Assignment 5 Solutions, MATH 240

prepared by Jonah Saks

Q1 First, we check if this relation is reflexive. Let (a, b) ∈ Z × (Z \ {0}). We’d like to check if (a, b) ∼ (a, b).
Equivalently, we must verify if ab = ab. This is clearly true for any (a, b) ∈ Z × (Z \ {0}), hence this relation
is reflexive. Next, we check symmetry. Let (a, b), (c, d) ∈ Z × (Z \ {0}) such that (a, b) ∼ (c, d). Then, it
follows that (c, d) ∼ (a, b) since (a, b) ∼ (c, d) ⇒ ad = bc ⇒ cb = da ⇒ (c, d) ∼ (a, b), as desired. Finally,
for transitivity, let (a, b), (c, d), (e, f ) ∈ Z × (Z \ {0}) such that (a, b) ∼ (c, d) and (c, d) ∼ (e, f ). We’d like
to check that (a, b) ∼ (e, f ). From the hypothesis, we have that ad = bc and cf = de. Since d and b are
non-zero, we can divide the first equation by d and the second equation by b to obtain the equations a = bc d
and f = de bc de
c . Combining these equations, we see that af = d · c = be. Equivalently, (a, b) ∼ (e, f ).

Q2, (a) The relation R1 is the relation which says two binary sequences a = (a0 , a1 , a2 , ...) and b =
(b0 , b1 , b2 , ...) are related if, for all n ∈ N, either an or bn is 1. We claim this relation is not reflexive. Indeed,
consider the zero-sequence 0 := (0, 0, 0, 0, 0, ...), ie. the binary sequence which only takes the value 0. It’s
clear that 0 ̸∼R1 0. This proves that R1 is not reflexive. Next, we claim that R1 is symmetric. Indeed, for
any two binary sequences a and b, a ∼R1 b implies that for all n ∈ N, either an or bn is 1. Equivalently,
for all n ∈ N, either bn or an is 1. Hence, it follows that b ∼R1 a. Finally, we claim R1 is not transitive.
Consider the sequence a = (1, 0, 0, 0, ...)) and the sequence of only ones 1 := (1, 1, 1, 1, ...). Then it’s clear
that a ∼R1 1 and 1 ∼R1 0, however it is not the case that a ∼R1 0. Indeed, at the index n = 1, we have
that both sequences (an ) and 0 evaluate to 0.

Q2, (b) The relation R2 is clearly reflexive, since for any infinite binary sequence a, we have that F (a) =
F (a), hence in particular F (a) ≤ F (a), so that a ∼R2 a. We claim this relation is not symmetric. Indeed,
let a = (1, 0, 0, 0, ...) and let 0 denote the sequence only containing 0’s. Then F (a) = 0 and F (0) = ∞, so
we have that F (a) ≤ F (0), but F (0) ̸≤ F (a). In other words, we have that a ∼R2 0 but 0 ̸∼R2 a. This
shows that R2 is not symmetric. Finally, we claim that R2 is transitive. Let a, b, c be three infinite binary
sequences such that a ∼R2 b and b ∼R2 c. Then F (a) ≤ F (b) and F (b) ≤ F (c). If these were all natural
numbers, then we would be done since it would clearly follow that F (a) ≤ F (c) and hence a ∼R2 c. However,
we must be careful since the inequalities are also defined when F takes the value ∞. We can consider two
cases: whether or not F (b) is finite. First, if F (b) = m for some m ∈ N, then F (a) ≤ F (b) implies that
F (a) = n for some n ∈ N such that n ≤ m in the usual sense. Now, no matter if F (c) is finite or infinite, it
follows in any case that n ≤ m and m ≤ F (c) implies n ≤ F (c), as desired. In the case that F (b) = ∞, then
F (b) ≤ F (c) implies that F (c) = ∞. Then, no matter if F (a) is finite or not, we can in any case conclude
that F (a) ≤ F (c), as desired. Hence, in all cases, F (a) ≤ F (b) and F (b) ≤ F (c) implies that F (a) ≤ F (c).
Equivalently, R2 is transitive.

Q2, (c) An intuitive way of describing the relation R3 is that two infinite binary sequences a and b
are related by R3 if a and b differ at only finitely many places. Clearly, a ∼R3 a since an = an for all
n ∈ N, so {n ∈ N : an ̸= an } = ∅, and in particular this set is finite. The relation is symmetric since
a ∼R3 b ⇒ {n ∈ N : an ̸= bn } is finite ⇒ {n ∈ N : bn ̸= an } is finite ⇒ b ∼R3 a. Finally, we claim that the
relation is transitive. Let a, b, c be sequences such that a ∼R3 b and b ∼R3 c. Notice that:

{n ∈ N : an ̸= cn } ⊂ {n ∈ N : an ̸= bn } ∪ {n ∈ N : bn ̸= cn }

This follows because, for any n ∈ N, if an ̸= cn , then it cannot be that an = bn and bn = cn , (since this
would imply an = cn , a contradiction). Hence, it must be that either an ̸= bn or bn ̸= cn . Since this is

1
true for any n ∈ N, this proves the inclusion above. Finally, transitivity follows from this since a ∼R3 b
and b ∼R3 c means that the two sets on the right-hand side of the inclusion above are both finite. Hence,
{n ∈ N : an ̸= cn } must be finite, so a ∼R3 c.

Q3, (a) Since gcd a, b = 1, then we can apply Bézout’s Theorem to find integers r′ and s′ such that
ar′ + bs′ = 1. Multiplying the whole equation by n, we get anr′ + bns′ = n, so we can simply define r = nr′
and s = ns′ . Clearly, these are integers which satisfy ar + bs = n.

−s r


Q3, (b) We will use the hint! The length of the interval a , b on the real line is given by:

r −s r s ra + sb n
− = + = =
b a b a ab ab
n
Then, since n > ab, we know that ab > 1. Hence, the length of this interval on the real line is greater than
one. It’s clear that any interval on the real line with length > 1 must contain an integer. It follows that the
interval −s r


,
a b must contain an integer c. Equivalently, there must exist an integer c satisfying:

−s r
<c<
a b

Q3, (c) Starting with our inequality we obtained above, we can deduce the two inequalities −s < ac
and cb < r. Equivalently, 0 < s + ac and 0 < r − cb. Define u := s + ac and t := r − cb. These inequali-
ties we’ve obtained show that u and t are positive. Moreover, u and t are integers because they are linear
combinations of integers. Finally, we claim that u and t satisfy the equation ta + ub = n. Indeed:
(a)
ta + ub = (r − cb)a + (s + ac)b = ra − cba + sb + acb = ar − abc + bs + abc = ar + bs = n,

where the last equality follows from part (a).