question,
since |P(A)| = 2|A| .)
4. A function f : A → B is one-to-one if for all a1 6= a2 , f (a1 ) 6= f (a2 ) — or, to put it
differently, f (a2 ) = f (a2 ) implies a1 = a2 . It is onto if for all b ∈ B, there exists an a ∈ A
(not necessarily unique) such that f (a) = b.
Consider the following functions from N to N, where N = {0, 1, 2, 3, . . .} is the set of
natural numbers (non-negative integers). For each function, state whether it is onto, and
whether it is one-to-one. (Note that 2. and 3. can be easily expressed in terms of x’s binary
digit sequence.)

1. f (x) = x + 1

x + 1 if x is even
2. f (x) =
x − 1 if x is odd

1
 3. f (x) = x/2 (rounded down)

4. f (x) = (x − 4)2

Solution: 1. is 1-1 but not onto; 2. is 1-1 and onto; 3. is onto but not 1-1; and 4. is neither
1-1 nor onto.
5. (Exercise 0.7) A binary relation ∼ is reflexive if x ∼ x for all x, symmetric if x ∼ y if and
only if y ∼ x, and transitive if x ∼ y and y ∼ z implies x ∼ z.
Give examples of relations (from math, life, or anywhere else) which are

1. Reflexive and transitive but not symmetric

2. Transitive and symmetric but not reflexive

3. Symmetric and reflexive but not transitive

Solution (for instance): for 1. let x and y be people and say that x ∼ y if x = y or if x is an
ancestor of y. More generally, let x and y be vertices in a directed graph and say that x ∼ y
if there is a path of length 0 or more from x to y. This is transitive since your ancestor’s
ancestor is your ancestor, or since we can link paths x → y and y → z to make a path x → z.
For 2. let x and y be people and say x ∼ y if they are full brothers (i.e. sharing both
parents) but are not the same person. Or two people living in the same country, etc. This
would be an equivalence relation except that we forbid x ∼ x.
For 3. let x and y be people and say x ∼ y if they live within 1 mile of each other (including
if they are the same person). More generally, let x ∼ y be vertices in an undirected graph
and say x ∼ y if they are neighbors; this is not transitive since your neighbor’s neighbor
might not be your neighbor. (Another cute example is brothers, since your half-brother’s
half-brother might not share either parent with you — or in-laws.)
6. (Exercise 0.11) Find the error in the following proof that all horses are the same color.
Definition: A set of horses is uniform if all the horses in it are of the same color.
Lemma: if U , V , and U ∩ V are uniform, then so is U ∪ V . Proof: obvious.
Base case: a set of one horse is clearly uniform.
Induction step: suppose all sets of less than k horses are uniform. We wish to show that
any set of k horses is also. Let S be a set of k horses where k > 1. Then S is the union of
two sets U, V of size k − 1 (for instance, let U be S minus the first horse, and V be S minus
the second horse). Since U and V are smaller than S, they are uniform by the inductive
hypothesis, and since their intersection is also smaller than S, it is too. By the Lemma
above, U ∪ V = S is also uniform. QED.
Solution: The problem is with the induction step, and specifically with the Lemma
(beware of proofs that say “obvious”!) The null set is uniform (after all, it doesn’t contain
two horses of different colors) but if U and V are disjoint the Lemma is false. The problem
arises when k = 2: we then have two sets, each of size k − 1 = 1, but they can both be
uniform without their unions being uniform since their intersection is empty.

2
 P Q
7. Let (S) be the sumP of the elements
Q in a set, and (S) be their product.
P If S andQ T
are disjoint, what is (S ∪ T ) and (S ∪ T )? Conclude from this what (∅) and (∅)
should be (∅ is the empty set). P P P Q
Q Solution:
Q if S and T are disjoint, clearly (S ∪ T ) = (S) + (T ) and (S ∪ T ) =
(S) · (T ). (By the way, what if they are Pnot disjoint?)
P P Q Q Q
Now, since SP∪ ∅ = S for allQ
S, we have (S) + (∅) = (S) and (S) · (∅) = (S).
It follows that (S) = 0 and (S) = 1 — in other words, the sum of nothing is zero, but
the product of nothing is one!
8. True or false: every element of the empty is set is purple.
Solution: Of course! It contains no non-purple elements. Consider also the following: let
P be the set of purple things and G be the set of green things. Then every element of P ∩ G
is both purple and green. But (assuming for simplicity that each thing is only one color)
P ∩ G = ∅.
Note also that every element of the empty set is named Bob.
9. Suppose I have n cities, and the distances between them are given as n-bit numbers.
How many bits could the length of the shortest tour (or, for that matter, the longest tour)
possibly have?
Now, suppose I use the yes/no oracle for the decision version of the Travelling Salesperson
Problem to find the length of the shortest tour. Show that if I do this with a “linear scan”
(starting at 0 and searching upward) it might take exponential time (i.e. an exponential
number of calls to the oracle), but if I do a binary search it will only take polynomial time.
Remember that when I say “exponential” or “polynomial”, I mean as a function of n.
Solution: since a tour has length n, the longest the tour could be is n times the maximum
length. Since the lengths have n bits the maximum length is 2n , so the longest a tour could
be is Lmax = 2n × n. This number has log2 Lmax = log2 (2n n) = n + log2 n = O(n) bits, only
slightly more than the number of bits each distance has.
Now, if we start out at 0 and search upward towards Lmax linearly, it might take as many
as Lmax = 2n n steps, which is exponential in n. However, a binary search that starts at
Lmax /2 and then goes to Lmax /4 or 3Lmax /4, and so on, will only take log2 Lmax = O(n)
steps, which is polynomial in n.
10. Now that I know the length of the shortest tour, how do I find the shortest tour in a
polynomial number of calls to the oracle? (Hint: I can modify the problem, i.e. the set of
cities, the distances between them, and the desired distance, in any manner I desire.)
Solution: for each edge, increase its length by 1 and use the oracle to recalculate the length
of the new shortest tour. The edges used by the shortest tour are those where increasing
their length increases the length of the tour. (We need to be a little careful if there are
multiple tours of the same length — do you see how
to handle this case?)
n 2
Since a graph with n vertices has at most 2 = O(n ) edges, we just have to repeat
our procedure for finding the length of the shortest tour a polynomial number of times.
Since each time takes a polynomial number of calls to the oracle, and since the product of a
polynomial is a polynomial, the total number of calls to the oracle is still a polynomial (in
fact O(n3 )).