Solutions to Exercises in Discrete Mathematics

by: Norman L. Biggs: 2nd Edition 2002

4.1 The ‘laws of algebra’

4.1.1 Construct proofs of the following rules, using only

Axioms 1-9.
1 × n = n, (a + b)c = ac + bc.

Discussion We list each axiom, by number and by its common

name, as it is used:
• for example, associativity of × (Axiom 6)
tells us that, given a, b, c ∈ N, we get the same answers whether we
calculate a × b, and then multiply the answer from the right by c or
whether we calculate b × c, and then multiply the answer from the
left by a. Other stages in the reasoning are similarly indicated.
It is worth realising there are often several different valid proofs
that a given statement is True.

Solution One valid proof of the first statement is:

1×n = n×1 Axiom 5 commutativity of ×
so
1×n = n×1 = n Axiom 7 definition of multiplicative identity

A valid proof of the second statement is:

(a + b) × c = c × (a + b)
Axiom 5 commutativity of ×
so (a + b) × c = c × (a + b) = (c × a) + (c × b)
Axiom 9 left distributive law ;
but (c × a) = (a × c) and (c × b) = (b × c)
Axiom 5 commutativity of ×;
so (a + b) × c = c × (a + b) = (a × c) + (b × c) = ac + bc.

1
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.1.2 Show that if x is a multiple of y and y is a multiple of z then

x is a multiple of z.

Solution If x is a multiple of y then x = ry for some r ∈ N,

and if y is a multiple of z then y = sz for some s ∈ N, so that
x = r(sz) = (rs)z and is a multiple of z (using Axiom 6)
Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.1.3 N can be enlarged by introducing a new symbol 0, represent-

ing the ‘zero’ that we all know. 0 is assumed to satisfy the rules
0 + a = a, 0 × a = 0,
as well as Axioms 1-9 with one exception. What is this exception?

Discussion This is an example of an important aspect of abstract

mathematics: looking at a collection of axioms, and seeing which of
them do, or do not, apply to particular mathematical objects.

Solution One of the familiar conclusions of arithmetic is that ‘you

can’t cancel by nought (zero)’, whereas of course you can cancel
by non-zero numbers. Inspecting the list of Axioms, we see that
Axiom 8 says that in N we can cancel on the right by any z ∈ N
(and of course Axiom 5 then means we can cancel on the left by z
as well). From the second rule, it is clear that the symbol 0 cannot
satisfy Axiom 8, so that is the exception.

2
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.2 Putting the natural numbers in order

4.2.1 Using only the definitions and axioms given above, prove
that, for all natural numbers s, x, t,
if s > x and x > t then s > t.

Discussion Recall that, from the definitions in Section 4.2, if m, n ∈

N then m < n means that m + x = n for some x ∈ N, and n > m
means the same as m < n. There is a (very) brief solution using
Theorem 4.2 and the definition of > in terms of <, but the fuller
version better illustrates the logic of the arguments involved.

Solution Since s > x (i.e., x < s) there is some a ∈ N such that

x + a = s, and similarly there is some b ∈ N such that t + b = x.
It follows that (t + b) + a = s, so t + (b + a) = s (Axiom 4); since
(b + a) ∈ N (Axiom 1), it follows (definition of ,) that t < s, i.e.
s > t.

5
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.2.2 Prove that, for all m, n, r ∈ N, if m < n then m + r < n + r.

Solution Since m < n there is an element s ∈ N such that m + s =

n. It follows that (m + s) + r = n + r and (m + s) + r = m + (s + r).
Also (s + r) ∈ N (Axiom 1) and s + r = r + s (Axiom 3) so that
m + (s + r) = m + (r + s) = (m + r) + s (also using Axiom 4), and
hence (m + r) + s = (n + r), which means m + r < n + r.

6
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.2.3 Use the result of the preceding exercise, and Axiom 10, to
prove that if m + x = n + x for some natural numbers m, n, x, then
m = n.
Discussion Recall that Axiom 10 (from Section 4.2) states that, if
m, n ∈ N then exactly one of the following three statements is true:
m < n, m = n, n<m

Solution According to Ex. 2, if m < n then m + x < n + x, and if

n < m then n + x < m + x. According to Axiom 10, exactly one of
the three statements:
m + x < n + x, m+x = n+x and n+x < m+x
must be true. If if were the case that m < n then the true statement
would have to be m + x < n + x, and if were the case that n < m
then the true statement would have to be n + x < m + x. Since we
know that m + x = n + x, and (Axiom 10) that exactly one of the
three statements:
m < n, m = n, n<m
is true, the only possibility is m = n, which is therefore proved.

7
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.3 The principle of induction

4.3.1 Use the method of proof by induction to show that, for all
n ∈ N, n2 + n is an even number. (Another proof was given in
Section 1.5.)

Discussion When first using the method of proof by induction, it is

useful to spell out the various steps (see Ex. 2 and the solutions to
Ex. 1 and Ex. 2). The difficult step, in the terms of the description
of induction at Axiom 11, is usually in proving that P (k) implies
P (k + 1), although that notation is often unnecessary. In some cases
the formulation of the statements in the family {P (n) | n ∈ N} can
also require some thought: Example 2 in Section 4.4 illustrates this
potential difficulty.

Solution We wish to show that, for all n ∈ N, n2 + n is an even

number.
Induction basis If n = 1 then n2 + n = 2, so the statement is true
when n = 1.
Induction hypothesis k ∈ N and k 2 + k is an even number.
Induction step (k + 1)2 + (k + 1) = k 2 + 2k + 1 + k + 1 = (k 2 + k) +
2(k +1). By the induction hypothesis, k 2 +k = 2m for some m ∈ N,
so (k + 1)2 + (k + 1) = 2(m + k + 1), and is therefore also even.
So, by the principle of induction, for all n ∈ N, n2 + n is an even
number.

12
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.3.2 Use the principle of induction to prove the following state-

ment, identifying clearly the induction basis, the induction hypoth-
esis and the induction step.
for all n ∈ N, n3 + 3n2 + 2n is a multiple of 6.

Solution Induction basis If n = 1 then n3 + 3n2 + 2n = 6, so the

statement is true when n = 1.
Induction hypothesis k ∈ N and k 3 + 3k 2 + 2k is a multiple of 6.
Induction step (k + 1)3 + 3(k + 1)2 + 2(k + 1) =
(k 3 + 3k 2 + 3k + 1) + 3(k 2 + 2k + 1) + 2(k + 1) =
(k 3 + 3k 2 + 2k) + 3(k 2 + 1)2 + 3(k + 1).
By the induction hypothesis, (k 3 + 3k 2 + 2k) = 6m for some m ∈ N,
so:
(k + 1)3 + 3(k + 1)2 + 2(k + 1) = 6m + 3(k + 1)(k + 2).
Since one of k+1, k+2 must be even, this is therefore also a multiple
of 6.
So, by the principle of induction, for all n ∈ N, n3 + 3n2 + 2n is
a multiple of 6.

13
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.4 Summation formulae

4.4.1 Use the principle of induction to prove that

1
12 + 22 + · · · + n2 = 6
n(n + 1)(2n + 1)
for all natural numbers n.

Discussion There are two Examples of how to approach this type

of problem in Section 4.4, and further remarks in the Discussion
of Ex. 2.

Solution Induction basis If n = 1 then the left hand side (LHS)

is 1, and the RHS is (1 × 2 × 3)/6 = 1, so the formula is true for
n = 1.
Induction hypothesis k ∈ N and
12 + 22 + · · · + k 2 = 16 k(k + 1)(2k + 1).
Induction step It follows that
12 + 22 + · · · + k 2 + (k + 1)2 = 61 k(k + 1)(2k + 1) + (k + 1)2 =
(k + 1){k(2k + 1) + 6(k + 1)}/6 =
(k + 1){2k 2 + k + 6k + 6}/6 =
(k + 1){(k + 2)(2k + 3)}/6,
which is the required RHS in the case that n = k + 1.
So, by the principle of induction, for all n ∈ N,
12 + 22 + · · · + n2 = 16 n(n + 1)(2n + 1).

14
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.4.2 Use the method of induction to show that the following state-
ment is true for all natural numbers n:
Pn 1
r=1 r(r + 2)(r + 4) = 4 n(n + 1)(n + 4)(n + 5).

Discussion This example illustrates the methods of:

• substituting n = k into the formulae on each side of the
equality;
• adding the next term of the summation to one side;
• manipulating the new expression to obtain the n = k + 1
version of the other side.
The principle of induction then yields that the required equality is
true for all n ∈ N.

Solution Induction basis If n = 1 then the LHS is 15 and the

RHS is (1 × 2 × 5 × 6)/4, that is 15, so the statement is true when
n = 1.
Induction hypothesis k ∈ N and
Pk 1
r=1 r(r + 2)(r + 4) = 4 k(k + 1)(k + 4)(k + 5).

Induction step When n = k + 1 the LHS is

{k(k + 1)(k + 4)(k + 5) + 4(k + 1)(k + 1 + 2)(k + 1 + 4)}/4 =

(k + 1)(k + 5){k(k + 4) + 4(k + 3)}/4 =
(k + 1)(k + 5){k 2 + 8k + 12}/4 =
(k + 1)(k + 1 + 4)(k + 2)(k + 6)/4 =
(k + 1)(k + 1 + 1)(k + 1 + 4)(k + 1 + 5)/4,

which is the required RHS in the case that n = k + 1.

So, by the principle of induction, for all n ∈ N,
Pn 1
r=1 r(r + 2)(r + 4) = 4 n(n + 1)(n + 4)(n + 5).

15
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.5 Recursive definitions

4.5.1 Calculate the values of u1 , u2 , u3 , u4 and u5 given by the

recursive definition
u1 = 1, u2 = 1, un+1 = un + 2un−1 (n ≥ 2).

Discussion This is useful practice: trying the first few values is very
often (though not always!) a useful way of getting a ‘feel’ for the
patterns involved in recursive definitions.

Solution The question states that u1 = 1 and u2 = 1, so that

u3 = 1 + 2 = 3, u4 = 3 + 2 × 1 = 5 and u5 = 5 + 2 × 3 = 11.

19
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.5.2 Write down formulae for the numbers un defined by the fol-
lowing equations.
(i) u1 = 1, un+1 = un + 3 (n ≥ 2).
2
(ii) u1 = 1, un+1 = (n + 1) un (n ≥ 2).

Discussion Again, a sensible initial approach is to try to find out

what is going on, either by calculating the first few values or by
some other method. By a formula is meant an expression into which
values of n can be substituted, in order to obtain the relevant value
of un .

Solution (i) The first few values are 1, 4, 7 and 10, and two ver-
sions of the correct formula are un = 3n − 2 and un = 3(n − 1) + 1.

(ii) The first few values are 1, 4, 36 and 36 × 42 = 576, and the
formula is un = (n!)2 .

20
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.5.3 Natural numbers qn are defined by the rule:

q1 = 2, qn+1 = qn2 − qn + 1.
Find a counter-example to the statement that qn is a prime for
all n ≥ 1.

Discussion When finding a counter-example to a claim like this, the

method of calculating successive values for qn is typically the best
approach. In this instance, it also helps to have a clear idea how to
recognise whether or not a natural number is prime!

Solution The first few values are: q1 = 2 (a prime); q2 = 4−2+1 =

3 (prime); q3 = 9 − 6 + 1 = 7 (prime); q4 = 49 − 7 + 1 = 43 (also
prime); q5 = 432 − 43 + 1 = 1807. Testing with a calculator, it turns
out that 1807 = 13 × 139, so it not prime: q5 = 1807 is therefore a
suitable counter-example.

21
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.6 Other forms of the principle of induction

4.6.1 In the following cases find the appropriate values of n0 and

show that the statement is true for all n ≥ n0 .
(i) n2 − 6n + 8 ≥ 0; (ii) n3 ≥ 6n2 .

Discussion Both parts can be approached along the lines of Exam-

ple 1 in Section 4.6. It is probably useful to try to identify n0 by
successively substituting the values n = 1, n = 2, . . . until one can
make a sensible judgment.

Solution (i) Let f (n) = n2 − 6n + 8. The first few values of f (n)

are: f (1) = 3; f (2) = 0; f (3) = −1; f (4) = 0 and f (5) = 3.
This suggests that n0 = 4, and (writing out induction arguments
less formally than previously) we have:
f (4) ≥ 0;
Suppose that k ∈ N, k ≥ 4 and f (k) ≥ 0. Then
f (k + 1) = (k + 1)2 − 6(k + 1) + 8 =
(k 2 − 6k + 8) + (2k + 1 − 6) = f (k) + (2k − 5).
Since k ≥ 4 we know that 2k−5 ≥ 3 > 0, so that f (k+1) > f (k) ≥ 0
(again using ‘transitivity’, as discussed in Ex. 4.2.5), and the result
follows by the principle of induction.
(ii) Let g(n) = n3 −6n2 ; the statement we are required to prove is
equivalent to g(n) ≥ 0 for n ∈ N and n ≥ n0 . It is clear that g(n) =
n2 × (n − 6), which immediately suggests that n0 = 6. Certainly
g(6) ≥ 0, so suppose that k ∈ N, that k ≥ 6 and that g(k) ≥ 0.
Then g(k + 1) = (k + 1)2 × (k + 1 − 6), so that g(k + 1) = g(k) +
k 2 + (2k + 1) × (k − 5) ≥ g(k) since k > 6 means k − 5 > 0.

23
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.6.2 Find the least number n0 for which it is true that n! ≥ 2n .

Taking the case n = n0 as the induction basis, show that the state-
ment is true for all n ≥ n0 .

Discussion As in Ex. 4.6.1, it helps to find a function (h(n)) such

that the given statement is equivalent to h(n) ≥ 0. To illustrate
how the original version of the principal of induction works in these
modified circumstances, the solution which follows, like those for
Section 4.4, explicitly identifies the the successive stages of the in-
duction argument.

Solution Let h(n) = n! − 2n , so that the given statement is equiva-

lent to h(n) ≥ 0. The first few values of h(n) are: h(1) = 1−2 = −1;
h(2) = 2 − 4 = −2; h(3) = 6 − 8 = −1; h(4) = 24 − 16 = 8, so
that 4 is the obvious candidate for the role of n0 .

Now the claim: if n ∈ N and n ≥ 4 then n! ≥ 2n ;

is equivalent to: if n ∈ N and n ≥ 4 then h(n) ≥ 0,
and to: if r ∈ N then h(r + 3) ≥ 0,
which is the version we shall prove.
Induction basis 1 ∈ N and h(4) = 8, so the statement is true for
r = 1.
Induction hypothesis k ∈ N and h(k + 3) ≥ 0.
Induction step h(k + 1 + 3) = (k + 4)! − 2k+4 =
(k + 4) × (k + 3)! − 2 × 2k+3 ≥
2 × {(k + 3)! − 2k+3 } =
2 × h(k + 3),

and 2 × h(k + 3) ≥ 0 by the inductive hypothesis, so h(k + 4) ≥ 0,

and the r = k + 1 version of the statement is true. ‘Translating’
back to the original problem, by the principle of induction
n! ≥ 2n for all n ≥ n0 = 4.

24
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.6.3 Use the strong form of the principle of induction to show that
if un is defined recursively by the rules

u1 = 3, u2 = 5, un+1 = 3un − 2un−1 (n ≥ 2),

then un = 2n + 1 for all natural numbers n.

Discussion The ‘strong principle of induction’ is discussed just be-

fore Example 1 in Section 4.6. Example 2 in that section illustrates
the method to be used.

Solution We need a (formally) stronger type of Induction hypoth-

esis:

if k, r ∈ N, k ≥ 2 and 1 ≤ r ≤ k then ur = 2r + 1.

Because the induction hypothesis has a modified form, we need a

different type of Induction basis:
u1 = 3 = 21 + 1 and u2 = 5 = 22 + 1;
hence the induction hypothesis is true in the case k = 2. We then
have the
Induction step Given the induction hypothesis, uk+1 = 3 × (2k +
1) − 2 × (2k−1 + 1) = (3 × 2k − 2k ) + (3 − 2) = 2k+1 − 1, and the
initial claim therefore holds in the case n = k + 1. By the strong
principle of induction, un = 2n + 1 for all n ∈ N.

25
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.7 Greatest and least members

4.7.1 Suppose that there is a greatest member of N: call it g.

What can we say about g + 1? Write out this argument as a Proof
by Contradiction of the statement that N has no greatest member.

Discussion The definition of m < n, for m, n ∈ N, is at the start

of Section 4.2: m < n if m + x = n for some x ∈ N. This implies
that g < g + 1, contradicting the assumption that g is the greatest
member of N. The following solution is presented fairly formally,
and in the ‘Theorem, Proof’ style.

Theorem The natural numbers N do not contain a greatest

element g.

Proof (by Contradiction) Suppose that g ∈ N is the greatest

element of N. By Axioms 1 and 7, g + 1 ∈ N, and by the definition
of <, g < g + 1, which contradicts the assumption that g is the
greatest element of N. It follows that the natural numbers N do not
contain any such greatest element.

Remark Proof by Contradiction was discussed in Section 3.5, and

in the notation of the end of that section we have shown that
(¬s) ⇒ ¬(¬s), and hence (¬s) ⇒ ((¬s) ∧ (¬(¬s)), where s is the
statement in the theorem.

28
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.7.2 For each of the following sets find the least member, and
find the greatest member if it exists.
{n ∈ N | n2 ≤ 38},
{n ∈ N | n is a multiple of 3},
{n ∈ N | n2 ≤ 100n},
{n ∈ N | n is a prime}.

Discussion To find least members, as efficient a method as any

is to guess a plausible value, and then use trial and error: for non-
empty sets, the existence of a least element is assured (Theorem 4.7).
Deciding whether or not a set has a greatest element uses general
mathematical skills; if the set has a largest element, it is usually
possible to find a plausible value, and then use trial and error until
it is definitely identified.

Solution For {n ∈ N | n2 ≤ 38}, it is easy to list the elements:

{n ∈ N | n2 ≤ 38} = {1, 2, 3, 4, 5, 6}, using 72 = 49 > 38. The least
and greatest elements are 1 and 6 respectively.
For {n ∈ N | n is a multiple of 3}, it is clear that:
{n ∈ N | n is a multiple of 3} = {3, 6, 9, . . .},
and that the least element is 3 but that there is no greatest element
(ultimately, because 3 × (g + 1) > 3 × g for any g ∈ N).
For {n ∈ N | n2 ≤ 100n}, it is clear that the first few natural
numbers, 1, 2, 3, . . . are all in the set, and that this pattern continues
until we get to the vicinity of n = 100. Now 992 < 100 × 99 and
1002 = 100 × 100; it follows that 99 and 100 are in the set. On the
other hand 1012 > 100 × 101, so 101 is not in it, and clearly the
same applies to all larger natural numbers. Hence the least element
of the set is 1 and the greatest is 100.
For {n ∈ N | n is a prime}, the smallest element is clearly 2,
while Theorem 1.6 shows that the set of primes has no greatest
element.

29
 Solutions to Exercises in Discrete Mathematics
by: Norman L. Biggs: 2nd Edition 2002

4.7.3 Suppose that subsets X1 , X2 of N both have least members,

denoted by min X1 and min X2 respectively, and greatest members,
denoted by max X1 , max X2 . What can you say about the min and
max of X1 ∪ X2 and X1 ∩ X2 ?

Discussion This is a useful exercise in understanding the relation-

ships between subsets of N, and is not as easy as it sounds. For
example, do we know whether min X1 ∪ X2 or max X1 ∩ X2 exists?
Many people find it helpful to ‘translate’ the arguments used into
words (of ordinary language).

Solution Since X1 , X2 have least elements neither is empty, and

therefore X1 ∪ X2 is non-empty too, so by Theorem 4.7 X1 ∪ X2 also
has a least element.
This can also be shown directly, by a method which in addition
identifies min X1 ∪ X2 . Let x = min {min X1 , min X2 } and let
y ∈ X1 ∪X2 . If y ∈ X1 then x ≤ min X1 ≤ y and if y ∈ X2 \X1 then
x ≤ min X2 ≤ y, so by the transitivity of ≤, x ≤ y in either case.
But clearly x ∈ X1 ∪ X2 , so min X1 ∪ X2 = min {min X1 , min X2 }.
A similar argument (with the inequalities reversed) shows that
max X1 ∪ X2 exists and is equal to max {max X1 , max X2 }.
The case of X1 ∩ X2 is more complicated: for instance, if X1 =
{1, 3, 5, 7, 9} and X2 = {2, 4, 6, 8, 10} then X1 ∩ X2 = ∅, and there-
fore has neither a least nor a greatest element. So suppose that
X1 ∩ X2 6= ∅, and let z be its least element (Theorem 4.7). Clearly
z ≥ min X1 , since z ∈ X1 , and similarly z ≥ min X2 , so z =
min X1 ∩ X2 ≥ max {min X1 , min X2 }.
However, if X1 = {1, 3, 5} and X2 = {2, 5} then min X1 = 1 and
min X2 = 2 but X1 ∩ X2 = {5}, so min X1 ∩ X2 = 5, which is
strictly greater than max {min X1 , min X2 } = 2.
When X1 ∩ X2 6= ∅, it is easy to show that it is a finite set,
and then the elements can be listed in increasing order, so there
must be a maximal element, say w. Again with the inequalities

30
reversed, a similar argument to that used for min X1 ∩ X2 shows
that w ≤ min {max X1 , max X2 }.

Remark It is not difficult to devise an example, analogous to that

already given for min X1 ∩ X2 , to show that the strict inequality
w = max X1 ∩ X2 < min {max X1 , max X2 } can be satisfied.

31