lOMoARcPSD|54546769

Compsci 120 coursebook

Computer Science 120 Mathematics for Computer Science (University of Auckland)

Scan to open on Studocu

Studocu is not sponsored or endorsed by any college or university

Downloaded by Raven Ebony (ravenebony133@gmail.com)
 lOMoARcPSD|54546769

COMPSCI 120
Mathematics for Computer Science

Above: How to tile a 16 × 16 grid minus its top-right corner with shapes.

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Contents

How To Read This Coursebook 2

Expected Background Knowledge 4

0.1 What Does Compsci 120 Expect? . . . . . . . . . . . . . . . 4

Numbers 6
1.1 Integers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 Even and Odd Integers . . . . . . . . . . . . . . . . . . . . . 8
1.3 Divisibility and Primes . . . . . . . . . . . . . . . . . . . . . 11
1.4 Modular Arithmetic . . . . . . . . . . . . . . . . . . . . . . . 14
1.5 % and Arithmetic . . . . . . . . . . . . . . . . . . . . . . . . 18
1.6 Other Number Systems . . . . . . . . . . . . . . . . . . . . . 22
1.7 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 24

Sets and Strings 26

2.1 Strings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2 Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 35

Combinatorics and Probability 36

3.1 How to Count . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.2 Ordered Choice . . . . . . . . . . . . . . . . . . . . . . . . . . 40
3.3 Unordered Choice . . . . . . . . . . . . . . . . . . . . . . . . 42
3.4 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 47

Algorithms and Functions 50

4.1 Functions in General . . . . . . . . . . . . . . . . . . . . . . . 50
4.2 Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.3 Recursion, Composition, and Algorithms . . . . . . . . . . 59
4.4 Runtime and Algorithms . . . . . . . . . . . . . . . . . . . . 63
4.5 Comparing Runtimes: Limits . . . . . . . . . . . . . . . . . 68
4.6 Limit Techniques and Heuristics . . . . . . . . . . . . . . . . 70
4.7 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 74

Graphs 76
5.1 Graph Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.2 Graphs: Useful Concepts . . . . . . . . . . . . . . . . . . . . 79
5.3 Walks and connectedness . . . . . . . . . . . . . . . . . . . . 81
5.4 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 85

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Trees 86
6.1 Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
6.2 Useful Results on Trees . . . . . . . . . . . . . . . . . . . . . 88
6.3 Rooted Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
6.4 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 92

Proofs 94
7.1 Proofs: Motivation and Fundamentals . . . . . . . . . . . . 94
7.2 Direct Proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
7.3 Proof by Cases . . . . . . . . . . . . . . . . . . . . . . . . . . 99
7.4 Proof by Contradiction . . . . . . . . . . . . . . . . . . . . . 101
7.5 Common Contradiction Mistakes: Not Understanding Nega-
tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
7.6 Proof by Construction . . . . . . . . . . . . . . . . . . . . . . 106
7.7 Proof by Induction: First Examples . . . . . . . . . . . . . 109
7.8 Induction: Two Recurrence Relations . . . . . . . . . . . . 113
7.9 Induction, Graphs, and Trees . . . . . . . . . . . . . . . . . 114
7.10 Proof Methods: How to Choose . . . . . . . . . . . . . . . . 116
7.11 Practice Problems . . . . . . . . . . . . . . . . . . . . . . . . 120

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

How To Read This Coursebook

2019 Chapter -1

Welcome to Compsci 120: Mathematics for Computer Science! We’re

excited to have you in this paper, and cannot wait to show you some
of the fundamental mathematics that underpins the field of computer
science.
This coursebook is possibly a bit different to ones you have seen in classes
before. It has the following features:
❼ Each chapter opens up with a set of “motivating problems.” These
are exercises and puzzles related to the theme of each chapter,
chosen to illustrate how the mathematical content covered in this
paper can be used to solve “real-world” tasks.
After reading these problems, try pausing for a bit and solving
them on your own before going on to the rest of the chapter! Doing
so will help you understand the concepts in each section better
than immediately reading through everything. (It will also help
you appreciate the solutions to these puzzles when they show up
later in the chapter!)
❼ We’ve placed large margins on the side of each page, for you to
A mathematical poem, to start off write notes in. When you’re in class or reading the book at home,
your margin notes: use these margins to write down questions that come to mind,
A dozen, a gross, and a score,
useful ideas from class, or as scratchwork when you’re trying to
plus three times the square root of work out an exercise.
four,
❼ Finally, we’ve placed even more exercises at the end of each chap-
divided by seven,
plus five times eleven, ter. Easier/more straightforward problems are labeled (-), while
is nine squared and not a bit more! trickier ones are labeled with a (+). Try working on these problems
to test your understanding!
Some exercises are labeled (++); these are exceptionally hard
and/or open problems in mathematics. We’re not expecting any-
one to solve these problems; instead, we’re listing them so that
you can see the kinds of tasks that you might study in a Ph.D
programme. (If you do solve any of these, though, please let us
know!)
While solutions to these end-of-chapter problems are not present
in this coursebook, your lecturers are happy to give you hints or
walk you through the solutions to any of these problems in office
hours or on Piazza (a discussion forum you can find on Canvas!)
Have fun, and enjoy the course!

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Expected Background Knowledge

2019 Chapter 0

0.1 What Does Compsci 120 Expect?

We are assuming that students entering this paper are relatively math-
ematically confident, having done well in NCEA Level 3 mathematics.
To be precise: we’re hoping that you’ve received a ‘merit’ or higher on
one of the three externally-assessed NCEA L3 mathematics standards,
In other systems, this is roughly equiv- namely differentiation, integration, or complex numbers.
alent to a passing mark in CIE A2
mathematics, or a C or better in CIE
Compsci 120 is a paper that assumes you are already comfortable with a
AS mathematics, or a 3/7 or higher in number of mathematical concepts and conventions. Trying to take this
IB mathematics. course without a solid background in mathematics is a bad idea. If you
do not have this background, you should take Maths 102! Maths
102 is a course ran every semester (including summer semesters) at the
University of Auckland, and has no prerequisites. It is designed to give
you the skills to succeed in CS 120!
To help clarify some of the specific concepts we’re expecting students to
understand (in case you’re coming from overseas, or it’s been a while
since you finished high school), here is a particular set of skills that we
are hoping you’ve acquired over your career. This list is not exhaustive,
but is just meant to point out the most common stumbling blocks that
less-prepared students encounter in Compsci 120. Again, if you do not
feel comfortable with these calculations, please enrol in Maths 102!
❼ Exponents. You should know what ab is for any two integers a, b,
By “comfortable with,” we mean that and know how to work with exponents. For example, the following
you should be able to see these calcu- calculations are ones you should be comfortable with:
lations done in lecture without further
explanation, and be able perform these 25 = 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 = 32, 100 = 1,
calculations yourself on an exam with-
out a calculator.
23 ⋅ 24 = 23+4 = 27 = 128, ab ⋅ ac = ab+c ,

(22 ) = 22⋅3 = 26 = 64, (ab ) = abc ,

3 c

1 1 1
7−2 = 2
= , a−b = .
7 49 ab
❼ Logarithms. You should know what log2 (n) and log10 (n) mean,
and be comfortable with calculations like the following:

log2 (32) = log2 (25 ) = 5, log10 (1) = 0, log2 (2n ) = n, 2log2 (a) = a.

❼ Expanding polynomials. You should be able to expand prod-

ucts of polynomials and other expressions. For example, the fol-
lowing calculations are ones that you should be able to read and
understand without further explanation:
(x + 1)3 = x3 + 3x2 + 3x + 1,

(x − y)(xn + xn−1 y + xn−2 y 2 + . . . + xy n−1 + y n ) = xn+1 + xn y + xn−1 y 2 + . . . + xy n

− xn y − xn−1 y 2 − . . . − xy n − y n−1
= xn+1 − y n+1 ,

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ Fractions. You should be able to combine ratios and fractions

with various arithmetic operations. For instance, you should be
comfortable with the following calculations:
1 4 1 ⋅ 8 + (−4) ⋅ 3 (−45) ⋅ (−5) 225
(− ) ⋅ (− ) =
4 45 5
− = =− , = ,
3 8 24 24 11 6 11 ⋅ 6 66
7 113 ⋅ 3 + 7 346 a d ac + d
3+ = = , + = ,
113 113 113 b bc bc
√ √
√ =√ ⋅√
1 1 + −x + x x 1 1 x+a x+a
= =1+ , = .
1−x 1−x 1−x x−a x−a x + a x − a2

❼ Solving equalities and inequalities. Given an equation or in-

equality in one or more variables, you should be able to rearrange
it to “solve” for one variable in terms of the others. For example,
the following processes should be ones you’re comfortable with:

16
3x − 4 = 12 ⇒ 3x = 16 ⇒ x= ,
3

(a2 − 1)b + 1 = a (a2 − 1)b = a − 1

a−1 1
⇒ ⇒ b= 2
= , if a ≠ ±1
a −1 a+1

−5x − 7 ≤ 3 ⇒ −5x ≤ 10 ⇒ x ≥ −2,

x2 + 1 1
∣x∣ < 3,
20
>2 ⇒ < ⇒ x2 + 1 < 10 ⇒
x2 +1 20 2

x2 − 3x + 2 > 0 ⇒ (x − 2)(x − 1) > 0 ⇒ (x − 2) and (x − 1) are both < 0, or (x − 2) and (x − 1) are both > 0
⇒ (x < 2 and x < 1), or (x > 2 and x > 1)
⇒ (x < 1) or (x > 2).

❼ Substitution. Given a function f (x), you should be able to plug

in values and expressions in to this function, and get the correct
output. For example, you should be capable of the following:

If f (x) = x2 + 1 then f (4) = 42 + 1 = 17, and f (x + 1) = (x + 1)2 + 1 = x2 + 2x + 2.

n(n + 1) (n + 1)(n + 2) (2n)(2n + 1)

If g(n) = then g(n + 1) = , and g(2n) = = n(2n + 1).
2 2 2

If h(x) = 1 − x then h(y) = 1 − y, and h(h(y)) = h(1 − y) = 1 − (1 − y) = y.

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

= a 2x1 domino
Numbers
2019 Chapter 1

Exercise 1.1. Take an 8 × 8 chessboard. Suppose you have a bunch of

2 × 1 dominoes lying around. Can you completely cover your chessboard
with dominoes, so that the dominoes don’t overlap or hang off of the
board? Now, suppose you have a younger sibling who has eaten the top-
left square of your chessboard. Can you completely cover without overlap
this chessboard with 2×1 dominoes?
What if they also ate the bottom-right square? Does this change your
answer?

Exercise 1.2. A mistake people often make when adding fractions is the
following:
1 1 ? 1
+ =
x y x+y
1 1 y x
Now, we know that this isn’t right, and that + is actually + =
x y xy xy
x+y
. Sometimes, however, even a formula that is wrong in general
xy
might be right in a specific situation! (Think of the old adage “a stopped
clock is right twice a day.”)
Does this ever happen with this mistake? In other words: are there any
1 1 1
values of x, y such that + = ? Or is this false for literally every
x y x+y
value of x and y?

1.1 Integers
We start our coursebook by studying the integers! In case you have
not seen the word “integer” before, we define it here:
In this course and in mathematics in
general, we will define lots of useful Definition 1.1. The integers are the collection of all whole numbers:
concepts. When we do so, we’ll label
these things by writing “Definition” that is, they consist of the whole positive numbers 1, 2, 3, 4, . . ., together
in bold, and then give you a carefully- with the whole negative numbers −1, −2, −3, −4, . . ., and the number 0.
written and precise definition of that We denote this set by writing the symbol Z.
word.
When studying for this class, it is a
good idea to make sure that you know The symbol Z comes from the German word “Zahl,” which means “num-
all of the definitions in this coursebook, ber,” in case you were curious.
as well as some examples and nonex-
amples for each definition where appro- Example 1.1. √ The numbers 1, −7, 10000, 78, 45, 0, −345678 are integers,
priate. but things like 2, −2.787878787, 81 and π are not.

In some form or another, integers have been used by humans for almost
as long as humans have existed theirselves. The Lebombo bone, one
of the oldest human artifacts, is a device on which people used tally
marks to count the number of days in the lunar cycle. The base-10
system and idea of numerals took a bit longer for us to discover, but we
can trace these concepts back to at least the Egyptians and Sumerians
around 4,000-3,000 BC. Finally, negative numbers and zero have been
with us for at least two millenia: historical documents tracking back
to at least 1000 BC describe how Chinese mathematicians used negative

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

numbers solve concrete problems on bounding areas of fields, exchanging

commodities, and debt.
In short, integers are quite handy! To this day, we use them to model
all sorts of problems. To give a pair of examples, let’s solve the first two
exercises from this chapter:
Answer to Exercise 1.1. By just trying things out by hand, it’s rel-
atively straightforward to find a tiling of a 8 × 8 board with no squares

?
missing; one solution is drawn in the margin. When you remove a square,
however, things get a bit more interesting! Try it for a while; no mat-
ter how you do this, you’ll always have at least one square left over
uncovered.
However, saying “I tried it a lot and it didn’t work” is not a very persua-
sive argument. Suppose that you had a boss that said that Chad from
Marketing said this was totally possible, and to have a working version
on their desk by 5pm or you’ll be fired. What do you do?
Well, maybe you start revising your resumé and looking for other work.
Before that, though, you may want to try to make a logical argument to
your boss that shows that it’s impossible to come up with such a tiling
— a set of reasons so airtight that no matter what they think or would
respond with, they’d have to agree that you’re right. In mathematics
and computer science, we call such logical arguments proofs!
In this class, we’re going to write these kinds of arguments for almost
all of the claims that we make. We’re going to leave the rules for what
constitutes a “proof” a little vague at first; in our first few chapters, we’re
going to just try to write a solid argument that tells us why something
is true, and anything that does that we’ll count as a proof. (Later in
this course, we’ll approach proofs a bit more formally: feel free to read
ahead if you can’t wait!)
For this problem, let’s prove that an 8 × 8 chessboard without a single
square cannot be covered with 2 × 1 dominoes as follows:

Proof. First, notice that each time we place a 2 × 1 domino on a board,

it covers two squares worth of area. Because dominoes cannot overlap,
this means that if we have placed k dominoes on our board, there are
2k squares of our board covered by dominoes.
If we have an 8 × 8 chessboard that’s missing a square, that board has
8 ⋅ 8 − 1 = 63 squares on it in total. Therefore, if we’ve completely covered
our chessboard with dominoes, the number k of dominoes we used must
satisfy the equation 2k = 63.
In other words, we have k = 31.5; i.e. we’ve used half of a domino! This
means that the other half of that domino is sticking off of our board
or overlaps another domino, which means we’ve broken the rules of our
tiling problem. So this is impossible!
To denote the end of our arguments,
Notice how in the argument above, the only things we used were fac- we draw a square box symbol, like 2.
This means that we’ve reached the end
tual observations (i.e. the number of squares in our 8 × 8 grid minus of our argument. In physics or other
a square, the amount of area taken up by each 2 × 1 domino) and log- fields, people often use QED in place
ical combinations of those observations. This is how you write a solid of this. Feel free to use either in your
own working!
argument!
To get a bit more practice with this sort of thing, let’s try the last part
of our exercise. In this problem, we had a chessboard with its top-left
and bottom-right squares removed, and wanted to try to cover this with
2 × 1 dominoes.
On one hand, the argument above no longer tells us that this is impos-
sible. A 8 × 8 board minus two squares contains 62 squares in total;
therefore, it might be possible to do this with 31 dominoes! However, if

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

you try this for a while, you’ll find yourself quite stuck again. So: why
are we stuck? What is a compelling argument we could write that would
persuade someone that this is truly impossible, and not just that we’re
bad at tiling?
After some thought and clever observations, you might come up with
the following:

Proof.
i. Notice that each time we place a 2 × 1 domino on our grid, it’s not
just true that it covers two units of area: it also covers exactly one
unit of white area and one unit of black area! This is because our
chessboard has alternating white and black squares.
ii. Because dominoes cannot overlap, this means that if we have placed
k dominoes on our board, there are k white squares and k black
squares covered by our dominoes. In particular, this means that
we always have as much black area covered as white area!
iii. However, if we count the black and white area in our 8 × 8 board
with the top-left and bottom-right squares removed, we can see
that there are 32 black squares and 30 white squares. These num-
bers are different! Therefore, it is impossible to cover this board
with dominoes as well.

Notice that when we wrote this argument, we used observations from

our earlier work, and modified them to suit this new problem! This is a
good problem-solving technique in general: always see if you can modify
a pre-existing solution before trying to come up with something entirely
new.

In our working above, the fact that we could not write 63 as an integer
multiple of 2 was a very useful observation! We can generalize this into
the concepts of even and odd integers:

1.2 Even and Odd Integers

Definition 1.2. We say that an integer is even if we can write it as 2
times another integer; in other words, we say that an integer n is even
if we can find an integer k such that n = 2k.
Similarly, we say that an integer is odd if we can write it as one plus an
even number; in other words; we say that an integer n is odd if we can
find an integer k such that n = 2k + 1.

Example 1.2. 2 is even, because we can write 2 = 2 ⋅ 1, and 1 is an

integer. Similarly, 4 is an even number, because we can write 4 = 2 ⋅ 2
and 2 is an integer.
As well, −2, −4 and −6 are all even, because we can write −2 = 2 ⋅ (−1),
−4 = 2 ⋅ (−2) and −6 = 2 ⋅ (−3), where −1, −2, −3 are all integers.
3 is odd, because we can write 3 = 2 + 1 and 2 is even (as shown above.)
Similarly, because 5 = 4 + 1 and 7 = 6 + 1, we have that 5 and 7 are also
odd. As well, −1, −3, −5 are all odd, as −1 = −2 + 1, −3 = −4 + 1, and
−5 = −6 + 1.

Exercise 1.3. Using the definition above, is 0 even or odd?

To get some practice writing logical arguments, let’s look at a few prop-
erties of even and odd numbers that you already know:

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Claim 1.1. The sum of any two odd numbers is even.

Before getting into a “good” solution for this claim, let’s first study an
argument that doesn’t work.
“Bad” solution: Well, 1 + 1 = 2 is even, 3 + 7 = 10 is even, −13 + 5 = −8 is
even, and 1001 + 2077 = 3078 is even. Certainly seems to be true!
2
A defense of the “bad” solution: This might seem like a silly argument,
but suppose we’d listed a thousand examples, or a billion examples, or
set a computer program to work overnight and had it check all of the
pairs of numbers below 1012 . In many other fields of study, that would
be enough to “show” a claim is true! (Think about science labs: there,
we prove claims via experimentation, and any theory you could test a
billion times and get the same result would certainly seem very true!)
Why this argument is not acceptable in mathematics and computer sci-
ence: When we make a claim about “any” number, or say that something
is true for “all” values, we want to really mean it. If we have not liter-
ally shown that the claim holds for every possible case, we don’t believe
that this is sufficient!
This is not just because computer scientists are fussy. In the world of
numbers, there are tons of “eventual” counterexamples out there:
❼ Consider the following claim: “The sequence of numbers “12, 121,
1211, 12111, 121111, . . . ” are all not prime.” Skip ahead a bit to definition 1.4 if
you haven’t encountered prime num-
If you were to just go through and check things by hand, you’d bers before!
probably be persuaded by the first few entries: 12 = 3 ⋅ 4, 121 =
11 ⋅ 11, 1211 = 173 ⋅ 7, 12111 = 367 ⋅ 11, 12111 = 431 ⋅ 281, . . .

³¹¹ ¹ ¹ ¹ ¹ ¹ ·¹¹ ¹ ¹ ¹ ¹ ¹ µ
136 1′ s

However, when you get to 12 111 . . . 1, that one’s prime! This is well
beyond the range of any reasonable human’s ability to calculate
things, and yet something that could come up in the context of
computer programming and information security (where we make
heavy use of 500+ digit primes all the time.)
❼ Here’s a fun exercise one of the writers of this coursebook saw in
the wild on Facebook, shared by one of our troll-ier friends:
Exercise 1.4. (++) Can you find positive integer values for ,
and so that the equation + + = 4 holds?
+ + +

On its surface, this looks simple, right? Just solve it. Roughly 10106 years, or 3 ⋅ 10113 sec-
onds, or about 6 ⋅ 10130 floating-point
And yet: suppose you set a computer program to work at this operations on the world’s fastest super-
problem, by just bashing out all possible triples of numbers. To be computer as of early 2019.
fair, let’s give you access to the world’s fastest supercomputer. By
the end of a week, you wouldn’t have found an answer. By the end
of a year, you wouldn’t have found an answer! Indeed: by the time
we reached the heat death of the universe, you’d still have found
nothing. You’d be tempted to say that no answer exists, right?
And yet, an answer exists! The math required to find this goes way
beyond the scope of this course, but if you’re curious: the smallest
known answer is to do the following:

=154476802108746166441951315019919837485664325669565431700026634898253202035277999,
= 36875131794129999827197811565225474825492979968971970996283137471637224634055579,
and
= 4373612677928697257861252602371390152816537558161613618621437993378423467772036.

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

In general, mathematics and computer science is full of things like this!

Huge numbers that are prime or that satisfy equations similar to the
ones above can be incredibly useful in information security (as you’ll see
in later papers in Compsci/Maths!)
Examples alone, then, are not enough for a good argument. What would
a good solution look like here? Here is one of many possible answers:

Claim 1.1. The sum of any two odd numbers is even.

Proof. Take any two odd numbers. Let’s give them names, for ease of
reference: let’s call them M and N . By definition, because M and N
are odd, we can write M = 2k + 1 and N = 2l + 1, for two integers k, l.
Therefore, M + N = (2k + 1) + (2l + 1) = 2k + 2l + 2 = 2(k + l + 1). In
particular, this means that M + N is an even number, as we’ve written
it as a multiple of 2!

The argument above might look a little awkwardly-written to you at a

first glance! This is because mathematics is something of a language
in its own right: there are certain phrases and constructions with very
specific meanings in mathematics, that we use when making arguments
to ensure that they’re completely rigorous. Let’s talk about some of
those phrases and concepts that came up in the argument above:
❼ We worked in general ! That is, we didn’t just look at a few exam-
ples, but instead considered arbitrary values! This is what writing
a phrase like “Take any two odd numbers” does for us: it doesn’t
lock us into specific odd numbers, but instead tells the reader that
we’re going to show that this works for anything that could ever
come up.
❼ We defined our variables! That is, we didn’t just try to use pro-
nouns to refer to our odd numbers the whole way through: instead,
we gave them variable names, by calling them N and M . This
makes writing our arguments much cleaner, as it’s much easier to
refer to something if it has a name!
Note also that we defined these variables only after saying what
kind of object they were! This is like when you’re writing code:
you typically have to declare the type of variable you’re working
with, i.e. you’d write something like “int numA” instead of just
saying “numA.”
❼ We used words to describe what we were doing and why it worked!
Mathematics, surprisingly, has a lot of words and sentences in it.
You should find that the number of sentences in most solutions
you make in this class exceeds the number of equations!
To get a bit more practice with this, let’s try out a few more claims:

Claim 1.2. The product of any two odd numbers is odd.

Proof. As before, let’s start by taking any two odd numbers. Let’s give
them names, and call them M and N respectively.
Also as before, let’s consult our definitions! By definition, because M, N
are both odd, we can again write M = 2k + 1 and N = 2l + 1, for two
integers k, l.
Therefore, M ⋅ N = (2k + 1) ⋅ (2l + 1). Expanding this product gives you
4kl + 2k + 2l + 1, which you can regroup as (4kl + 2k + 2l) + 1. Factoring
a 2 out of the left-hand part gives you 2(2kl + k + l) + 1.
Because k, l are integers, the expression 2kl + k + l inside the parentheses
is an integer as well, as any product or sum of integers is still an integer.

10

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Therefore, we’ve written M ⋅ N in the form 2 ⋅ (an integer) + 1; i.e. we’ve

shown that M ⋅ N satisfies the definition of being an odd integer!

Claim 1.3. No integer is both even and odd at the same time.

Proof. As before, let’s start by working in general. Take any integer N .

As a thought experiment, let’s think about what it would mean for N
to be both odd and even at the same time. If N was even, then by Phrases like “Take any integer n” or
definition we could write N = 2k for some integer k; as well, if N is also “take any number x” are useful; they
start by both saying that we’re going to
odd, then by definition we should be able to write N = 2l + 1 for some work in general with any number, and
integer l. also assign a variable to that number
Note that we had to pick different letters k, l when we applied the even so that we can refer to it.
and odd definitions! If we had used the same letter k for both, that
would imply that the same integer is being used in both definitions, and
this might not be true.
As a result, we have N = 2k and N = 2l+1. Combining gives us 2k = 2l+1,
which we can rearrange into 2(k − l) = 1; i.e. k − l = 12 .
But this is clearly not possible! k and l are both integers, and so their
difference must an integer as well; it cannot be 12 !
As a result, we’ve shown that it is impossible for an integer N to be
equal to 2k and 2l + 1 at the same time if k, l are both integers; that is,
it is impossible for N to be both odd and even!

By definition, you can think of our even/odd split above as classifying

every number as either “a multiple by 2” or “not a multiple of 2.” We
expand on this idea in the following section, where we study divisibility
and prime numbers.

1.3 Divisibility and Primes

Definition 1.3. Given two integers a, b, we say that a divides b if there
is some integer k such that ak = b.
There are many synonyms for “divides”: each of the phrases
❼ “a is a divisor of b”,
❼ “a is a factor of b”,
❼ “b is a multiple of a”,
❼ “b can be divided by a,” and
❼ a∣b
all mean the same thing as “a divides b.”

Here’s a string of examples, to make this clearer:

Example 1.3.
❼ 4 divides 12; this is because we can multiply 4 by 3 to get 12.
❼ 72 can be divided by −6; this is because we can multiply −6 by −12
to get 72.
❼ 2 does not divide 15; this is because for any integer k, 2k is an
even number, and so is in particular never equal to an odd integer
like 15.
❼ n is a multiple of 1 for any integer n; this is because we can always
multiply 1 by n to get n.
❼ n is a factor of 0 for any integer n; this is because we can always
multiply n by 0 to get 0.

11

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Example 1.4. Note that the phrases “divides” and “can be divided
by,” while quite similar-sounding in English, have almost the opposite
meaning in mathematics! For instance, 3 divides 9 and 9 can be
divided by 3 are the same claims.

To get some practice with making abstract arguments about divisibility,

let’s study a simple claim about factors and divisibility:

Claim 1.4. Let a, b, c be three integers. If a divides b and b divides c,

then a divides c.

Proof. Again, we need to work in abstract! That is: we cannot just

check this for a few values and say that “well, when a = 4, b = 12, c = 36
this all works out.” Instead, we must consider any three integers a, b, c,
In theoretical computer science / and work in general without knowing what a, b, c are.
mathematics, the words “any” and
“every” are often used interchangeably.
By definition, if a divides b, we can write ak = b for some integer k.
That is, if someone says to you “For Similarly, if b divides c, then we can write bl = c for some integer l.
any integer n, n2 ≥ 0,” this means the
Now, take the equation ak = b, and use this to substitute in ak for b
same thing as “For every integer n,
n2 ≥ 0.” in our second equation bl = c. This gives us akl = c, i.e. a(kl) = c.
Because k, l are both integers, their product is an integer; as a result,
we’ve written a ⋅ (an integer) = c. In other words, by definition we have
shown that a is a factor of c, as desired.

A particularly useful concept related to divisibility is the concept of a

prime number:

Definition 1.4. A prime number is any positive integer with only two
distinct positive factors; namely, 1 and itself.

Example 1.5. The first few prime numbers are 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, . . .

Observation 1.1. 1 is not a prime number.

Proof. This is because 1’s only factor is 1, and so 1 does not have two
distinct positive integer factors.

Observation 1.2. 2 is the only even prime number.

Proof. This is because every other positive even number by definition

has the form 2k, and so has at least 1, 2, k, 2k as its set of factors (and
thus has more than 2 distinct positive factors.)

Definition 1.5. A composite number is any positive integer n that

can be written as the product of two integers a, b, both of which are at
least 2 (and thus both of which are strictly smaller than n.)

Example 1.6. 6 = 2 ⋅ 3, 9 = 3 ⋅ 3, and 24 = 2 ⋅ 12 are all composite.

Observation 1.3. By definition, any positive integer is either a prime

number, a composite number, or 1.

Definition 1.6. Given a positive integer n, a prime factorization of

n is any way to write n as a product of prime numbers.

Example 1.7. Here are a few prime factorizations:

❼ 120 = 23 ⋅ 3 ⋅ 5,
❼ 243 = 35 ,
❼ 30031 = 59 ⋅ 509

Here are a pair of useful observations about prime numbers:

12

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Theorem 1.1. Every positive integer can be factorized into a product To see a proof of Theorem 1.1, take
of prime numbers in exactly one way, up to the ordering of those prime Compsci 225!
factors.
You might object to Theorem 1.1 by
In other words, this theorem is saying the following: saying that 1 cannot be factored into a
product of prime numbers. If so, good
❼ Every positive integer can be factored into primes, as illustrated thinking! For now, regard 1 as a spe-
above. So, for example, we can take 60 and write it as 2 ⋅ 2 ⋅ 3 ⋅ 5. cial case.
Later on, though, we’ll consider the
❼ No number can be factored into primes in two different ways, up idea of an “empty product” when we
to the ordering. That is: while you could write 60 as 5 ⋅ 2 ⋅ 3 ⋅ 2 or get to the factorial and exponential
5 ⋅ 3 ⋅ 2 ⋅ 2, you’re never going to write a prime factorization of 60 functions. At that time, we’ll argue
that has a 7 as one of its prime factors, or doesn’t have a 5. that the “empty product” or “product
of no numbers” should be considered
Proving this theorem is a bit beyond our skill set at the moment. Instead, to be 1, because 1 is the multiplica-
let’s mention a second useful fact about primes: tive identity. If you believe this, then
1 can indeed be written as the product
Theorem 1.2. There are infinitely many primes. of prime numbers; it specifically can
be written as the product of no prime
numbers!
This proof is also a bit beyond us for now. However, if you skip ahead
to the proof by contradiction section of our notes / wait a few weeks,
you’ll see a proof of this in our course!
Prime numbers (as you’ll see in Compsci 110) are incredibly useful for
communicating securely. Using processes like the RSA cryptosystem,
one of the first public-key cryptosystems to be developed, you can use
prime numbers to communicate secretly over public channels.
Prime numbers are also quite baffling objects, in that despite having
studied them since at least 300 BC there are still so many things we do
not know about them! Here are a few particularly outstanding problems,
the solutions for which would earn you an instant Ph.D/professorship
basically anywhere you like in the world:
Exercise 1.5. (++)(Goldbach conjecture.) Show that every even integer
greater than 2 can be written as the sum of two prime numbers. For
example, we can write 8 = 3 + 5, 14 = 11 + 3, 24 = 17 + 7, 6 = 3 + 3, . . .
Exercise 1.6. (++)(Twin prime conjecture.) A pair of prime numbers
are called twin primes if one is exactly two larger than the other. For
example, (5, 7), (11, 13) and (41, 43) are twin primes. Show that there
are infinitely many twin primes.

In general, working with prime numbers can get tricky very fast; even
simple problems can get out of hand! With that said, there are some
claims that are approachable. We study two such statements here:
Claim 1.5. Let ab be a two-digit positive integer (where b is that num-
ber’s ones’ digit and a is its tens’ digit.) Show that the number abab is
not prime.
ab abab prime factorization of abab
Proof. As noted before, examples alone aren’t enough for a solution. 10 1010 2 ⋅ 5 ⋅ 101
However, if you’re stuck (as many of us would be on first seeing a problem 98 9898 2 ⋅ 72 ⋅ 101
like this,) they can be useful for helping us find a place to start! 21 2121 3 ⋅ 7 ⋅ 101
88 8888 23 ⋅ 11 ⋅ 101
So: let’s create a bunch of two-digit numbers and factor them (say, via 92 9292 22 ⋅ 23 ⋅ 101
WolframAlpha.) If our claim is wrong, then maybe we’ll stumble across 43 4343 43 ⋅ 101
a number whose only factors are 1 and itself. Conversely, if our claim is
right, maybe we’ll see a pattern we can generalize! We do this at right.
As we do this, a pattern quickly emerges: it looks like all of these numbers
are multiples of 101! This isn’t a solution yet: we just checked six
numbers out of quite a few possibilities. It does, however, tell us how to
write a proper argument here:
Notice that for any two-digit number ab, ab⋅101 = ab⋅100+ab = ab00+ab =
abab. Therefore, any number of the form abab is a multiple of 101 and

13

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

also a multiple of ab, by definition. Because ab is a two-digit number by

definition, this means that abab has at least two factors other than 1.
This means that abab cannot be a prime number, as claimed!
√
Claim 1.6. Let n be any positive integer greater √ than 1. Let k = ⌊ n⌋
denote the number given by rounding down n. If n does not have any
of the numbers 2, 3, . . . k as factors, then n is prime.

Proof. We use the “suppose we’re wrong” technique from Claim 1.3.
That is: take any number n with the “no factors in the set 2, 3, . . . k”
property listed above. There are two possibilities:
❼ n is prime. This is what we want: if this case holds, we’re done!
❼ n is not prime. We want to show that this case cannot hold; if we
can do this, then we are left with only the case above, and have
thus proven our claim!
To do this: note that if n is not prime, then, it must have more
than 2 factors. Let a be one of those positive factors that is not 1
or n; by the definition of factor, then, we can write n = ab for some
positive integer b, and thus have that b is also another factor of n.
We know that because a ≠ n that b ≠ 1. We also know that be-
cause both a, b are factors of n, that by our “suppose we’re wrong”
√ these
assumption that a, b ≠ 2, 3, 4, . . . k. Therefore, by combining
results, both a, b > k; that is, both a, b are greater than n.
√ √
But this means that ab > n ⋅ n = n. But this is impossible, as
ab = n.
Therefore the only possibility left is that n is prime, as desired.

This last claim is particularly useful! We know that by definition, a

positive integer n is prime if it has exactly two positive integer factors:
1 and itself. Naively, then, to check if a number n is a prime, you might
think that you would have to check all of the numbers 2, 3, 4, . . . n − 1
each individually, and see if any of them divide n.
However, Claim 1.6 tells us that we don’t actually √ have to check all the
numbers up to n − 1; we only have to go up to n. This can save us a
lot of time: as you’ll see later in this coursebook in our runtime-analysis
√
section / in papers like Compsci 130 and 220, a “runtime” of checking
n cases is considerably better than a “runtime” of checking n cases!
For instance, to see if n = 101 is prime, Claim 1.6 tells us that because
101 is not a multiple of 2,3,4,5,6,7,8,9 or 10, then 101 is itself prime.
(Much faster than checking all of the numbers up to 100.)
We leave a few more exercises along these lines for the end of this chap-
ter. For now, we turn to an extension of the ideas behind divisibility:
modular arithmetic!

1.4 Modular Arithmetic

Modular arithmetic is something you’ve been working with since you
were a child! Specifically, think about how you tell and measure time:
❼ Suppose that it is currently 6am and 3 hours pass. Because 6+3 = 9,
we know that the answer is 9am.
Now, suppose that it is 11am and 18 hours pass. What is the time
now?

14

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Unlike the process above, you would not find the answer here by 12
11 1
adding 18 to 11 and getting “29 o’clock.” Instead, you’d find 11 + 10 2
18 = 29, and then subtract off 24 to get the right answer, 5 a.m. 9 3

In general, when you’re adding hours together to tell time, you 8 4

7 5
6
do so on a circle (as drawn at right,) where measurements “wrap
around” every 12 hours. That is: if it is 7 o’clock now and 38
hours pass, then the hour hand on a clock would point to a 9; this +6 hours
is because 7 + 36 = 45, and 45 − 12 − 12 − 12 = 9.
12
❼ In the above example, we saw that hours wrapped around in units 11 1
10 2
of 12. Similarly, we know that days of the week repeat in groups of
9 3
7. That is: let Monday = 1, Tuesday = 2, and so on/so forth until 8 4
we get to Sunday = 7. Then, suppose that it is Thursday and we 7
6
5
want to know what day of the week it is in 10 days. We can see
that because “Thursday +10”= 4 + 10 = 14 and 14 − 7 = 7, the day
must be Sunday. Using “99” on your microwave: one of
the less-useful “life pro tips.”
❼ Similarly, seconds wrap around in groups of 60. That is: if you
type in “99” on your microwave and walk away, the same amount
of time passes as if you had entered “1:39”, because 99 − 60 = 39
and thus 99 seconds is a minute and 39 seconds.
All of these calculations involved arithmetic where our numbers “wrapped
around” once they got past a certain point. This idea is the basis for
modular arithmetic, which is arguably the most useful mathematical
concept you’ll encounter in computer science.
To make this rigorous, let’s make a definition:

Definition 1.7. Take any two integers a, n, where n > 0. We define the
number a % n, pronounced “a mod n”, by the following algorithm : An algorithm is a step-by-step pro-
cess for solving a problem or perform-
❼ If a ≥ n, repeatedly subtract n from a until a < n. The result of ing a calculation! We will study al-
this process is a % n. gorithms in more depth later in this
class; you’ll also see them everywhere
❼ If a < 0, repeatedly add n to a until a > 0. The result of this process in Compsci 110 / 130 / 220 / 225 / es-
is a % n. sentially, every paper you’ll see in your
Compsci degree!
❼ If neither of these cases apply, then by definition 0 ≤ a < n. In this
case, a % n is simply a (that is, we don’t have to do anything!) Note: if n < 0, we can use the same pro-
We call % the modulus operator. Most programming languages im- cess as above to calculate a % n. The
only change is that we replace n with
∣n∣ in our steps, to compensate for the
plement it as described here, though (as always) you should read the
documentation for details. fact that n is negative.

To get a handle on this, we calculate a number of examples:

Example 1.8. ❼ 6 % 2 = 0. To see why, simply run the process

above. Because 4 > 2, we just subtract copies of 2 from 4 until we
subtract 2 subtract 2 subtract 2
get something less than 2: 6 ÐÐÐÐÐ→ 4 ÐÐÐÐÐ→ 2 ÐÐÐÐÐ→ 0 .
❼ −9 % 2 = 1. To see why, simply run the process above. Because
−9 < 0, we just add copies of 2 from 4 until we get something
add 2 add 2 add 2 add 2 add 2
nonnegative: −9 ÐÐÐ→ −7 ÐÐÐ→ −5 ÐÐÐ→ −3 ÐÐÐ→ −1 ÐÐÐ→ 1 .
❼ In general, if n is any even number, then n % 2 = 0. This is because
if we start with any even number n = 2k, the process above will
reduce 2k to 0 after subtracting/adding 2 k times in a row.
❼ Similarly, if n is any odd number, then n % 2 = 1. These two
observations come in handy when working with binary arithmetic:
as you’ll see in Compsci 110, these facts tell you that you can tell
if a number is even or odd by looking at its last digit in binary!
❼ 3 % 4 = 3. This is an easy one to calculate: because 0 ≤ 3 < 4, we
don’t have to do anything here!

15

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

We also study a quick argument-based problem, to make sure that we’re

comfortable with a more “abstract” way of working with the modulus
operator % :

Claim 1.7. If a, n are any two integers with n > 0, the quantity a % n
exists and is between 0 and n − 1. That is: the algorithm given above
to calculate % will never “crash” nor “run forever,” and it will always
generate an output between 0 and n − 1.

Proof. As always, because this is a claim about any two integers, we

need to work in general. That is: we cannot pick examples for a and n,
and instead need to consider every possible pair of values for a, n with
n > 0.
Our algorithm had three different cases: one for when a ≥ n, one for
when a < 0, and one for when 0 ≤ a < n. As such, our argument will
likely want three cases as well:
❼ a ≥ n. In this case, our algorithm repeatedly subtracted copies of
n from a until it got to something less than n.
This process clearly only runs for finitely many steps, as a starts
off as greater than n and decreases by a fixed nonzero amount at
each step. In particular, it must eventually be less than n, as that
was the condition we gave for this process to stop.
As well, if a ≥ n, then a − n ≥ 0; as a result, when this process ends,
it cannot end at a negative number.
Therefore, at the end of our process we’ve reduced a so that it’s
nonnegative and less than n, as desired.
❼ a < 0. In this case, our algorithm repeatedly added copies of n
from a until it got to something positive.
As before, this process clearly only runs for finitely many steps;
this is because a starts off as a negative number and increases by
a fixed amount at each step. In particular, it must eventually be
nonnegative, as that was the condition we gave for this process to
stop.
As well, if a < 0, then a + n < n; as a result, when this process ends,
it cannot end at a value equal to or greater than n.
Therefore, at the end of our process we’ve reduced a so that it’s
nonnegative and less than n, again as desired.
❼ 0 ≤ a < n. In this case our algorithm just outputs a, which again
has the desired properties.
Because every integer a falls into one of these three cases, we’ve proven
our claim for all values of a and all n > 0, as desired.

This sort of argument (that a given algorithm “works”) is one that we

will make repeatedly in this class! More generally, these are the most
common sorts of arguments you’ll want to make in computer science:
when working with any problems that require nontrivial algorithms or
thought, you’ll often want to be able to write proofs that the process
Some useful notation for rounding: should be bug-free in theory.
given any number x, we say that ⌊x⌋ is
“x rounded down,” ⌈x⌉ is “x rounded Modular arithmetic comes up often in computer science and mathemat-
up,” and [x] is “x rounded to the near- ics:
est integer, with .5 rounding up.”
For example, ⌊π⌋ = 3, ⌈π⌉ = 4, [π] = 3, ❼ Remainders. Take any two positive integers a, b. We say that
and [3.5] = 4. the quotient of a on division by b is just “ ab rounded down:” that
is, we say that the quotient of 14
3
is 4, as 14 3
= 4.66666 . . . rounds
down to 4.

16

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Based on this, we say that the remainder of ab is just a, minus

(the quotient of ab ) ⋅ (b). That is, the remainder of ab is everything
“left over” after we try to divide through by b.
a
Notice that by definition, a % b is the remainder of b
! So, for
example, we have the following:
14
– Because 14 % 12 = 2, we have that the remainder of 12
is 2.
26
– Because 26 % 13 = 0, we have that the remainder of 13
is 0.
❼ Binary arithmetic. Computers are built off of the binary number
system: i.e. instead of storing numbers in decimal notation where
we use the digits 0,1,2,. . . 9, computers store everything with just
0’s and 1’s (i.e. on and off, which computers are much better at
storing than something as imprecise as a decimal digit!) To work
with numbers in binary, we often work modulo 2, as you’ll see in
classes like Compsci 110.
❼ Overflow errors. In many programming languages, you have
to tell the computer the “type” of any variable that you want to
work with. That is, you can’t just tell the computer that “x = 3;”
instead, you have to first tell the computer that x is an integer by
writing something like “int x;” and then say “x = 3.”
However, when you declare that x is an integer, your computer
does not automatically assume that x can literally be any integer!
This is because when a computer declares a variable, it typically
has to set aside a block of space to store the information corre-
sponding to that variable. As a result, the computer has to make
an assumption about a reasonable range of values that your integer
will fall between (a common range is −231 to 231 − 1, i.e. you use
32 bits to describe your integer in binary, with one of those used
to record ±.)
Suppose you did this, though, and later on tried to increase the
value stored in x past its maximum value: i.e. you had x = 231 − 1,
and you tried to add one to x. The result (if you ignore warnings /
errors produced by your compiler) won’t be 231 , because this value
is too large to store! Instead, it will often “overflow” and wrap
around to become −231 instead. That is: computers technically
do a lot of their arithmetic “mod 232 ,” and this can trip you up if
you’re not being careful about the sizes of your variables! Check
out Wikipedia’s integer overflow page for a bunch of examples and
discussion around how to handle this situation.
❼ Checksums. On almost every ID number or tag (i.e. barcodes,
bank account numbers, library book numbers, etc) the last digit or
two is a “checksum” digit, found by adding up the previous num-
bers and applying the % n operator to the result (with n usually
being 10, but sometimes something stranger like 97 in the case of
bank account numbers.)
The use of these checksum digits is that they let a computer pro-
gram spot typos: if someone makes a mistake when entering a
number, they’ll usually do so in a way that changes the modulus
of the sum! As a result, we can spot this error by rechecking the
modulus, and alert the user that they’ve made a mistake.
❼ Cryptography. Calculating remainders modulo n has been key
to the entire field of cryptography, from its inception in ancient
Rome to its present-day uses. There is almost no way to keep
information secret in the modern world that does not use modular
arithmetic in some way.

17

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

As an example, let’s look at how modular arithmetic is used in one

of the first forms of cryptography: the Caesar cipher, otherwise
known as a shift cipher!

– To set this up, take the alphabet {a, b, . . . z}, and label each
letter with the numbers {0, 1, . . . 25}, as shown at right. Also
a b c d e f g h i j k l m
↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓
choose a secret key k from the set of numbers {0, 1, . . . 25};
0 1 2 3 4 5 6 7 8 9 10 11 12
n o p q r s t u v w x y z
↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ for this problem, let’s pick our secret key to be k = 15.
13 14 15 16 17 18 19 20 21 22 23 24 25
– Now, take the message you want to send: for example, let’s
send the phrase chain fusion .
– Swap the letters for numbers, to get 2.7.0.8.13 5.20.18.8.14.13 .
– Take each number in your code and add our secret code to it:
this gives us 17.22.15.23.28 20.35.33.23.29.28 .
– Now replace each number in our code with its value modulo
26; this gives us 17.22.15.23.2 20.9.7.23.3.2 .
– Translate this back to letters, to get the encrypted message
rwpxc ujhxdc . To an outsider, this would look like gibber-
ish! In ancient Roman times, most people who would inter-
cept such a message would assume that it was written in a
foreign language and not be able to translate it.
– But, if you knew the secret key, you could just translate this
message back to numbers and subtract 15 from each letter
modulo 26. This would give you the original message, as
desired.

This specific cipher is easy to crack: a simple brute-force approach

of trying all 26 keys one-by-one on the encrypted text would lead
you to the right answer pretty quickly. (Try it out by hand or by
writing a program; it’s not too bad.)

There are better ways to keep information secure, though! See

courses like Compsci 110 and Maths 328 for some more modern
approaches (that still use modular arithmetic at their core!)

1.5 % and Arithmetic

As of 2019, according to a variety of ar-
ticles + some back-of-the-envelope es- Consider the following problem:
timation by the coursebook’s authors.
By way of comparison, the total data
storage of humanity was about 108 ter- Question 1. What is
abytes in 2008. To put that figure in
modern perspective: YouTube broad-
casted more than this much data in the 1234567891011121314151617181912345678910111213141516171819 ?
past six months.

If you tried to use a computer to directly expand this problem, the size
29
of the resulting number (approximately 1010 ) would take up ≈ 1017
terabytes, i.e. several orders of magnitude in excess of the current total
storage capacity of humanity.
And yet, if you pop over to WolframAlpha and ask it this question, you’ll
get the following answer in about three seconds:

18

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

So: how is WolframAlpha able to calculate this so quickly, if the precise

result is so staggeringly huge?
Partly, this is because the “powers-of-ten” approximation can be made
by replacing the values here with just large powers of 10: i.e. things like
12345678910111213141516171819 ≈ 1028 can make approximating things
a bit easier. However, WolframAlpha also seems to know the last few
digits of this number, which seems a little bit spooky: how can you
calculate this without understanding the entire number?
The trick here is the modulus operation, which we can use to perform
certain arithmetic tasks very quickly. To show how this is possible, we
introduce the closely related idea of congruency modulo n:

Definition 1.8. Take any three integers a, b, n. We say that a is con-

gruent to b modulo n, and write a ≡ b mod n, if a − b is a multiple of
n. The phrase “a is equivalent to b mod-
ulo n” is also frequently used for this
concept.
To get an idea for how this works, we calculate a number of examples
here:

Example 1.9.

❼ 21 ≡ 5 mod 8; this is because 21 − 5 = 16 = 2 ⋅ 8 is a multiple of 8.

❼ 21 ≡ 13 mod 8; this is because 21 − 13 = 8 is a multiple of 8. Notice
that it’s possible for a number to be congruent to many other
possible values: i.e. 21 was congruent to both 5 and 13 modulo 8!
❼ −19 ≡ 7 mod 2; this is because −19−7 = −26 = (−13)⋅2 is a multiple
of 2.
❼ 14 ≡/ 18 mod 5; this is because 14 − 18 = −4 is not a multiple of 5.
❼ For any two integers a, b, a ≡ b mod 1; this is because a−b is always
a multiple of 1 (as any integer is a multiple of 1!)

19

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ For any two integers a, b, we have a ≡ b mod 2 if and only if a, b are

both even or a, b are both odd. This is because a ≡ b mod 2 holds
if and only if a − b is a multiple of 2, i.e. a − b is even, and this only
happens when a, b are the same parity: i.e. when they’re both
even or both odd.

Congruency and our modulus operator are very closely linked:

Claim 1.8. Take any three values a, b, n such that n ≠ 0. Then the
In English/mathematics, we say that following two statements are equivalent:
two statements are equivalent if they
hold in precisely the same situations: 1. a % n = b % n.
i.e. whenever one is true, the other 2. a − b is a multiple of n; i.e. a ≡ b mod n.
is true, and vice-versa. For example,
the two statements “n is an even prime
number” and “n = 2” are equivalent, as We reserve proving this claim for your practice problems at the end of
they both describe the same thing! this chapter (see exercise 7). Instead, let’s talk about how we can use
this:
Claim 1.9. Suppose that a, b, c, d, n are any set of integers with n ≠ 0,
such that a % n = b % n and c % n = d % n.
Then we have the following properties:
❼ (a + c) % n = (b + d) % n.
❼ (ac) % n = (bd) % n.

This claim is basically just saying that we can do “arithmetic modulo

n!” That is: for numbers a, b, c, d, you know that if a = b and c = d then
ac = bd and a + c = b + d, by just combining these equalities with the
addition and multiplication operations. This claim is saying that if your
values are “equal modulo n,” the same tricks work!
We prove this claim here:

Proof. We need to prove our claim in general here, as we’re making a

claim about any possible set of values, not a specific set of values! To do
so, we let a, b, c, d, n be any set of integers such that a % n = b % n and
c % n = d % n.
If we now use Claim 1.8, we get the following: a − b is a multiple of n,
and c − d is a multiple of n. By definition, this means that we can write
a − b = kn and c − d = ln for some pair of integers k, l.
By adding these equations together, we get (a + c) − (b + d) = kn + ln =
(k + l)n; that is, that (a + c) − (b + d) is a multiple of n. By Definition
1.8, this means that a + c ≡ b + d mod n! Applying Claim 1.8 gives us
(a + c) % n = (b + d) % n, as claimed.
If we take a − b = kn and multiply both sides by c, we get ac − bc = kcn;
similarly, if we take c − d = ln and multiply by b we get bc − bd = bln.
Adding these equations together gives us ac − bc + bc − bd = kcn + bln =
(kc + bl)n; i.e. ac − bd is a multiple of n. This means that ac ≡ bd mod n!
Again, applying Claim 1.8 then tells us that (ac) % n = (bd) % n, as
desired.

More English/mathematics word def- A quick corollary to Claim 1.9 is the following:
initions! A corollary is a claim that
follows quickly from the claim that just Corollary 1.1. If a % n = b % n, then for any positive integer k, we have
came before. That is, a corollary is a (ak ) % n = (bk ) % n.
“consequence” of the earlier claim, or
something that you get “for free” once
you’ve proven some earlier result. We leave proving this for Exercise 9; try proving this by using the mul-
tiplication property in Claim 1.9!
Instead of spending more time with this sort of abstract stuff, let’s do
something practical: let’s talk about how we could solve our Wolfra-
mAlpha problem! Specifically, let’s solve the following problem:

20

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Claim 1.10. The last digit of 213047129314 is 9.

Proof. The clever thing here is not that we can calculate this (again,
if we just wanted an answer, we could plug this into WolframAlpha),
but rather that we can calculate this easily! That is: we can use mod-
ular arithmetic to find this number without needing a calculator, with
relatively little work overall.
To do so, just make the following observations:
❼ First, 213047 % 10 = 7, and in general any positive number is con-
gruent to its last digit modulo 10. This is by definition: if you
take any number a greater than 10, subtracting 10 from it does
not change its last digit! Therefore, the process we defined to cal-
culate a % 10 will never change the last digit of a, and thus its
output at the end is precisely a’s last digit.
❼ Therefore, we have that (213047129314 ) % 10 = (7129314 ) % 10, by
using our “exponentiation” result from earlier.
❼ Now, notice that (72 ) % 10 = 49 % 10 = 9, and thus that (74 ) % 10 =
(72 ⋅ 72 ) % 10 = (9 ⋅ 9) % 10 = 81 % 10 = 1.

As a result, we have that for any k, (74k ) % 10 = (74 )k % 10 ≡

1k % 10 = 1.
❼ Therefore, because 129314 = 129300+12+2, and any multiple of 100
is a multiple of 4, we have that (7129314 ) % 10 = (72 ⋅ 7a multiple of 4 ≡ 72 ⋅ 1) % 10 =
9 ⋅ 1 ⋅ 1 = 9.
In other words, our number’s last digit is 9 !

It’s worth noting that this trick isn’t just useful for humans, but for
computers as well: you can use tricks like this to massively speed up
calculations in which you only care about the number’s last few digits,
and/or other pieces of partial information.
We can use this trick to basically calculate any number to any other
number’s last digit! For example, the task of finding the last digit of

1234567891011121314151617181912345678910111213141516171819

is now something we can do very quickly:

❼ Like before, we only care about the last digit of the thing we’re
exponentiating: i.e. we can reduce this problem to just finding
9123...19 ’s last digit.
❼ (92 ) % 10 = 81 % 10 = 1; therefore, for any even number 2k, we
have (92k ) % 10 = ((92 )k ) % 10 = 1k = 1.
❼ Therefore, for any odd number 2k + 1, we have (92k+1 ) % 10 =
(9 ⋅ 92k ) % 10 = 9 ⋅ 1 = 9.
❼ So, because the last digit of our base is 9 and the exponent is odd,
the last digit of our entire number is 9 !
Generalizations of this process (i.e. working modulo 100 to find the last
two digits, etc) can be used to quickly find the last few digits of any
number. This can be quite useful when writing computer algorithms:
most of the time when working with large numbers, you only need the
first few and last digits and an approximation for its size, not the entire
exact value!

21

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

1.6 Other Number Systems

In the past five sections, we’ve focused on studying the integers; i.e.
whole numbers. However, there are lots of problems in real life whose
answers cannot be given with just integers alone. Even going back to
primary school, you’ve seen and solved problems like the following:
1. Suppose that three hedgehogs come across two sausages while wan-
dering through a backyard barbie, and want to share them equally.
How many sausages should each hedgehog get?
2. If you have one and a half pavlovas left over from the barbeque,
Note: pavlovas are likely not part of a and you eat a third of one for breakfast, how much pav do you
balanced breakfast. have left?
3. How many dollars are there in ten cents?
The answers to each of these ( 32 , 23 − 31 = 76 and 100
10 1
= 10 , respectively) are
all relatively easy to calculate, and are tasks you’ve seen in school for
years now! However, their solutions are not things that we can express
as integers.
To work with them, we need some new sets of numbers! We start with
one such set that you’re already familiar with:
Definition 1.9. A rational number is any number that we can express
x
as a ratio , where x, y are integers and y is nonzero. We let Q denote
y
the collection of all rational numbers.
Definition 1.10. Given a rational number x, we say that x is the nu-
merator of xy , and that y is the denominator.

Example 1.10. The numbers 21 , −2 , 3 and 07 are all rational numbers.

13 6
3
The numerator of 6 is 3, and the denominator is 6; similarly, the numer-
ator of 70 is 0, and the denominator is 7.
Notice that every integer x can be written as a rational number, because
x = x1 .

There are several operations we know how to perform on rational num-

bers:
Fact 1.1. If ab , dc are a pair of rational numbers, then we can add and
multiply them! Specifically, we say that
a c ad + bc a c ac
+ = , and ⋅ = .
b d bd b d bd
Notice that these are both still rational numbers, as their numerator and
denominators are still integers, and the denominator is nonzero because
b, d ≠ 0.

With these operations in mind, we can answer our second exercise:

Answer to Exercise 1.2. In this problem, we’re asking if it is ever
possible for x1 + y1 to be equal to x+y
1
.
If you’re ever given a problem like this (i.e. someone hands you an
equation and asks you “are there any solutions,”) the first thing you
should try to do is solve the equation! That is: try to rearrange the
equation for one of your variables in terms of the other and/or some
constants, and hope that this tells you something.
In this situation, we know that x, y and x + y must be nonzero if the
fractions x1 , y1 , x+y
1
are defined. Therefore, the numbers x1 , y1 are rationals,
and so their sum is just x+y xy
as noted above.

22

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

If x+y
xy
1
= x+y , then because x, y and x + y are all nonzero, we can multiply
both sides by the denominators (i.e. multiply both sides by xy(x + y).)
Doing so will get rid of our fractions, as

(xy)(x + y) ( ) = (xy)(x + y) ( ) (x + y)2 = xy

x+y 1
⇒
xy x+y
⇒ x2 + 2xy + y 2 = xy
⇒ x2 + xy + y 2 = 0.

This is a simpler equation! In particular, if we think of x as our variable

and y as a constant, we have a quadratic equation. In general, a
quadratic
√
equation of the form ax2 + bx + c = 0 has the two solutions
−b± b2 −4ac
2a
when b2 −4ac is nonnegative, and has no real-valued solutions
if b2 − 4ac is negative.
In this situation, the quadratic formula above tells us that a √= 1, b =
−y± y 2 −4y 2
y, c = y 2 , and thus that our solutions (if they exist) would be 2
.
However, we know that y 2 − 4y 2 = −3y 2 is always a negative number; this
is because −3 is negative, while y 2 is a nonzero number squared (and
thus is positive.) Therefore we have no solutions!
There are other number systems that you’ve likely seen before in high
school as well:
Definition 1.11. The natural numbers, denoted N, is the collection
of all nonnegative integers. That is, N = {0, 1, 2, 3, 4, 5, . . .}.
Definition 1.12. The real numbers, denoted R, is the collection of
all numbers that you can write out with a (possibly infinite) decimal
expansion: i.e. it’s the collection of things like
❼ 2.1,
❼ −724,
❼ 0.111111 . . . = 0.1, and
❼ −3.1415926535 . . .

So, for example, the real number contain all of the rational numbers,
because you can do things like write
❼ 1
2
= 0.5,
❼ 1
3
= 0.333333 = 0.3 . . ., and
❼ 22
7
= 3.142857142857142857 . . . = 3.142857.
However, there are also real numbers that are not rationals: i.e. there
are quantities out there in the world that we cannot express as a ratio
of integers! We call such numbers irrational.
Observation 1.4. Notice that every real number, by definition, is either
rational or irrational.

It’s a bit beyond the skills we have at the moment, but numbers like
π = 3.1415926535897932384626433832795028841971693 . . .
and

√
e = 2.7182818284590452353602874713526624977572470 . . .
are both irrational numbers. Later on in this class, we’ll prove that 2
is also an irrational number (check out the proof by contradiction section
of this book if you cannot wait!)
With that said, though, it would be a bit of a shame to not show you at
least one irrational number. So, let’s close this chapter by doing this!

23

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Claim 1.11. The number log2 (3) is not rational.

Proof. We prove this claim by using the “suppose we’re wrong” structure
we’ve successfully used in Claims 1.6 and 1.3. That is: what would
happen if log2 (3) was rational?
Well: we could write log2 (3) = xy , for two integers x, y. Note that because
log2 (3) is positive (it’s greater than log2 (2) = 1), we can have x, y > 0.
Exponentiating both sides of this equation by 2 gives us 2log2 (3) = 2x/y .
By definition, log2 (⋆) and 2⋆ cancel each other out, so this simplifies to
3 = 2x/y .
Now, raising both sides to the y-th power gives us 3y = (2x/y ) . By using
y

our known rules for exponentiation, this simplifies to 3y = 2x .

On the left-hand-side, we have an odd number: it’s just y copies of 3
multiplied together! On the right-hand-side, however, we have an even
number: it’s x copies of 2 multiplied together.
These two values cannot be equal! In other words, our initial assumption
that we could write log2 (3) = xy must have been flawed, as this assump-
tion would imply that a number can be even and odd at the same time.
Therefore log2 (3) is irrational, as desired.

We close this chapter with some exercises for you to try out. Give these
an attempt, and post your answers / attempts on Piazza!

1.7 Practice Problems

1. (-) Let m be even. Show that −m is also even.
2. Show that if a and b are both odd, then ab is also odd
3. (+) Show that if a is an even number and k is a positive integer,
then ak is even.
4. In Claim 1.3, we showed that no number is both even and odd at
the same time. Prove the converse of this claim: that is, prove that
every integer is either even or odd (i.e. you can’t have an integer
that is both not even and not odd.)
5. (-) Write down all of the numbers between 1 and 100 that are
congruent to 2 modulo 12.
6. How many numbers between 1 and 1000 are congruent to 0 mod-
ulo 3? How many are congruent to 0 modulo 4? Can you make
a formula to count how many numbers between 1 and 1000 are
congruent to 0 modulo n, that works for any positive integer n?
7. (+) Prove Claim 6.2.
8. Take any three-digit number abc (where c is the ones digit, b is the
tens digit, and a is the hundreds digit.) Show that abcabc is never
a prime number.
9. Prove Corollary 1.1.
10. (+) Suppose that p1 , p2 , . . . pn are all prime numbers. Must the
number N = 1 + (p1 ⋅ p2 ⋅ . . . ⋅ pn ) be prime? Either explain why, or
find a set of prime numbers such that the value N defined above
is not prime.
11. (-) Show that n2 − n is always even, for any integer n.
12. Show that if n is an odd number, then n2 − 1 is a multiple of 8.
13. Find the last digit of 1234567890 without using a computer or cal-
culator.

24

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

14. (+) Find the last digit of 987654321 without using a computer or
calculator.
15. (+) What is the longest English word that can be shift-ciphered
into another English word?
16. (++) Take an integer n. If it’s odd, replace it with 3n + 1; if it’s
even, replace it with n/2. Repeat this until the number is equal to
1. Will this process always eventually stop?
a c
17. Consider the following claim: “Let , be a pair of rational num-
b d
a c
bers. Suppose that ad > bc. Then we have > .”
b d
Is this claim true? If it is, explain why. If it is not, find a pair of
fractions that demonstrate that this claim is false.
18. If x is rational and y is irrational, must x + y be irrational? Either
explain why or find a counterexample (i.e. find a pair of numbers (from https://xkcd.com/710/)
x, y such that x is rational, y is irrational, and x + y is rational.)
19. If x and y are both irrational numbers, must x + y be irrational?
Either explain why or find a counterexample (i.e. find a pair of
numbers x, y such that x, y are irrational, but x + y is rational.)
20. (+) Show that if a, b, c are all odd integers, then there is no rational
number x such that the equation ax2 + bx + c = 0 holds.
21. (++) Is π + e irrational?

25

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Sets and Strings

2019 Chapter 2

Exercise 2.1. You’re trying to break into a safe that has a PIN lock.
The safe has two buttons: 0 and 1. The PIN you’re trying to guess is a
three-digit sequence of binary numbers, and accepts the last three digits
you’ve typed in without needing you to hit enter: i.e. if you typed in
“00010,” the safe would open if the pin was either “000” or “001” or
“010”.
Sounds easy, right? There’s only eight possible PINs to check (two pos-
sibilities per digit, three digits in the PIN ⇒ 23 = 8 possible pins), so we
should be able to brute-force the lock by checking all possibilities.
However, the safe is wired to call the cops if more than ten buttons are
pressed and the correct PIN is not entered. As such, we can’t use our
brute-force approach: that could take 8 ⋅ 3 entries!
Is there an approach that is guaranteed to break us into the safe?

Exercise 2.2. You’re a geneticist! As such, you’re working with DNA

strands, which we can think of as long strings over the alphabet {A, C, G, T },
if we let these letters represent the nucleotides adenine, cytosine, guanine
and thymine.
You’ve designed a clever little combination of DNA restriction+polymerase
enzymes that do the following: given any string s of DNA strands, every
time there’s a substring of the form “. . . AC . . . ” in s, that substring
gets cut out and replaced with “. . . CCA . . . ”
So, for example, if your DNA strand was “ACGT,” it would get turned
into “CCAGT” and then would stay stable from there. If your strand
was “ACCT”, however, it would first turn into “CCACT”, and then
“CCCCAT.”
Suppose you’re originally working with strings of DNA all of the form
“AAAAC,” and you dump them into a bath with your enzymes in it.
What would you expect to see at the end of this process?

Earlier in this coursebook, we discussed various properties about num-

bers (divisibility, modular arithmetic, etc) that are very useful in com-
puter science!
However, numbers are not the only things that we work with in com-
puting systems. We also work heavily with things like passwords, user
IDs, databases full of names: i.e. strings! We study these objects in
the following section:

2.1 Strings
First, let’s define what an alphabet is:

Definition 2.1. An alphabet is any collection of symbols.

Example 2.1. You’re already familiar with the Roman alphabet:

A,B,C,D,E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T,U,V,W,X,Y,Z
Another alphabet that comes in handy is the collection of decimal digits!
We use this to describe numbers:

26

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

0,1,2,3,4,5,6,7,8,9
If we’re working in binary, we use a much smaller alphabet:
0,1
If we’re working with DNA, we’d use
A,C,G,T
as our alphabet, where these represent the four nucleotide bases cytosine
[C], guanine [G], adenine [A] or thymine [T].

There are other alphabets that are too big to write down here: for
example, the set of all Unicode symbols, or the set of all emojis!
Given an alphabet, it’s often useful to be able to refer to the whole
thing with a symbol. We’ll do this by writing something like Σ =
{0, 1, 2, 3, 4, 5, 6, 7, 8, 9}. This notation, where we list our symbols be-
tween a pair of curly braces and separate them with commas, tells us
that Σ is an alphabet containing the ten symbols 0,1,. . . 9.
With the definition of an alphabet in hand, we can define strings:
Definition 2.2. Take any alphabet Σ. A string over the alphabet Σ is
any sequence of letters in an alphabet. Some people refer to strings as
“words:” if you see an author referring
Example 2.2. If we let Σ be the Roman alphabet described earlier, then to a collection of words over a given
“cat,” “mongoose,” and “ssssssssssss” are all strings over this alphabet. alphabet, this is just a synonym for
strings!
Note that these strings don’t have to correspond to any particular mean-
ing; they’re just sequences of symbols!
If we let Σ be the decimal alphabet, then “123,” “00012,” and “999”
are each possible strings over this alphabet. Again, these don’t always
have to correspond to numbers! In particular, notice that as strings
we think that “00012” and “12” are different things. Even though as
numbers they’re different, as strings they’re quite different: “00012”
has zeroes in it, while “12”does not. (That is, think about entering a
password on your phone. There, if someone has a password of “00012,”
entering “12” shouldn’t unlock your phone!)
We will sometimes not specify an alphabet, and instead just refer to
strings by listing their entries. If so, we assume that their alphabet is
the most reasonable one to work with that string in (usually either the
Roman alphabet, decimal, or binary.)
A particularly useful string to refer to is the empty string “”, i.e. the
string containing no symbols. We denote this string by writing λ.

Strings are incredibly useful in computer science! Essentially every pro-

gram we have works with data in the form of strings, in the form of ID
numbers, names, IP addresses, and just simply the binary strings that
encode literally everything that a computer does.
Perhaps the simplest operation to define on strings is length:
Definition 2.3. The length of any string is the number of characters
in that string.
Note that when we’re working
with strings, writing something like
Example 2.3. The string “abcdef” has length 6, the string “00000” has “s1 s2 s3 . . . sn ” does not mean that
length 5, and the string “0123” has length 4. we’re multiplying these things all
together as if they were numbers!
The idea of length is useful when we’re trying to describe a general string! That is: the string “0123” is not the
same thing as the product 0 ⋅ 1 ⋅ 2 ⋅ 3 = 0.
Many arguments involving strings will start with the sentence “Take a This is why it’s important to keep
string s over the alphabet Σ. Let n be the length of s, and write s as track of the type of thing you’re work-
s1 s2 s3 . . . sn .” ing with / in general, at the start of
problems, to define your variables and
In particular, we can use this to define what it means for two strings to notation.
be equal:

27

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Definition 2.4. Take any two strings s = s1 s2 . . . sn and t = t1 t2 . . . tn of

the same length We say that s and t are equal if s1 = t1 , s2 = t2 , . . . and
s n = tn .

In other words, two strings are equal if and only if they are literally
character-for-character identical! Note that two strings of different lengths
are always nonequal.

Example 2.4.
❼ The strings “00001” and “1” are different. Even though the un-
derlying numbers they represent are the same, these are different-
length strings.
❼ If we take the alphabet given by all characters on a keyboard, the
strings “12+23” and “10+25” are different. Even though these are
the same length and represent the same underlying integer, the
characters are different in some places: for instance, the second
character of the first string is 2, while the second character of the
second string is 0.

A particularly useful operation on strings in computer science is con-

catenation:

Definition 2.5. Take any two strings s = s1 s2 . . . sn and t = t1 t2 . . . tm .

The concatenation of s and t, written st, is the string s1 s2 . . . sn t1 t2 . . . tm .

Example 2.5.
❼ Let s =“song” and t =“bird”. Then st is the string “songbird”.
❼ Let s =“12” and t =“0”. Then st is equal to “120”. Notice that
this is very different to what we would mean by writing st if we
thought of s, t as integers; there, st would denote 12 ⋅ 0 = 0!
In general, if you’re using string concatenation on strings of num-
bers, make sure to indicate this to your reader repeatedly through
your working so that they know what you’re doing. The use of
quotation marks can help keep things clear: that is, because we
wrote s =“12” and t =“0”, we’ve told you that we are thinking of
s, t as strings, and thus that concatenation is the appropriate way
to combine them.
❼ You can concatenate multiple strings at once: i.e. if s =“3”, t =“.”,
and u =“14159265...”, then stu is just “3.14159265...”

Notice that if s has length n and t has length m, then st has length
n + m.
Concatenation is used in tons of practical applications:
❼ Every bank account number is a concatenation of a bank code
(telling you what company you bank with,) an account number
(which tells the bank who owns this account,) and an account type
code (telling you what kind of account that number is attached to.)
❼ We saw that many ID numbers have “check digits” when we worked
with modular arithmetic! As such, your full ID number is usu-
ally created by concatenating your account number with the check
digit.
Related to the idea of concatenation, we have the concepts of prefixes,
substring and suffixes:

Definition 2.6. Let s and t be strings. We say that s is a prefix of t

if t is just s with some additional stuff possibly tacked on the end: i.e. if
we can find a third string u such that su = t.

28

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Similarly, we say that s is a suffix of t if t is just s with some additional

stuff possibly tacked on the front: i.e. if we can find a third string u such
that us = t.
Finally, we say that s is a substring (alternately, an “infix”) of t if t
is just s with some stuff possibly tacked on both the front and end: i.e.
if we can find strings u, v such that usv = t.

Example 2.6.
❼ If t = “snowball,” then “snow” is a prefix of t, “ball” is a suffix of
t, and “now” is a substring of t.
❼ If t = “112323411,” then “112” is a prefix of t, “323411” is a suffix
of t, and “2” is a substring of t.

As a bit of practice with writing arguments, we study a few claims:

Claim 2.1. The empty string λ is a prefix, suffix, and substring of every
string t.

Proof. Take any string t. Notice that t = λt = tλ = λtλ, because attaching

the empty string to the start or end of any string doesn’t change it.
Therefore λ meets the definition of being a prefix, suffix, and substring
for any other string t!

Claim 2.2. If s is a prefix of t, then s is a substring of t.

Proof. If s is a prefix of t, then there is some string u such that su = t.

Therefore, we have λsu = t as well, because concatenating the empty
string λ with any string doesn’t change it! This shows us that s satisfies
the definition of substring, as claimed.

We can use this idea of a “substring” to answer our safe-cracking prob-

lem:
Answer to Exercise 2.1. Think of the sequence of keys we’re entering 0 0 0 1 0 1 1 1 0 0
into the safe as a string s. If we do this, then the properties we want 0 0 0 1 0 1 1 1 0 0
s to have are the following: we want every three-digit binary string to 0 0 0 1 0 1 1 1 0 0
0 0 0 1 0 1 1 1 0 0
occur as a substring of s, and we want s to have length at most 10. 0 0 0 1 0 1 1 1 0 0
As it turns out, we can do this! Enter the following string: “0001011100.” 0 0 0 1 0 1 1 1 0 0
This string has length 10, and contains all possible three-digit pins as 0 0 0 1 0 1 1 1 0 0
0 0 0 1 0 1 1 1 0 0
subsequences, as shown in the margins. Success!
We can also use it to answer our DNA puzzle:
Answer to Exercise 2.2. On one hand, we could just brute-force the
answer, by repeatedly looking for “AC” substrings and replacing them
with “CCA” substrings:

AAA AC → AAA CCA ,

AA AC CA → AA CCA CA,
A AC C AC A → A CCA C CCA A,
AC C AC CCAA → CCA C CCA CCAA,
CC AC CC AC CAA → CC CCA CC CCA CAA,
CCCC AC CCC AC AA → CCCC CCA CCC CCA AA,
CCCCCC AC CCCCAAA → CCCCCC CCA CCCCAAA,
CCCCCCCC AC CCCAAA → CCCCCCCC CCA CCCAAA,
CCCCCCCCCC AC CCAAA → CCCCCCCCCC CCA CCAAA,
CCCCCCCCCCCC AC CAAA → CCCCCCCCCCCC CCA CAAA,
CCCCCCCCCCCCCC AC AAA → CCCCCCCCCCCCCC CCA AAA,

In other words: the result is a string of 16 C’s, followed by 4 A’s.

29

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Alternately, we could just notice the following: every time a C moves

past an A, we replace that C with two C. Therefore, if we move a C past
two A’s in a row, we’d expect to repeat this “doubling” process twice,
and have four C’s; in general, if we move a C past n A’s in a row, we’d
expect to see 2n C’s at the end, as we’ve doubled our C’s n times in this
process! This matches our results, as 24 = 16.

2.2 Sets
A second useful object, that we will often study in relation to strings, is
the concept of a set:
Definition 2.7. A set A is just a collection of things. We call those
things the elements of A, and write x ∈ A to denote with symbols the
statement “x is an element of A.”
To describe a set, we just list its elements between a pair of curly braces:
for example, {1, 2, 3} would be how we would describe the set consisting
of the three numbers 1, 2 and 3.

Basically every collection of things in real life can be thought of as a set:

Example 2.7.
❼ The collection of all strings in the Oxford English Dictionary is
a set. It contains elements like “heart” and “number,” but not
things like “arbleorble.”
❼ The collection of all words in Māori is a set. This set contains
elements like “tapawhā” and “tau” (the Māori words for rectangle
and number,) but does not contain strings like “123abc.”
❼ The collection of all commands in C is a set.
❼ The collection of all binary strings of length at most 2 is a set. We
could write this set out by listing its elements: {λ, 0, 1, 00, 01, 10, 11}.
❼ The “empty” set containing no elements {} is a set! We call this
the empty set, and refer to this by drawing the symbol ∅. This is
a fairly useful set to be able to refer to, for the same reasons that
0 is a useful number; it can be handy to talk about “nothing” in a
concrete way!
❼ The set of all prime numbers is a set: {2, 3, 5, 7, 11, 13, . . .}
❼ The set of integers Z, the set of rational numbers Q, the set of
natural numbers N, and the set of real numbers R are all sets.
❼ The set of all polynomials with degree at most 3 is a set: it contains
things like 2x − 4 and x3 − 3x2 + π.
❼ The set of all irrational numbers is a set.
❼ The set of all numbers that are solutions to the equation x3 − 3x2 +
3x − 1 = 0 is a set. (Specifically, because x3 − 3x2 + 3x − 1 = (x − 1)3 ,
this set is just {1}, the set containing only one object, namely 1.)

Notice that sets can be finite (in the case of things like “the collection
of all English words”) or infinite (in the case of the set of all prime
numbers!)
To make our lives easier when working with sets, let’s make a few nota-
tional conventions about how we should treat them:
❼ When we’re describing a set, we don’t care about the order in
which we list our elements: i.e. {cat, tag, tact} and {tag, cat, tact}
are both the “same” to us! This is because we only care about what

30

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

things are contained within a set; the order is something that we’ll
wind up changing a lot depending on the context (i.e. sometimes
alphabetical, sometimes by length...) and isn’t itself something we
want to care about.
❼ Similarly, when we’re describing a set, we only want to list each
element once. This is because otherwise it would be quite irri-
tating to try to look things up in our set: imagine a dictionary
that just listed the word “mongoose” forty times in a row!
As such, if someone gives you a set in which an element is repeated
twice, we just remove duplicates: i.e. we say { cat, tag, tact, tact,
tact } and {cat, tag, tact} are the same, and would never write
the first thing if we couldn’t help it.
❼ In the case of {cat, tag, tact}, we were able to describe our set by
just listing its elements. This works for small cases, but becomes
quite unwieldly for larger sets: imagine having to write out all of
the words in French before discussing the French language!
To deal with this, we have an alternate way of writing sets: you
can describe them by giving a property. For instance, when
we say “the set of all words in Māori” above, we’re giving you a
property that a given string of letters may or may not satisfy (i.e.
“is it a word in Māori”), and then taking the set of all words that
satisfy that property.
While the sentences we used in our examples above do work as
definitions for sets, you can also use the following more “math-y”
construction: to describe the set of all strings s with property blah,
you can just write

{s ∣ s has property blah}.

For instance, the set of all odd-length binary strings could be de-
scribed as the following: We use the notation “∈” as shorthand
for the word “in.”
{s ∣ length(s) = 2k + 1 for some k ∈ Z, and s is a binary string.}

The “s” on the left tells you the variable name, the divider ∣ just
separates the variable from its property, and the text at the right
gives the required property.
You can also use the left-hand part to describe the structure of
your set’s elements: i.e. something like

{concatenate(“001”, s) ∣ s is a binary string};

gives you all binary strings that start with the prefix “001.”
One useful concept when working with sets is a notion of “size:” In maths, the word cardinality is
used to refer to the size of a set. If you
Definition 2.8. A set A has size n if it contains precisely n different take papers like Maths 190 or Compsci
elements. If A contains infinitely many different elements, we say that 225, you can learn to study the idea of
A has “infinite” size. We denote the size of A by writing ∣A∣.
“different sizes of infinity” by working
with cardinality! In particular, using
the idea of a bijection in those courses,
Example 2.8. you can show that the integers, ratio-
❼ The set {1, 2, 3, π, 7} has size 5. nals, and natural numbers somehow all
have the same “countable” size of in-
❼ The set of all binary strings of length 2, i.e. {00, 01, 10, 11}, has finity, while the real numbers somehow
size 4. have a larger and “uncountable” size of
infinity. . .

Another useful concept when working with sets is the idea of a “subset:”
Definition 2.9. Take two sets A, B We say that B is a subset of A,
and write B ⊆ A, if every object in B is also an object in A.

31

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Example 2.9.
❼ Let A be the collection of all University of Auckland ID numbers,
and let B be the collection of all University of Auckland ID num-
bers corresponding to active Compsci 120 students. Then B is a
subset of A!
❼ Let A be the set of all binary strings of length 3, and let B be the
set of all binary strings with exactly two 1’s.
Then B is not a subset of A. This is because B contains things
like “11000”, which are not in A. Similarly, A is not a subset of
B, because A contains things like “000” that are not in B!
❼ Let A be the English language, and B be the collection of all
English words that rhyme with “avocado.” Then B is a subset of
A, as every word in B is by definition a word in A!

We also have a number of useful operations that we perform on sets:

Definition 2.10. Let A, B be a pair of sets. We define the union of

these two sets, A ∪ B, to be the collection of all elements that are in
either A or B or both.

Example 2.10.
❼ Let A be the collection of all English words with even length and
B be the collection of all English words with odd length. In this
case, A ∪ B is the collection of all English words.
❼ Let A be the collection of all Compsci 120 students that turned in
assignment 1, and B be the collection of all Compsci 120 students
that attended tutorial 1. Then A∪B is the collection of all Compsci
120 students who either attended tutorial 1 or turned in assignment
1, or both.
In general, unions work like “or” operations: the union of a set
defined by property A with a set defined by property B is just the
collection of all elements that satisfy property A or B.
❼ Let A be the collection of the 1000 most common phrases used in
spam emails ( things like “You be a Winner!!!1!!”) and B be a col-
lection of dodgy email addresses (e.g. “bi11.gates@micr0soft.ie”).
Then, the union A ∪ B is a good start for a “block list,” i.e. some-
thing that an email filter can use to automatically trash certain
emails.

Definition 2.11. Let A, B be a pair of sets. We define the intersection

of these two sets, A ∩ B, to be the collection of all elements that are in
both A and B at the same time.

Example 2.11.
❼ Let A be the English language and B be the German language.
Then A∩B is the set of words that are both in English and German
at the same time: i.e. words like “alphabet,” “computer” and “tag”
would be in A∩B, as they are all both English and German words.
❼ Let A be the set of numbers that are multiples of 3, and B be the
set of numbers who are multiples of 2. Then A ∩ B is the set of
numbers that are multiples of both 2 and 3; i.e. it’s the set of all
numbers that are multiples of 6!
Like how union was an “or,” intersection works like an “and” oper-
ation: that is, the intersection of a set defined by property A with
a set defined by property B is just the collection of all elements
that satisfy property A and B.

32

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ If A is the set consisting of ID numbers of current Compsci 120

students, and B is the set consisting of ID numbers of current
Compsci 720 students, then A ∩ B = ∅, the empty set. (This is
because there are no students simultaneously taking 120 and 720!)
Definition 2.12. Let A, B be a pair of sets. We define the difference
of these two sets, written A ∖ B or alternately A − B, to be the collection
of all elements that are both in A and not in B at the same time.
Example 2.12.
❼ If A was the set of ID numbers for all current Compsci 120 students,
and B was the set of ID numbers for Compsci 120 students who
attended at least eight tutorials, then A∖B is the set of ID numbers
for students who attended seven or fewer tutorials (i.e. the ID
numbers of students who will not have perfect marks for tutorials.
Don’t be in this set!)
❼ If A is the set of prime numbers, and B is set of odd integers,
then A ∖ B is the collection of all primes that are not odd: that is,
A ∩ B = {2}.
❼ Let A denote the set of all ASCII strings of length at least 10, B
be the set of all English words, and L3 be a list of the 10,000 most
common passwords. The set A ∖ (B ∪ L3 ) is a good start to a list
of “acceptable” passwords: i.e. if you were making a login system,
you could require all of your users to pick words in A ∖ (B ∪ L3 ).
Doing this would mean that they have to pick passwords that
– Have length at least 10 (i.e. are in A),
– Aren’t in a dictionary (i.e. not in B), and
– Aren’t commonly used (i.e. not in L3 ).
Useful!

Finally, we describe what it means for two sets to be equal:

Definition 2.13. We say that two sets A, B are equal if they both
consist of the same elements; that is, if
❼ Every element in A is a element in B, and
❼ Every element in B is also a element in A.

If you go back to our remarks earlier, this should make sense. We said
that the only thing we cared about for a set was the elements it con-
tained; i.e. we didn’t care about the order, and we ignored repeats/etc.
Therefore, two sets should be the same if they contain the same elements!
A useful proof technique, that we’ll often use to show that two sets are
the same is the following. Take two sets A, B that you want to show are
equal. Suppose you showed that
1. every element in A is a element in B, and also
2. every element in B is a element in A.
Then, by the definition above, we would know that A and B are equal!
As such, we can use this two-part approach to prove that many pairs
of objects are equal. We study a few examples here, to get the hang of
this:
Claim 2.3. Let A, B be any two sets such that A ⊆ B. Then A ∪ B = B.

Proof. We proceed as suggested above:

1. First, we show that every element in A ∪ B is in B. To do this,
we note that by definition A ∪ B is the set of all elements that are
in either A or B. Therefore, if we take any element s ∈ A ∪ B, we
either have s ∈ A or s ∈ B. This lets us work in cases:

33

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ If s ∈ A, then recall that we’ve assumed that A ⊆ B. By

definition, this means that every element in A is also in B.
Therefore, we have s ∈ B.
❼ If s ∈ B, then we trivially have s ∈ B.
Therefore, we’ve shown that for any s ∈ A ∪ B, we have s ∈ B, as
desired.
2. Second, we show that every element in B is in A ∪ B. This is not
too challenging.
Just notice that A∪B, by definition, contains all elements in either
A or B. Therefore, for any s ∈ B, we have s ∈ A ∪ B by definition.
This completes the two-way argument, as desired!

This is not the only way to prove that two sets are equal! As always,
simply expanding the definitions of both sets can often do the same trick:

Claim 2.4. Let A, B be any two sets. Then (A ∖ B) ∖ A = ∅.

Proof. We proceed by expanding our definitions:

1. First, notice that A ∖ B, by definition, is the collection of all ele-
ments in A that are not in B.
2. By definition again, (A ∖ B) ∖ A is “(the collection of all elements
in A that are not in B) that are also not in A.”
3. We can simplify this to “the collection of all elements in A that
are not in B or A.”
4. Every element of A is, um, in A.
5. Therefore, “the collection of all elements in A that are not in B or
A” is the empty set, as the “not in A” condition eliminates all of
the elements in A.
So we’ve proven our claim!

To close our chapter, we study a trickier example of this process:

Claim 2.5. If A, B, C are three sets, then A∖(B ∪C) = (A∖B)∩(A∖C).

Proof. We again proceed by expanding our definitions:

❼ On the left-hand-side, we have A ∖ (B ∪ C). By definition, A ∖ (B ∪
C) consists of all of the elements that are in A, but not in B ∪ C.
As well, by definition we know that B ∪ C is the set of all elements
that are in either B or C, or both.
Therefore, A ∖ (B ∪ C) can just be thought of as all of the elements
that are in A, but not in either B or C.
❼ On the right-hand-side, we can similarly use our definitions to no-
tice that A ∖ B is the set of all elements that are in A but not B,
and that A ∖ C is the set of all elements that are in A but not C.
As a consequence, we have that (A ∖ B) ∩ (A ∖ C) is the set of all
elements that are both (in A but not B) and (in A but not C).
Logically, we can simplify this sentence to the condition “in A, but
not in either B or C.”

These are the same statements; therefore we’ve shown that these sets
are equal!

34

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

2.3 Practice Problems

1. (-) Show that any suffix of a word t is a substring of t.
2. If s is both a prefix and a suffix of t, then must s = t? Either show
that this is true, or find a counterexample.
3. Show that if s is a substring of t and t is a substring of s, then
s = t.
4. (+) Suppose our safe in Exercise 2.1 had PIN numbers with decimal
digits, not binary (i.e. they could be any digit from 0-9, instead of
just 0 and 1.)
What is the smallest number of buttons we would have to press to
guarantee that the safe would open in this situation?
5. Suppose that A and B are two sets with the following property:
every string in A is a prefix of a string in B. Is it possible for A
to contain more elements than B? Either find such an example, or
explain why this is impossible.
6. (-) Explain why ∅ ∪ L = L for any set L.
7. Show that for any three set A, B, C, that A ∖ (B ∩ C) = (A ∖ B) ∪
(A ∖ C). (Try doing so using the method from Claim 2.3, and then
try with the method from Claim 2.4! Which do you prefer?)
8. Take two binary strings s, t of the same length. We say that s and t
are orthogonal if they disagree at precisely half of their locations:
for example, s =“1111” and t =“1100” are orthogonal.
(a) (-) Show that if s, t are odd-length strings, then s and t cannot
be orthogonal.
(b) Find a set consisting of four length-4 strings that are all or-
thogonal to each other (i.e. every possible pair of strings in
your set should be orthogonal.
(c) (+) Find a set consisting of 2n length-2n strings that are all
orthogonal to each other.
(d) (++) What is the largest set of orthogonal length-668 strings
that you can make?

35

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Combinatorics and Probability

2019 Chapter 3

3.1 How to Count

This might seem like a silly section title; counting, after all, is something
that you learned how to do at a very young age! So let’s clarify what we
mean by “counting.”
On one hand, it is easy to see that there are four elements in a set like

(3,3) (3,3) A = {3, 5, 7, Snape}.

We’re not going to practice counting things like this! Instead, consider
the following four exercises:
Exercise 3.1. How many strings of five letters are palindromes (i.e.
(0,0) (0,0) can be read the same way forwards and backwards?)
Exercise 3.2. A lattice path in the plane R2 is a path joining integer
points via steps of length 1 either upward or rightward. How many lattice
paths are there from (0, 0) to (3, 3)?
Exercise 3.3. How many seven-digit phone numbers exist in which the
digits are all nondecreasing?
Exercise 3.4. In how many ways can you roll a six-sided die three times
and get different values each time?

All of these are “counting” problems, in that they’re asking you to figure
out how many objects of a specific kind exist. However, because the sets
in question are trickily defined, these problems are much harder than
our “how many elements are in {3, 5, 7, Snape} question.
To approach them, we’ll need some new counting techniques! This is
the goal of this chapter: we’re going to study combinatorics, the art
of counting, and develop techniques for solving problems like the ones
above.
Let’s start off with a simple task:
Problem 3.1. Suppose that we have 3 different postcards and 2 friends.
In how many ways can we mail out all of our postcards to our friends
while we’re on vacation?

Answer. Let’s give our cards names A, B, C, and also give our friends
names X and Y , for easy reference.
In the setup above, a valid “way” to mail postcards to friends is some
way to assign each postcard to a friend (because we’re mailing out all
three of our postcards.) To do this, think of going through each card
A, B, C one-by-one and choosing a friend X, Y for each card.
By using brute-force, we can just enumerate all of the possibilities:
❼ X gets A, B, C, Y gets nothing.
❼ X gets A, B, Y gets C.
❼ X gets A, C, Y gets B.

36

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ X gets B, C, Y gets A.
❼ X gets A, Y gets B, C.
❼ X gets B, Y gets A, C.
❼ X gets C, Y gets A, B.
❼ X gets nothing, Y gets A, B, C.
This works, and tells us that there are 8 different ways to do this.
However, if we had more cards this brute-force process seems like a bad
idea:
Problem 3.2. Suppose that we have 10 different postcards and 4 friends.
In how many ways can we mail out all of our postcards to our friends
while we’re on vacation?

Answer. We could try a brute-force approach like before. It would

work, if you had enough time! However, if you start doing this you’ll
quickly find yourself bored out of your mind; there are tons of ways to
do this, and just trying to list all of them is exhausting.
Instead, to do this, let’s think about a process to generate ways to send
out cards. That is: think about going through each card one-by-one and
choosing a friend for each card:

4 possible friends 4 possible friends . . . 4 possible friends

´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
card 1 card 2 card 10

When doing this process, we have 4 choices of friend for each card, and
any combination of those choices will give us a valid way to send out
cards. Therefore, we should have

4 ⋅ 4 ⋅ . . . ⋅ 4 = 410
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¸′¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
10 4 s

total ways in which we can send out our cards.

This looks like an excellent answer to a counting problem, and also a
much better method than our first approach! You’ll see this called the multiplica-
tion principle in other combinatorics
Alongside our answer, we also came up with a fairly interesting method resources online!
for counting at the same time. Specifically, we had a set A of the follow-
ing form:
1. Each element f of A could be constructed by making k choices in
a row.
2. There was a fixed number ni of possibilities for the i-th choice By “fixed,” we mean the following:
made in constructing any such element of A. the number of such choices is not af-
fected by our other choices. That is,
3. Therefore, we had n1 ⋅ n2 ⋅ . . . ⋅ nk total elements in A.
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶
we’ll always have ni options for our i-
th choice, no matter what our earlier
k n′ s
choices actually were.
To get some practice with this principle, we study a number of examples
of it in action:
Problem 3.3. How many binary strings of length 5 exist?

Answer. A binary string of length 5 has five characters in it, each of

which is 0 or 1. Therefore, making such a string is the same thing as
making five choices in a row, each of which has two options:
0 or 1 0 or 1 0 or 1 0 or 1 0 or 1
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹¶
entry 1 entry 2 entry 3 entry 4 entry 5

By the multiplication principle, this can be done in 25 ways.

37

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Problem 3.4. Take a six-sided die, and roll it twice in a row. In how
many ways can you do this and not see the same value repeat?

Answer. We can again think of this process as making two choices in

a row:

1, 2, 3, 4, 5 or 6 1, 2, 3, 4, 5 or 6, ≠ roll 1
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
roll 1 roll 2

There are 6 possibilities for the first roll. Once this is done, there are 5
possibilities for the second roll, as we can have any value show up other
than the one that occurs in the first roll. Therefore, there are 6 ⋅ 5 = 30
possibilities in total by the multiplication principle.

It is worth noting that the multiplication principle does not apply to all
possible situations! Consider the following counting problem:

Problem 3.5. How many binary strings of length at most 3 exist?

Answer. Using the same logic as in Problem 3.3, we know that

❼ There are 23 = 8 binary strings of length exactly 3,
❼ there are 22 = 4 binary strings of length exactly 2, and
❼ there are 21 = 2 binary strings of length exactly 1.
Also, we know that there is one binary string of length 0, namely the
empty string λ.
Therefore, in total, there are 8 + 4 + 2 + 1 = 15 strings in total.
You’ll sometimes see this “add disjoint
things” process referred to as the ad- Notice how at the end of the process above, we didn’t multiply these
dition principle.
numbers together! This is because the numbers 8, 4, 2, 1 aren’t all part
of creating one giant mega-string. Instead, they each came from disjoint
processes: that is, the strings of length 3 have no overlap with the strings
of length 2, and so on/so forth. As such, we added them together!
To get some practice with spotting places where we use addition in count-
ing, let’s look at another problem:

Problem 3.6. Suppose that you have four friends Aang, Korra, Zuko,
and Toph. In how many ways can you arrange them in a row, if Aang
and Zuko are currently fighting and can’t be placed next to each other?

Answer. If we just had the problem “In how many ways can we ar-
range Aang, Korra, Zuko, and Toph in a row,” our problem is pretty
straightforward! By using our multiple-choice framework from before,

A,K,Z,T A,K,Z,T A,K,Z,T A,K,Z,T

´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ ´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
friend 1 friend 2 friend 3 friend 4

we can see that we have 4 choices for the first friend in our order, 3 for
our second (as we can’t repeat a friend), 2 for our third (can’t choose
any of the two previously-placed friends), and 1 for our last (everyone
else is already in place!) So, in total, we have 4 ⋅ 3 ⋅ 2 ⋅ 1 = 24 ways to do
this.
However, in this process we’ve allowed all of the possible arrangements,
including ones like “A,Z,T,K” where Aang and Zuko are together. How
do we correct for this?
Well, let’s try to count all of the ways in which Aang and Zuko could
be placed together! There are two ways in which this can happen:

38

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ Aang and Zuko are placed in the order “AZ.” To find all of the
arrangements where this could happen, think of all of the ways to
place three things T , K , AZ in order! There are 3 ⋅ 2 ⋅ 1 = 6
ways to do this, by the same reasoning as above.
❼ Aang and Zuko are placed in the order “ZA.” To find all of the
arrangements where this could happen, think of all of the ways to
place three things T , K , ZA in order! There are again 3⋅2⋅1 = 6
ways to do this.
Because the “AZ” and “ZA” cases are completely distinct, we add them
to get that there are 6 + 6 = 12 ways for Aang and Zuko to be placed
together in either order.
Finally, in any way of arranging our friends at all, we either have Aang
and Zuko together or not. Therefore, we add these situations together,
to get
all arrangements = arrangements with Aang and Zuko together + arrangements with Aang and Zuko separate
24 = 12 + arrangements with Aang and Zuko separate .

In other words, there are 24 − 12 = 12 arrangements where Aang and

Zuko are separate!
One particularly useful thing we can do with counting arguments is mea-
sure probability. That is: suppose that we have an experiment in
which all of the individual outcomes are
❼ equally likely, and
❼ mutually exclusive.
To give some examples of experiments like this:
❼ Suppose that you were rolling a fair 6-sided die. In this case,
there are 6 outcomes: the die could show a 1,2,3,4,5, or 6. These
outcomes are all equally likely if our die is fair: they’re also all
mutually exclusive, as a die can only show one number at a time.
❼ Suppose that you flipped a fair coin. In this case, there are 2 out-
comes: heads or tails. They’re equally likely if the coin is fair, and
they’re both mutually exclusive, as a coin cannot simultaneously
be both heads and tails.
❼ Suppose that you have a standard 52-card deck, and reveal the
top card. In this case, there are 52 possible outcomes, i.e. the 52
possible cards in the deck, and each of them are mutually exclusive
(because you’re only revealing one card.)
❼ Suppose that you have a bag containing 12 cookies, and you take
one cookie out. In this case, there are 12 events (the 12 possible
cookies you could have picked) and they’re all mutually exclusive
(because you are only taking one cookie.)
In situations like this, we have the following very useful observation:
Observation 3.5. Take any experiment in which all of the individual
outcomes are equally likely and mutually exclusive. Let n be the num-
ber of all possible outcomes. Then, the probability that any individual
outcome happens when you run that experiment is n1 .
More generally, take any set A of outcomes and let a denote the number
of elements in A. Then the probability of running your experiment and
having an outcome from A happening is na .

This is very useful! By using our newfound ability to count complicated

objects, we can now determine the likelihood of various events happen-
ing.

39

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This probably feels somewhat abstract, so let’s illustrate this with a few
examples:
Problem 3.7. Take two fair 6-sided dice and roll them. What is the
probability that the sum of these dice is 7?

Answer. On one hand, we know that there are six possible ways for
our dice to sum to 7: the six pairs (1, 6), (2, 5), (3, 4), (4, 3), (5, 2), (6, 1)
each correspond for a way for to roll two dice and get a 7. (Notice that
we count (1, 6) and (6, 1) differently! This is because our two dice are
different, and so rolling a 1 on the first die and a a 6 on the second is a
different event than rolling a 6 followed by a 1.)
On the other hand, we know that there are 36 possible outcomes for a
given roll of our two dice; this is because there are 6 choices for what
the first die can be, 6 for the second, and we multiply these together.
6
Thus, the odds that we roll two dice and they sum to 7 is 36 = 61 .

Problem 3.8. You and three of your friends are graduating. At gradu-
ation, you each throw your hats in the air to celebrate! At the end, you
each pick up a hat from the ground. What are the odds that each of you
pick up your own hat? (Assume that you’re picking up hats at random,
i.e. you’re as likely to pick up your own hat as anyone else’s hat: also
assume that for some reason no-one else is graduating, and so there are
only four possible hats to pick up.)

Answer. It takes a bit more thought to come up with the experiment

structure here. To do so, let’s label you and your friends 1,2,3 and 4, and
similarly label the hats 1,2,3 and 4. If we do so, then we can describe any
way for you and your friends to pick up hats by just listing the numbers
1,2,3,4 in some order: that is, (2, 1, 3, 4) would describe the situation
where 1 has 2’s hat, 2 has 1’s hat, and 3 and 4 have their own hats.
In this setting, there are as many ways to pick up hats as there are ways
to write the numbers 1,2,3,4 in some order. We know from our work
earlier that there are 24 ways for this to happen in (it’s the same thing
as ordering our four friends in Problem 3.6!)
There is also only one way for everyone to get their own hat: namely,
the ordering (1,2,3,4). Therefore the probability that everyone picks up
1
their own hat is 24 .

3.2 Ordered Choice

A special case of the counting processes we studied earlier is the follow-
ing:
Observation 3.6. (Ordered choice with repetition.) Suppose that
you are choosing k objects from a set of n things, where you care about
the order in which you choose your objects and can repeatedly pick the
same thing if desired. There are nk many ways to make such a choice.

This principle is remarkably handy! We can use it to answer one of our

problems from the start of this chapter:
Answer to Exercise 3.1. In this problem, we’re trying to count the
number of five-letter strings that are palindromes (i.e. can be read the
same way forwards and backwards.)
To do this, start by noticing that any five-letter palindrome can be con-
structed by taking an arbitrary three-letter string and sticking its second

40

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

and first characters at the end of the string: e.g. you can transform “rad”
to “radar,” “ten” to “tenet” and “eev” to “eevee.” This process is clearly
reversible: i.e. you can take any five-letter palindrome and cut off its
last two letters to get a 3-letter string.
Any such three-letter string is formed by making three choices in a row,
and we have 26 choices for each letter; this gives us 263 many such
strings, and thus 263 many five-letter palindromes.
This can get a bit more complex, however. Let’s try changing our post-
card problem a bit from before:
Problem 3.9. Suppose that we have k different kinds of postcards, n
friends, and that we want to mail these postcards to our friends. Last
time, however, it was possible that we just mailed all of our postcards
to the same friend. That’s a bit silly, so let’s add in a new restriction:
let’s never send any friend more than one postcard.
In how many ways can we mail out postcards now?

Answer. We can still describe each way of sending postcards as a se-

quence of choices:

? choices ⋅ ? choices ⋅ . . . ⋅ ? choices

´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
k total slots
A common question students ask
As before, we still have n possibilities for who we can send our first about this calculation: why do we go
card to. However, the “ordered choice with repetition” principle doesn’t to n − 1 and not n in the product
above? Well: we have k total slots.
immediately apply here: because we don’t want to repeat any of our In the first slot, none of our choices
friends, we only have n − 1 choices for our second slot, instead of n as were eliminated yet! In the second slot,
before! In general, we have the following sequence of choices: however, we’ve eliminated one choice
with our first slot. By the third slot
we’ve eliminated two possibilities, by
n choices ⋅ n-1 choices ⋅ n-2 choices ⋅ . . . ⋅ n- (k-1) choices , the fourth we’ve eliminated three pos-
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶ sibilities, and in general in the i-th slot
k total slots we’ve eliminated i−1 possibilities. This
leaves us with k − 1 possibilities elim-
which translates into inated by the time we get to the k-th

n ⋅ (n − 1) ⋅ (n − 2) ⋅ . . . ⋅ (n − (k − 1))
slot!

many choices in total.

We can simplify this expression considerably by introducing a useful
function:
Definition 3.1. For any nonnegative integer n, we define the factorial
of n, written n!, as the following product:

n! = 1 ⋅ 2 ⋅ 3 ⋅ 4 ⋅ . . . ⋅ n

If n = 0, we think of this as the “empty product,” and say that 0! = 1.

With this notation in mind, we can simplify our answer to the postcard
problem considerably: notice that

(n ⋅ (n − 1) ⋅ . . . ⋅ (n − (k − 1))) ⋅ ((n − k) ⋅ (n − (k + 1)) ⋅ . . . ⋅ 3 ⋅ 2 ⋅ 1)

n ⋅ (n − 1) ⋅ . . . ⋅ (n − (k − 1)) =
((n − k) ⋅ (n − (k + 1)) ⋅ . . . ⋅ 3 ⋅ 2 ⋅ 1)

n!
(n − k)!
= .

We can use this idea to describe another counting method:

41

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Observation 3.7. (Ordered choice without repetition.) Suppose

that you are choosing k objects from a set of n things, where you care
about the order in which you choose your objects, but can only pick an
n!
(n − k)!
object at most once. There are many ways to make such a

choice if k ≤ n, and 0 ways otherwise (as we’ll run out of choices!)

To practice using this observation, let’s work a pair of problems:

Problem 3.10. If you take a six-sided die and roll it three times, what
is the probability that you get different values each time?

Answer. If we keep track of what our first/second/third roll of our die

is, then there are 63 many possible ways for us to roll a 6-sided die
three times: we have 6 possible outcomes, we’re choosing one of those
outcomes 3 times, and the order matters + repeats are allowed.
Conversely, if we want to know the number of ways to roll a 6-sided die
6!
and have no repeated values, this is = 6 ⋅ 5 ⋅ 4, by our order matters
3!
+ no repeats process described above. (Note that this answers Exercise
3.4!)
6⋅5⋅4
Therefore, in total, we have a 63
≈ 0.56 = 56 % chance of this hap-
pening.

Problem 3.11. Suppose that you have six lightbulbs: two identical red
bulbs, and one green / one blue / one white / one pink.
In how many ways can you screw these bulbs into a string of six lightbulb
sockets? (Assume that the order in which we string these bulbs matter:
i.e. we consider “RRGBW P ” and “P W BGRR” to be different.)

Answer. First, we notice that if we could distinguish between our two

red bulbs, this problem is a straightforward ordered choice without rep-
etition problem: we would have 6 different bulbs, 6 sockets, and thus
6 ⋅ 5 ⋅ 4 ⋅ 3 ⋅ 2 ⋅ 1 = 6! ways to do this.
However, these red bulbs are identical! This creates a bit of a problem:
if we label our red bulbs R1 , R2 , then the method above thinks that
“R1 W GR2 BP ” and “R2 W GR1 BP ” are different strings, even though
to us they are identical.
How can we correct for this? Well, just notice the following: given any
string where our red bulbs are not labeled, there are exactly two ways to
label them: either label the leftmost red bulb “R1 ” and the other “R2 ”,
or vice-versa. In other words, there are twice as many strings where we
have labelings as strings where we don’t!
6!
Therefore, we can just divide our earlier answer by 2 to get as our
2
final answer.

3.3 Unordered Choice

In the section above, we discussed how to choose things in situations
where we both could and could not “repeat” our choice. In both scenar-
ios, however, the order in which we made these choices mattered!
This does not always happen in real life. Consider the following example:

42

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Problem 3.12. Suppose that we have just one friend that we want to
send postcards to. We still have n different kinds of postcards, but now
want to send that one friend k different postcards in a bundle (say as a
gift!) In how many ways can we pick out a set of k cards to send our
friend? In mathematics: whenever you have a
problem at hand, constantly look for
In this problem, we have n different kinds of postcards, and we want to modifications like these to make to the
problem! If you’re stuck, it can give
find out how many ways to send k different cards to a given friend. At you different avenues to approach or
first glance, you might think that this is the same as the answer to our think about the problem; conversely, if
second puzzle: i.e. we have k slots, and we clearly have n choices for the you think you understand the problem,
this can be a way to test and deepen
first slot, n − 1 choices for the second slot, and so on/so forth until we
have n − (k − 1) choices for our last slot.
that understanding.

n!
(n − k)!
This would certainly seem to indicate that there are many ways
to assign cards. However, our situation from before is not quite the same
as the one we have now! In particular: notice that the order in which
we pick our postcards to send to this one friend does not matter to our
friend, as they will receive them all at once anyways! Therefore, our
process above is over-counting the total number of ways to send out
postcards: it would think that sending card X and Y is a different action
to sending card Y and card X!
To fix this, we need to correct for our over-counting errors above. Notice
that for any given set of k distinct cards, there are k! different ways to
order that set: this is because in ordering a set of k things, you make
k choices for where to place the first element, k − 1 choices for where to
place the 2nd element, and so on / so forth until you have just 1 choice
for the k-th element.
Therefore, if we are looking at the collection of ordered length-k se-
quences of cards, each unordered sequence of k cards corresponds to
k! elements in this ordered sequence! That is, we have the following
equality:

Unordered ways to Ordered ways to

n!
(n
pick k cards ⋅k! = pick k cards =
from n choices from n choices − k)!

Therefore, if we want to only count the number of unordered ways to

pick k cards from n choices, we can simply divide both sides of the above
equation by k!, to get

Unordered ways to
n!
pick k cards =
from n choices k!(n − k)!

This concept — given a set of n things, in how many ways can we pick k
of them, if we don’t care about the order in which we pick those elements
— is an incredibly useful one, and as such leads itself to the following
definition:
Definition 3.2. (Unordered choice without repetition.) The bi-
nomial coefficient (nk) is the number of ways to choose k things from n
choices, if repeated choices are not allowed and the order of those choices
does not matter.

Observation 3.8. By the working above, we can see that ( ) =

n n!
k k!(n − k)!
for any natural numbers k, n with k ≤ n. For k > n, we have (nk) = 0 by
the same reasoning as before: we cannot choose more than n distinct
things from a set of n possibilities!

43

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This observation lets us solve a few more problems:

(3,3) (3,3)
Answer to Exercise 3.2. In this problem, we’re trying to count the
number of lattice paths from (0, 0) to (3, 3). How can we do this?
Well: notice that any path from (0, 0) to (3, 3) will need to take 6 steps;
of those 6 steps, precisely 3 must be to the right and the remaining 3 must
be upward. Therefore, we can create any such path by just “choosing”
(0,0) (0,0) 3 out of our 6 steps to be the rightward steps!
There are (63) many ways to choose such a set of rightward steps; this is
because we don’t repeat any of these choices (i.e. we have to pick three
different places to go right), and because the order in which we decide
which steps are rightward steps doesn’t matter (i.e. picking steps 2, 3
and 5 to be rightward steps is the same as picking steps 5, 2 and 3 to be
rightward steps.) Therefore, we have our answer: it’s (63) = 3!3!
6! 6⋅5⋅4
= 3⋅2⋅1 =
20 .
Problem 3.13. Let’s return to Problem 3.8, and tweak its setup a bit:
what’s the probability that no-one picks up their own hat?

Answer. Like in Problem 3.8, we still have the same universe of 24

possible ways for hats to be distributed. However, the set of situations
in which no-one picks up their own hat is harder to measure!
One solution here could be to just list out via brute-force all of the
possible ways to shuffle hats around without repeats. This could work;
try it, if you want!
A second, clever trick (that would let us solve this problem for any
number of friends) is the following:

Ways in which for All ways to

All ways
no-one gets = - return hats where
to return hats
their own hat someone gets their own hat

On one hand, this is kind of a dumb observation: we’re just saying

that the set of situations where our experiment succeeds is all of the
possibilities where our experiment does not fail. On the other hand,
though, this idea of “think about what happens where we do not succeed”
has already been proven to be quite useful: both in Problem 3.6, and
more generally when we’ve used the “suppose we’re wrong” argument
method in our first chapter! It’s a good trick, and it will serve us well
here.
In particular, suppose that we group our situations by cases, related to
the number of people who get their own hats back:
❼ 4 hats back to original owners: There’s just 1 way to do this,
as noted in our previous problem.
❼ 3 hats back to original owners: This is impossible! If three
people get their own hat back, then there’s only one hat left for
the fourth person, i.e. their own hat. So there are 0 ways for this
to happen.
❼ 2 hats back to original owners: To count this, first count the
number of ways to pick the lucky two people who get their own
hat back. This is (42) = 2!2!
4!
= 6, because we’re picking 2 people to
get their hats back from a set of 4, and order doesn’t matter / we
don’t repeat people.
With this done, there’s exactly one way for these hats to go back
to their owners in the desired fashion for each such pair: the two
chosen people get their own hat, and the other two swap hats.
Therefore, there are 6 ways for this to happen in total.

44

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ 1 hat back to its original owner: We count this in the same

way! There are 4 ways to pick a person to get their own hat back:
just pick one of the four people.
With this done, in how many ways can our remaining people shuffle
hats? Well: of the three remaining people, the first can choose
either of the other two people’s hats. The person whose hat was
not chosen then has no choice: to avoid taking their own hat, they
must take the first person’s hat. This leaves the third person with
no choice as well. Therefore, there are 2 ⋅ 1 ⋅ 1 ways for this to
happen for a given chosen person.
Across our four possible ways to choose people, then, there are
2 ⋅ 4 = 8 many ways for this to happen.
Therefore, the number of ways in which people get their own hats is
1 + 6 + 8 = 15, and so the number of ways in which this does not happen
9
is 24 − 15 = 9. Thus, our probability is 24 = 38 .
To illustrate one last counting process, we return to our postcard prob-
lem:
Problem 3.14. Suppose that we are at the shops and want to buy a
bunch of postcards to send out to our friends. The shop sells n different
kinds of postcards, and has tons of each kind. We want to buy k cards
(possibly with repetitions, if there’s a specific card design we like and
want to send to many people.) In how many ways can we pick out a set
of k cards to buy?

It first bears noting that this problem does not fall under the situations
of our earlier problems. In this problem, our choices are unordered:
i.e. we’re just picking out a bundle of cards to buy, and the order in
which they’re bought is irrelevant. Therefore, we cannot use the “or-
dered choice with repetitions” observation we made earlier, as this would
massively overcount things (i.e. we’d count orders of the same cards as
different if the cashier rang things up in a different order, which is silly.)
However, unlike our two “ordered/unordered choice without repeats”
situations, we can repeat choices! This means that this is not at all
like those situations: in particular, k can be larger than n and we will
still have lots of possibilities here, whereas in in the “without repeats”
situations this was always impossible. So we need a new method!
To develop this method, think of the n different kinds of postcards as n
“bins.” Here’s a visualization for when n = 5:

card card card card card

type 1 type 2 type 3 type 4 type 5

Picking out k cards to buy, then, can be thought of as pulling a few cards
from the first bin, a few from the second, and so on/so forth until we’ve
pulled out k cards in total. In other words, this is the same problem
as distributing k balls amongst n bins:

card card card card card

type 1 type 2 type 3 type 4 type 5

To do this task, replace the n bins with n − 1 “dividers” between our

choices. This separates our choices just as well as the bins did, so this
is still the same problem.

45

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Now, forget the difference between objects and dividers! That is,
take the diagram above and suppose that you cannot tell the difference
between an object and a divider between our choices.

? ? ? ? ? ? ? ? ? ? ? ? ?

How can we return this back to a way to choose k things from n choices?
Well: take the set of k + (n − 1) objects, of which k used to be things
and n − 1 were dividers. Now choose n − 1 of them to be dividers! This
returns us back to a way to pick out k things from n choices.

? ? ? ? ? ? ? ? ? ? ? ? ?

In particular, note that given any set of k + (n − 1) placeholders, we

can turn it into a way to choose k things from n choices with repetition
by performing such a choice! Therefore, there are as many ways to
make such choices as there are ways to choose n − 1 things from a set of
k + (n − 1) options to be placeholders. This second choice is unordered
and without repetition (we want all of the placeholders to be different,
and don’t care about the order in which we pick the placeholders: just
the elements that are chosen!) Therefore, we can use our “unordered
choice without repetition” principle to see that there are (k+n−1
n−1
) many
ways to do this!
In other words, we have the following observation:

Observation 3.9. (Unordered choice with repetition.) The number

of ways to choose k things from n choices, where we do not care about
the order in which we make our choices but allow choices to be repeated,
is (k+n−1
n−1
).

To practice this, let’s answer our last two problems of this chapter:

Problem 3.15. Suppose that you have ten identical cookies, and want
to distribute them to four of our friends, so that each friend gets at least
one cookie. In how many ways can we do this?

Answer. First, if every friend gets at least one cookie, we can start by
just distributing one cookie to each friend! This leaves us with 6 cookies
left over to distribute further to our friends.
If we think of taking each cookie and “choosing” a friend to give it to, this
is unordered choice with repetition: the order doesn’t matter because the
cookies are identical, and repeats are allowed because we can give one
friend multiple cookies.
By our formula above, then, there are (6+4−1
4−1
) = (93) = 9⋅8⋅7
3⋅2
= 84 ways
for this to happen.

Answer to Exercise 3.3. In this problem, we’re trying to determine

how many seven-digit phone numbers exist, in which the digits are non-
decreasing? (By “nondecreasing” here, we just mean that each digit is at
least as big to the digit to its left; i.e. 122-2559 is a valid phone number,
but 321-1234 would not be.)

46

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

With this understood, we claim that our problem can be reduced to an

“unordered choice with repetition” task as follows: consider any way to
choose seven numbers from the set of digits {0, 1, . . . 9}, without caring
about the order and with repetition allowed.
On one hand, we claim that any such choice can be turned into a non-
decreasing phone number! Just list the digits here in order of their size;
i.e. if you picked three 1’s, a 2, a 3, and two 5’s, write down 111-2355.
This process also clearly generates any such phone number (just pick its
digits!), and so the number of seven-digit nondecreasing phone numbers
is just the number of unordered ways to choose seven things from the
set of digits {0, 1, . . . 9} with repetition.
By our above formula, there are (7+10−1 10−1
) = (16
9
) = 11440 many such
numbers. Success!

3.4 Practice Problems

1. (-) How many binary strings of length 10 exist?
2. How many three-digit numbers exist where all of the digits are
different? (Hint: the answer is not 10 ⋅ 9 ⋅ 8. . . )
3. You have ten indistinguishable cookies that you’ve baked for four
of your friends Jianbei, Sione, Sina and Julia. You want to give
away all ten of your cookies to your friends, and for each friend to
get at least one cookie. You also know that Sione wants an even
number of cookies, so he can share them with a friend.
In how many ways can you give away your cookies?
4. You’re a Pokémon master! You have collected exactly one of all 151
Pokémon in the base game. You want to assemble a team to take on
the Elite Four: doing this involves choosing a team of 6 Pokémon
(in which the order does not matter) and then designating one of
those six to be the “lead” on your team (i.e. the first one to go
out.) In how many ways can you do this?
5. For any two natural numbers n, k ∈ N, let C(n, k) denote the num-
ber of ways to choose k unordered objects from a set of n distinct
objects without repeats, and let P (n, k) denote the number of ways
to choose k ordered objects from a set of n distinct objects with-
out repeats.
For what values of n, k is C(n, k) = P (n, k)?
6. (-) Take the collection of all length-10 binary strings and pick one
at random. What are the odds that the sum of all digits in your
string is even? Does this change if you were looking at length-11
binary strings?
7. Take a standard 52-card deck of playing cards, shuffle it, and draw
a hand of five cards. What is the probability that your hand has
a three-of-a-kind?
8. Can you find three events A, B, C such that
❼ the probability of A and B happening at the same time is 12 ,
❼ the probability of A and C happening at the same time is 21 ,
❼ the probability of B and C happening at the same time is 21 ,
and yet somehow
❼ the probability that all three of A, B, C happening at the
same time is 0?

47

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

9. (+) Suppose that there’s an election! Two candidates, Sherlock

and Moriarty, are running for office. Suppose that Sherlock receives
8 votes and Moriarty receives 7 votes, and that these votes are
being counted up one-by-one to create a running total.
What is the probability that Sherlock is never behind in this run-
ning total? In general, if Sherlock got s votes and Moriarty got m
votes, what is this probability?
10. (+) Auckland has about 1 million residents. Suppose each resident
has a jar with 100 coins in it.
Two jars are considered to be “equivalent” if they have the same
number of 10c, 20c, 50c, ✩1 and ✩2 coins in them.
How many different (i.e. nonequivalent) jars of coins exist? Is it
theoretically possible that every person in Auckland has a different
jar of coins?

48

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

49

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Algorithms and Functions

2019 Chapter 4

Exercise 4.1. Two processes that you can use to multiply two nonneg-
An example run of Algorithm 4.1 when ative integers a, b together are listed below:
a = 3, b = 12:
Algorithm 4.1.
step a b prod
1 3 12 0 1. Define a new number prod, and initialize it (i.e. set it equal) to 0.
2
3 2 12 2. If a = 0, stop, and return the number prod.
2
3. Otherwise, add b to prod, and subtract 1 from a. Then go to 2.
3 1 24
2
Algorithm 4.2. 1. Define a new number prod, and initialize it (i.e.
3 0 36
2 (halt!) set it equal) to 0.
2. If a = 0, stop, and return the number prod.
Similarly, an example run of Algorithm
4.2 when a = 3, b = 12: 3. Otherwise, if a is odd, subtract 1 from a and set prod = prod + b.
step a b prod 4. Divide a by 2, and multiply b by 2.
1 3 12 0
2
5. Go to step 2.
3 2 12
4 1 24 In general, which of these is the faster way to multiply two numbers?
2 Why?
3 0 36
4 0 48
2 (halt!)
4.1 Functions in General
In your high-school mathematics classes, you’ve likely seen functions de-
scribed as things like “f (x) = 2x +3” or “g(x) = max(x, y).” When we’re
writing code, however, we don’t do this! That is: in most programming
languages, you can’t just type in expressions like the ones before and
trust that the computer will understand what you mean. Instead, you’ll
often write something like the following:

/* function returning the max between two numbers */

int max ( int num1 , int num2 )
{
/* local variable declaration */
int result ;

if ( num1 > num2 )

result = num1 ;
else
result = num2 ;

return result ;
}

Notice how in this example we didn’t just get to define the rules for our
function: we also had to specify the kind of inputs the function should
expect, and also the type of outputs the function will generate! On one
hand, this creates more work for us at the start: we can’t just tell people
our rules, and we’ll often find ourselves having to go back and edit our
functions as we learn what sorts of inputs we actually want to generate.
On the other hand, though, this lets us describe a much broader class
of functions than what we could do before! Under our old high-school

50

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

definition for a function, we just assumed that functions took in num-

bers and returned numbers. With the above idea, though, we can have
functions take in and output anything: multiple numbers, arrays, text
files, whatever!
On the third(?) hand, enforcing certain restrictions on the types of
inputs and outputs to a function is also a much more secure way to
think about functions in computer science. If you’re writing code in a
real-life situation, you should always expect malicious or just clueless
users to try to input the worst possible data to your functions. As a
result, you can’t just have a function defined by the rule f (x) = x1 and
trust that your users out of the goodness of their hearts will never input
0 just to see what happens! They’ll do it immediately (as well as lots of
other horrifying inputs, like , 10 , 1 − 0.9, . . . ) just to see what happens.
https://xkcd.com/327/
This is why many programming languages enforce type systems: i.e.
rules around their functions that specifically force you to declare the
kinds of inputs and outputs ahead of time, like we’ve done above! Doing
this is an important part of writing bug-free and secure code.
As this is a computer science class, we should have a definition of function
that matches this concept. We provide this here:

Definition 4.1. Formally, a function consists of three parts:

❼ A collection A of possible inputs. We call this the domain of our
function.
❼ A collection B describing the type of outputs that our function
will generate. We call this the codomain of our function.
❼ A rule f that takes in inputs from A and generates outputs in B.
Furthermore, in order for this all to be a function, we need it to satisfy
the following property:

For every potential input a from A, there should be exactly one b in

B such that f (a) = b.

In other words, we never have a value a in A for which f (a) is undefined,

as that would cause our programs to crash! As well, we also do not allow
for a value a ∈ A to generate “multiple” outputs; i.e. we want to be able
to rely on f (a) not changing on us without warning, if we keep a the
same.

Example 4.1. Typically, to define a function we’ll write something like

“Consider the function f ∶ Z → Q, defined by the rule f (n) = n21+1 . This
definition tells you three things: what the domain is (the set the arrow
starts from, which is the integers in this case), what the codomain is
(the set the arrow points to, which is the rational numbers in this case),
and the rule used to define f .

Example 4.2. f ∶ Z → Z defined by the rule “f (x) = y if and only if

x = y 2 ” is not a function. There are many reasons for this:
❼ There are values in the domain that do not get mapped to any
values in the codomain by our rule. For instance, consider x = −1 ∈
Z. There is no value y ∈ R such that −1 = y 2 , because no integer
when squared is negative! Therefore, x is not mapped to any value
y in the codomain, and so we do not regard f as a function.
❼ There are also values in the domain that get mapped to multiple
values in the codomain by our rule. For instance, consider x = 1 ∈ Z.
Because 1 = y 2 has the two solutions y = ±1, this rule maps x = 1
to the two values y = ±1. This is another reason why f is not a
function!

51

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Example 4.3. Let A be the set of all students at the University of

Auckland, and B be the set of all integers. We can define a function f ∶
A → B by defining f (a) to be equal to that student’s current ID number.
This is a function, because each student has a unique ID number!
However, if we tried to define a function g ∶ B → A by the rule g(b) = the
student whose ID number is b, we would fail! This is because there are
many integers that are not ID numbers: for example, no student has ID
number −1, or 10100 .

While the objects above have had relatively “nice” rules, not all functions
can be described quite so cleanly! Consider the following example:

Example 4.4. Let A be the collection { , , , } of the Canada,

cat, bird, and kiwi emojis, and let B be the collection { , , , }
Domain of the octopus, radiation, bee, and man emojis. Consider f ∶ A → B,
f defined by the rules
Codomain
f( )= , f( )= , f( )= , f( )= , f( )=

This is a function, because we have given an output for every possible

input, and also never sends an input to multiple different outputs. It’s
not a function with a simple algebraic rule like “x2 + 2x − 1”, but that’s
OK!
A useful way to visualize functions defined in this piece-by-piece fashion
is with a diagram: draw the domain at left, the codomain at right, and
draw an arrow from each x in the domain to its corresponding element
f (x) in the codomain.

Alongside the domain/codomain ideas above, another useful idea here is

the concept of range:

Definition 4.2. Take any function f ∶ A → B. We define the range

of f as the set of all values in the codomain that our function actually
sends values in the domain to. In other words, the range of f is the
following set:

{b ∈ B ∣ there is some a ∈ A such that f (a) = b}

Note that the range is usually different to the codomain! In the examples
we studied earlier, we saw the following:
❼ f ∶ Z → Q defined by the rule f (n) = 2
1
does not output every
n +1
rational number! Amongst other values, it will never output any
1 1
number greater than 1 (as 2 ≤ 2 = 1 for every integer n.)
n +1 0 +1
As such, its codomain (Q) is not equal to its range.
❼ The function f ∶ A → B from Example 4.3, that takes in any
student at the University of Auckland and outputs their student
ID, does not output every integer: amongst other values, it will
never output a negative integer! As such, its codomain (B) is not
equal to its range.

❼ The emoji function in Example 4.4 never outputs , even though

it’s in the codomain.
Intuitively: we think of the codomain as letting us know what type of
outputs we should expect. That is: in both mathematics or computer
science, often just knowing that the output is “an integer” or “a binary
string of length at most 20” or “a Unicode character” is enough for our
compiler to work. As such, it’s often much faster to just describe the

52

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

type of possible outputs, instead of laboriously finding out exactly what

outputs are possible!
However, in some cases we will want to know precisely what values we get
as outputs, and in that situation we will want to find the actual outputs:
i.e. the range. To illustrate this, let’s consider a few examples:
Example 4.5. Consider the function f ∶ Z → Z given by the rule f (x) =
2∣x∣ + 2. This function has range equal to all even numbers that are at
least 2.
To see why, simply notice that for any integer x, ∣x∣ is the “absolute
value” of x: i.e. it’s x if we remove its sign so that it’s always nonneg-
ative. As a result, 2∣x∣ is always a nonnegative even number, and this
2∣x∣ + 2 must be a nonnegative even number that’s at least 2.
That tells us that the only possible outputs are even numbers that are
at least 2! However, we still don’t know that all of those outputs are
ones that we actually can get as outputs.
To see why this is true: take any even number that is at least 2. By
definition, we can write this as 2k, for some k ≥ 1. Rewrite this as
2(k − 1) + 2; if we do so, then we can see that f (k − 1) = 2∣k − 1∣ + 2 =
2(k − 1) + 2 (because if k ≥ 1, then k − 1 ≥ 0 and so ∣k − 1∣ = k − 1.) As a
result, we’ve shown that f (k − 1) = 2k for any k ≥ 1, and thus that we
Domain
can actually get any even number that’s at least 2 as an output. f Codomain

Example 4.6. Consider the emoji function from Example 4.4. If we look
at the diagram we drew before, we can see that our function generates
three possible outputs: , and . Therefore, the collection of
these three emojis is our range!
Example 4.7. Let A be the collection of all pairs of words in the En-
glish language, and B be the two values {true, false}. Define the func-
tion f ∶ A → B by saying that f (w1 , w2 ) = true if the words w1 , w2
rhyme, and false otherwise. For example, f (cat, bat) = true, while f (cat,
cataclysm) = is false.
The range of this function is {true, false}, i.e. the same as its codomain!
It is possible for the range and codomain to agree. (If this happens, we
call such a function a surjective function. We’re not going to focus
on these functions here, but you’ll see more about them in courses like
Compsci 225 and Maths 120/130!)
Example 4.8. Let R denote the set of all real numbers (i.e. all numbers
regardless of whether they’re rational or irrational; alternately, anything
you can describe with a decimal expansion.) Define the function f ∶ R →
R by the rule f (x) = 2x . 3

This function has range equal to the set of all positive numbers! This 2

takes more math to see than we currently have: again, take things like
Maths 130 to see the “why” behind this. However, if you draw a graph 1

of 2x you’ll see that the outputs (i.e. y-values) range over all of the 0
-6 -5 -4 -3 -2 -1 0 1 2

possible positive numbers, as claimed.

To close this section, we give a useful bit of notation for talking about
functions defined in terms of other functions: function composition.
Fact 4.2. Given any two functions f ∶ B → C, g ∶ A → B, we can combine
these functions via function composition: that is, we can define the
function f ○ g ∶ A → C, defined by the rule f ○ g(x) = f (g(x)). We
pronounce the small open circle symbol ○ as “composed with.”
Example 4.9. ❼ If f, g ∶ R → R are defined by the rules f (x) = x + 1
and g(x) = x2 −1, then we would have g○f (x) = g(f (x)) = g(x+1) =
(x + 1)2 − 1 = x2 + 2x.

53

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ Notice that this is different to f ○ g(x) = f (g(x)) = f (x2 − 1) =

(x2 − 1) + 1 = x2 !
In general, f ○g and g ○f are usually different functions: make sure
to be careful with the order in which you compose functions.
❼ If f, g ∶ R → R are defined by the rules f (x) = 3x−1 and g(x) = x+1
,
then f ○ g(x) = f (g(x)) = f ( x+1 ) = 3 x+1 − 1 = (x + 1) − 1 = x.
3
Here, R+ denotes the set of all positive 3 3
real numbers.
❼ If f ∶ R → R+ and g ∶ R+ → R are defined by the rules f (x) = 2x +1
2

and g(x) = log2 (x) − 1, then g ○ f (x) = g(f (x)) = g (2x +1 ) =

2

log2 (2x ) − 1 = (x2 + 1) − 1 = x2 .

2
+1

Handy!
Notice that in the definition above, we required that the domain of f was
the codomain of g. That is: if we wanted to study f ○ g(x) = f (g(x)),
we needed to ensure that every output of g is a valid input to f .
This makes sense! If you tried to compose functions whose domains and
codomains did not match up in this fashion, you’d get nonsense / crashes
when the inner function g returns an output at which the outer function
is undefined. For example:
Example 4.10. ❼ If f ∶ R ∖ {0} → R is defined by the rule f (x) = x1
and g ∶ R → R is defined by the rule g(x) = x2 − 1, then you might
think that f ○ g(x) = f (g(x)) = x21−1 .
However, this is not a function! When x = ±1, for example, we
have f (g(±1)) = (±1)12 −1 = 10 , which is undefined. This is why we
insist that the codomain of g is the domain of f ; we need all of g’s
outputs to be valid inputs to f .
❼ Let A be the set of all people in your tutorial room, B be the set
of all ID numbers of UoA students, and C be the set of all ID
numbers of Compsci 120 students. Then f ∶ C → R defined by
taking any Compsci 120 student’s ID and outputting their grade
on the mid-sem test is a function; as well, g ∶ A → B, defined by
mapping each person in your tutorial room to their ID number is
a function.
However, f ○ g, the function that tries to take each person in your
tutorial room and output their mid-sem test score, is undefined! In
particular, your tutor is someone in your tutorial room, who even
though they do have an ID number, will not have a score on the
mid-sem test. Another reason to insist that the codomain of g is
the domain of f !

4.2 Algorithms
In the previous section, we came up with a “general” concept for function
that we claimed would be better for our needs in computer science, as
it would let us think of things like

/* function returning the max between two numbers */

int max ( int num1 , int num2 )
{
/* local variable declaration */
int result ;

if ( num1 > num2 )

result = num1 ;

54

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

else
result = num2 ;

return result ;
}

as a function. However, most of the examples we studied in this chapter

didn’t feel too much like the code above: they were either fairly mathe-
matical in nature (i.e. f (n) = n21+1 ) or word-problem-oriented (i.e. the
function that sent UoA students to their ID numbers.)
To fix this issue, this section will focus on the idea of an algorithm:
that is, a way of describing in general a step-by-step problem-solving
process that we can easily turn into code. An algorithm for pancakes!
We start by describing what an algorithm is:
1. Acquire and measure out the
Definition 4.3. An algorithm is a precise and unambiguous set of following ingredients:
instructions. ❼ 2 cups of buttermilk, or 1.5
cups milk + .5 cups yoghurt
Typically, people think of algorithms as a set of instructions for solving whisked together.
❼ 2 cups of flour.
some problem; when they do so, they typically have some restrictions in ❼ 2 tablespoons of sugar.
mind for the kinds of instructions they consider to be valid. For example, ❼ 2 teaspoons of baking powder.
consider the following algorithm for proving the Riemann hypothesis: ❼ 1/2 teaspoon of baking soda.
❼ 1/2 teaspoon of salt.
1. Prove the Riemann hypothesis. ❼ 1 large egg.
2. Rejoice! ❼ 3 tablespoons butter.
❼ Additional butter.
On one hand, this is a fairly precise and unambiguous set of instructions: ❼ Maple syrup.
step 1 has us come up with a proof of the Riemann hypothesis, and step 2. Whisk the flour, sugar, baking
2 tells us to rejoice. powder, baking soda, and salt
in a medium bowl.
On the other hand: this is not a terribly useful algorithm. In particular, 3. Melt the 3 tablespoons of but-
its steps are in some sense “too big” to be of any use: they reduce the ter.
problem of proving the Riemann hypothesis to . . . proving the Riemann 4. Whisk the egg and melted but-
hypothesis. Typically, we’ll want to limit the steps in our algorithms ter into the milk until com-
bined.
to simple, mechanically-reproducible steps: i.e. operations that a com-
5. Pour the milk mixture into the
puter could easily perform, or operations that a person could do with dry ingredients, and whisk un-
relatively minimal training. til just combined (a few lumps
In practice, the definition of “simple” depends on the context in which should remain.)
6. Heat a nonstick griddle/fry-
you are creating your algorithm. Consider the algorithm for making
pan on medium heat until hot;
delicious pancakes, given at right. This algorithm’s notion of “simple” grease with a teaspoon or two
is someone who is (1) able to measure out quantities of various foods, of butter.
and (2) knows the meaning of various culinary operations like “whisk” 7. Pour 1/4 cup of batter onto the
and “flip.” If we wanted, we could make an algorithm that includes skillet. Repeat in various dis-
joint places until there is no
additional steps that define “whisking” and “flipping”. That is: at each spare room on the skillet. Leave
step where we told someone to whisk the flour, we could instead have gaps of 1cm between pancakes.
given them the following set of instructions: 8. Cook until large bubbles form
and the edges set (i.e. turn a
(a) Grab a whisk. If you do not know what a whisk is, go to this slightly darker color and are no
Wikipedia article and grab the closest thing to a whisk that you longer liquid,) about 2 minutes.
can find. A fork will work if it is all that you can find. 9. Using a spatula, flip pancakes,
and cook a little less than 2
(b) Insert the whisk into the object you are whisking. minutes longer, until golden
brown.
(c) Move the whisk around through the object you are whisking in
10. If there is still unused batter,
counterclockwise circles of varying diameter, in such a fashion to go to 5; else, top pancakes with
mix together the contents of the whisked object. maple syrup and butter, and
eat.
In this sense, we can extend our earlier algorithm to reflect a different
notion of “simple,” where we no longer assume that our person knows
how to whisk things. It still describes the same sets of steps, and in this This is a good recipe. Use it!
sense is still the “same” algorithm – it just has more detail now!

55

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This concept of “adding” or “removing” detail from an algorithm isn’t

something that will always work; some algorithms will simply demand
steps that cannot be implemented on some systems. For example, no
matter how many times you type “sudo apt-get 2 cups of flour,” your
laptop isn’t going to be able to implement our above pancake algorithm.
As well, there may be times where a step that was previously consid-
ered “simple” becomes hideously complex on the system you’re trying
to implement it on!
We’re not going to worry too much about the precise definition of “sim-
ple” in this class, because we’re not writing any code here (and so our
notion of “simple” isn’t one we can precisely nail down) — these are the
details we’ll leave for your more coding-intensive courses.
Instead, let’s just look at a few examples of algorithms! We’ve already
seen one in this class, when we defined the % operation:
Algorithm 4.3. This algorithm takes in any two integers a, n, where
n > 0. It then calculates a % n as follows:
❼ If a ≥ n, we repeatedly subtract n from a until a < n, and return
the end result.
❼ If a < 0, repeatedly add n to a until a > 0, and return the end
result.
❼ If neither of these cases apply, then we just return a.

Second, we can turn Claim 1.6 into an algorithm for how to tell if a
number is prime:
Algorithm 4.4. This is an algorithm that takes in a positive integer n,
and determines whether or not n is prime. It proceeds as follows:
❼ If n = 1, stop: n is not prime.
√
❼ Otherwise, if n > 1, find all of the numbers 2, 3, 4, . . . ⌊ n⌋. Take
each of these numbers, and test whether they divide n.
❼ If one of them does, then n is not prime!
❼ Otherwise, if none of them divide n, then by Claim 1.6, n is prime.

This is a step-by-step process that tells us if a number is a prime or not!

Notice that the algorithm itself didn’t need to contain a proof of Claim
1.6; it just has to give us instructions for how to complete a task, and not
justify why those instructions will work. It is good form to provide such a
justification where possible, as it will help others understand your code!
However, it is worth noting that such a justification is separate from the
algorithm itself: it is quite possible (and indeed, all too easy) to write
something that works even though you don’t necessarily understand why.
For a third example, let’s consider an algorithm to sort a list:
Algorithm 4.5. The following algorithm, InsertionSort(L), takes in
a list L = (l1 , l2 , . . . ln ) of n numbers and orders it from least to great-
est. For example, InsertionSort(1, 7, 1, 0) is (0, 1, 1, 7). It does this by
using the following algorithm:
1. If L contains at most one number, L is trivially sorted! In this
situation, stop.
2. Otherwise, L contains at least two numbers. Let L = (l1 , l2 , . . . ln ),
where n ≥ 2. Define a pair of values valmin , locmin , and set them
equal to the value and location of the first element in our list.
3. One by one, starting with the second entry in our list and work-
ing our way through our entire list L, compare the value stored in
valmin to the current value lk that we’re examining.

56

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

(a) If valmin > lk , update valmin to be equal to lk , and update

locmin to be equal to k.
(b) Otherwise, just go on to the next value.
4. At the end of this process, valmin and locmin describe the value and
the location of the smallest element in our list. Swap the first value
in our list with llocmin in our list: this makes the first value in our
list the smallest element in our list.
5. To finish, set the first element of our list aside and run InsertionSort
on the rest of our list.

Back in our first chapter, to understand our two previous algorithms 4.3
and 4.4 we started by running these algorithms on a few example inputs!
In general, this is a good tactic to use when studying algorithms; actually
plugging in some concrete inputs can make othewise-obscure instructions
much simpler to understand.
To do so here, let’s run our algorithm on the list (1, 7, 1, 0), following
each step as written:
list step locmin valmin current k current lk list step locmin valmin current k current lk
(1,7,1,0) 1 (0, 7, 1, 1 ) 3(a) 3 1
(1,7,1,0) 2 1 1 (0, 7, 1, 1 ) 3 4 1
(1,7,1,0) 3 2 7
(0, 1, 7, 1 ) 4
(1,7,1,0) 3 3 1
(0,1, 7, 1 ) 5
(1,7,1,0) 3 4 0
(1,7,1,0) 3(a) 4 0 (0,1, 7, 1 ) 1
(0,7,1,1) 4 (0,1, 7, 1 ) 2 3 7
(0, 7, 1, 1 ) 5 (0,1, 7, 1 ) 3 4 1
(0, 7, 1, 1 ) 1 (0,1, 7, 1 ) 3(a) 4 1
(0, 7, 1, 1 ) 2 2 7 (0,1, 1, 7 ) 4
(0, 7, 1, 1 ) 3 3 1 (0,1,1, 7 ) 5
(0,1,1, 7 ) 1

Here, we use the 7, 1, 1 , 7, 1 and 7 boxes to visualize the “rest of our

list” part of step 5 in our algorithm.
This worked! Moreover, doing this by hand can help us see an argument
for why this algorithm works:

Claim 4.1. InsertionSort (i.e. algorithm 4.5) works.

Proof. In general, there are three things we need to check to show that
a given algorithm works:
❼ The algorithm doesn’t have any bugs: i.e. every step of the
process is defined, you don’t have any division by zero things or
undefined cases, or stuff like that.
This is true here! The only steps we perform in this algorithm
are comparisons and swaps, and the only case we encounter is
“valmin > lk is true” or “valmin > lk is false,” which clearly covers
all possible situations. As such, there are no undefined cases or
undefined operations.
❼ The algorithm doesn’t run forever: i.e. given a finite input,
the algorithm will eventually stop and not enter an infinite loop.
This is also true here! To see why, let’s track the number of com-
parisons and write operations performed by this process, given a
list of length n as input:
– Step 1: one comparison (we checked the size of the list.)
– Step 2: two write operations (we defined valmin , locmin .)

57

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

– Step 3: (n − 1) comparisons and possibly 2(n − 1) write oper-

ations.
To see why: note that for all of the n−1 entries in our list from
l2 onwards, we looked up the value in lk and compared it to
the value we have in valmin , which gives us n − 1 comparisons.
If it was smaller, we rewrote the values in valmin and locmin
in 3(a).
– Step 4: 2 write operations (to swap these values.)
– Step 5: 1 write operation (to resize the list to set the first
element aside), and however many operations we need to sort
a list of n − 1 numbers with InsertionSort.
In total, then, we have the following formula: if InsertionSortSteps(n)
denotes the maximum number of operations in total needed to sort
a list with our algorithm, then

InsertionSortSteps(n) = 1 + 2 + (n − 1) + 2(n − 1) + 2 + 1 + InsertionSortSteps(n − 1)

= 3n + 3 + InsertionSortSteps(n − 1).
= 3(n + 1) + InsertionSortSteps(n − 1).

At first, this looks scary: our function is defined in terms of itself!

In practice, though, this is fine. We know that InsertionSortSteps(1) =
1, because step 1 immediately ends our program if the list has size
1.
Therefore, our formula above tells us that if we set n = 2, we have

InsertionSortSteps(2) = (3 ⋅ (2 + 1)) + InsertionSortSteps(2 − 1) = 9 + 1 = 10 ,

by using our previously-determined value for InsertionSortSteps(1).

Still finite!
We can do this again for n = 3, and get

InsertionSortSteps(3) = (3 ⋅ (3 + 1)) + InsertionSortSteps(3 − 1) = 12 + 10 = 22 ,

by using our previously-determined value for InsertionSortSteps(2).

Again, still finite!
In general, if this process can sort a list of size n − 1 in finitely
many steps, then it only takes us 3n + 3 more steps to sort a list
of size n. In particular, there is no point at which our algorithm
jumps to needing “infinitely many” steps to sort a list!
❼ The algorithm produces the desired output: in this case, it
produces a list ordered from least to greatest.
This happens! To see why, notice that in step 4, we always make
the first element of the list we’re currently sorting the smallest
element in our list. Therefore, on our first application of step 4,
we have ensured that l1 is the smallest element of our list.
On the second application of step 4, we were sorting the list starting
from its second element: doing so ensures that l2 is smaller than
all of the remaining elements.
On the third application of step 4, we had set the first two elements
aside and were sorting the list starting from its third element: doing
so ensures that l3 is smaller than all of the remaining elements.
. . . in general, on the k-th application of step 4, we had set the first
k − 1 elements aside and were sorting the list starting from its k-th
element! Again, doing so ensures that lk is smaller than all of the
remaining elements.

58

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

So, in total, what does this mean? Well: by the above, we know
that l1 ≤ l2 ≤ l3 ≤ . . . ≤ ln , as by definition each element is smaller
than all of the ones that come afterwards. In other words, this list
is sorted!

4.3 Recursion, Composition, and Algorithms

Algorithm 4.5 had an interesting element to its structure: in its fifth
step, it contained a reference to itself! This sort of thing might feel like
circular reasoning: how can you define an object in terms of itself?
However, if you think about it for a bit, this sort of thing is entirely
natural: many tasks and processes in life consist of “self-referential”
that are defined by self-reference! We call such definitions recursive
definitions, and give a few examples here: If you know some linear algebra, the
explicit formulas we are using here are
Example 4.11. Fractals, and the many plants and living things that the following: for each point (x, y) in
S, draw the four points
have fractal-like patterns built into theirselves, are examples of recursively-
❼ [ ] [ ] + [ ],
0.85 0.04 x 0
defined objects! For example, consider the following recursive process: −0.04 0.85 y 1.6
1. Start by drawing any shape S0 . For example, here’s a very stylized
❼ [ ][ ] + [ ]
0.2 −0.26 x 0
drawing of a cluster of fern spores: 0.23 0.22 y 1.6

❼ [ ][ ] + [ ]
−0.15 0.28 x 0
0.26 0.24 y 0.44

❼ [ ][ ] + [ ].
0 0 x 0
0 0.16 y 0
This is the precise math-y way of de-
scribing the operations we’re doing to
these rectangles!

2. Now, given any shape S, we define T (S) as the shape made by

making four copies of S and manipulating them as follows: if S
is a shape contained within the gray rectangle at left, we make
four copies of S appropriately scaled/stretched/etc to match the
four rectangles at right. (It can be hard to see the black rectangle,
because it’s so squished: it’s the stem-looking bit at the bottom-
middle.)

For example, T (S0 ), i.e. T applied to our fern spore shape, is the
following:

59

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

3. Using our “seed” shape S0 and our function T , we then recur-

sively define Sn as T (Sn−1 ) for every positive integer n. That is:
S1 = T (S0 ), S2 = T (S1 ), etc. This is a recursive definition because
we’re defining our shapes in terms of previous shapes! Using the
language of function composition, we can express this all at once
by writing Sn = T ○ T ○ . . . ○ T (S0 ).
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
n times
4. Now, notice what these shapes look like as we draw several of them
in a row:

Our seed grows into a fern!

This is not a biology class, and there are many open questions about
precisely how plants use DNA to grow from a seed. However, the idea
that many plants can be formed by taking a simple instruction (given one
copy of a thing, split it into appropriately stretched/placed copies of that
thing) and repeatedly applying it is one that should seem reasonable,
given the number of places you see it in the world!
Example 4.12. On the less beautiful but more practical side, recursion
is baked into many fundamental mathematical operations! For example,
think about how you’d calculate 11 ⋅ 13. By definition, because multipli-
cation is just repeated addition, you could calculate this as follows:

´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
11 ⋅ 13 = 13 + 13 + 13 + 13 + 13 + 13 + 13 + 13 + 13 + 13 + 13 = 143.
11 times

Now, however, suppose that someone asked you to calculate 12⋅13. While
you could use the process above to find this product, you could also
shortcut this process by noting that

12 ⋅ 13 = (1 + 11) ⋅ 13 = 13 + 11 ⋅ 13 = 13 + 143 = 156.

This sort of trick is essentially recursion! That is: if you want, you could
define the product n ⋅ k recursively for any nonnegative integer n by the
following two-step process:
❼ 0 ⋅ k = 0, for every k.
❼ For any n ≥ 1, n ⋅ k = k + (n − 1) ⋅ k.
The second bullet point is a recursive definition, because we defined
multiplication in terms of itself! In other words, when we say that 12⋅13 =
13 + 11 ⋅ 13, we’re really thinking of multiplication recursively: we’re not
trying to calculate the whole thing out all at once, but instead are trying
to just relate our problem to a previously-worked example.
Exponentiation does a similar thing! Because exponentiation is just
repeated multiplication, we know that

210 = 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 ⋅ 2 = 1024.
´¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¸¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¶
10 times

Similarly, though, we can define exponentiation recursively by saying

that for any nonnegative integer n and nonzero number a, that

60

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ a0 = 1, and
❼ For any n ≥ 1, an = a ⋅ an−1 . In classes like Compsci 220 / 320,
In other words, with this idea, we can just say that you’ll study the idea of efficiency in
depth, and come up with more sophis-
ticated ideas than this one! In most
211 = 2 ⋅ 210 = 2 ⋅ 1024 = 2048, practical real-life situations there are
better ways to implement multiplica-
instead of calculating the entire thing from scratch! tion and exponentiation than this re-
cursive idea; however, it can be useful
These ideas can be useful if you’re working with code where you’re calcu-
in some places, and the general prin-
lating a bunch of fairly similar products/exponents/etc. With recursion, ciple of “storing commonly-calculated
you can store some precalculated values and just do a few extra steps values and extrapolating other values
of working rather than doing the whole thing out by scratch each time. from those recursively” is one that does
come up in lots of places!
Efficiency!

Recursion also comes up in essentially every dynamical system that mod-

els the world around us! For instance, consider the following population
modelling problem:
Example 4.13. Suppose that we have a petri dish in which we’re grow-
ing a population of amoebae, each of which can be in two possible states
(small and large).
Amoebas grow as follows: if an amoeba is small at some time t, then at
time t + 1 it becomes large, by eating food around it. If an amoeba is
large at some time t, then at time t + 1 it splits into one large amoeba
and one small amoeba.
Suppose our petri dish starts out with one small amoeba at time t = 1.
How many amoebae in total will be in this dish at time t = n, for any
natural number n?

Answer. To help find an answer, let’s make a chart of our amoeba

populations over the first six time steps:

1 2 3 4 5 6
Large 0 1 1 2 3 5
Small 1 0 1 1 2 3
Total 1 1 2 3 5 8

This chart lets us make the following observations:

1. The number of large amoebae at time n is the total number of
amoebae at time n − 1. This is because every amoeba at time n − 1
either grows into a large amoeba, or already was a large amoeba!
2. The number of small amoebae at time n is the number of large
amoebae at time n − 1. This is because the only source of small
amoebae are the large amoebae from the earlier step when they
split!
3. By combining 1 and 2 together, we can observe that the number of
small amoebae at time n is the total number of amoebae at time
n − 2!
4. Consequently, because we can count the total number of amoebae
by adding the large amoebae to the small amoebae, we can con-
clude that the total number of amoebae at time n is the total
number of amoebae at time n−1, plus the total number of amoebae
at time n − 2. In symbols,

A(n) = A(n − 1) + A(n − 2).

We call relations like the one above recurrence relations, and will
study them in greater depth when we get to induction as a proof method.

61

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

For now, though, notice that they are remarkably useful objects! By
These are particularly cool because generalizing the model above (i.e. subtracting an−3 to allow for old age
they actually accurately model preda- killing off amoebas, or having a −ca2n term to denote that as the popu-
tor/prey relations in real life! See Isle
Royale, amongst other examples.
lation grows too large, predation or starvation will cause the population
to die off, etc.) one can basically model an arbitrarily-complicated pop-
For more details on these sorts of ulation over the long term.
population-modelling processes, see
the logistic map, the Lotka-Volterra
equations, and most of the applied
mathematics major here at Auckland!
It bears noting that the amoeba recurrence relation is not the first recur-
rence relation you’ve seen! If you go back to Algorithm 4.5, we proved
that

InsertionSortSteps(n) = 3(n + 1) + InsertionSortSteps(n − 1).

This is another recurrence relation: it describes the maximum number

of operations needed to sort a list of size n in terms of the maximum
number of operations needed to sort a list of size n − 1.
While recurrence relations are nice, they can be a little annoying to
work with directly. For example, suppose that someone asked us what
InsertionSortSteps(12) is. Because we don’t have a non-recursive
formula, the only thing we could do here is just keep recursively applying
our formula, to get

InsertionSortSteps(12) = 3(12 + 1) + InsertionSortSteps(11)

= 3(12 + 1) + 3(11 + 1) + InsertionSortSteps(10)
= 3(12 + 1) + 3(11 + 1) + 3(10 + 1) + InsertionSortSteps(9)
= ...
= 3(12 + 1) + 3(11 + 1) + 3(10 + 1) + . . . + 3(3 + 1) + 3(2 + 1) + InsertionSortSteps(1)
= 3(12 + 1) + 3(11 + 1) + 3(10 + 1) + . . . + 3(3 + 1) + 3(2 + 1) + 1 = 265 .

This is . . . kinda tedious. It would be nice if we had a direct formula for

this: something like InsertionSortSteps(n) = 2n or n2 − 4n + 17 that
To get these skills, either take Maths we could just plug n into and get an answer.
120 and study linear algebra / eigen-
values, or take Compsci 220/320, or
At this point in time, we don’t have enough mathematics to directly find
take Maths 326 and learn about gen- such a “closed” form. However, we can still sometimes find an answer
erating functions! by just guessing. To be a bit more specific: if we use our definition, we
can calculate the following values for InsertionSortSteps(n):

n 1 2 3 4 5 6 7
InsertionSortSteps(n) 1 10 22 37 55 76 100

If you plug this into the Online Encyclopedia of Integer Sequences, it

will give you the following guess by comparing it to all of the sequences
it knows:

62

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This turns out to be correct!

Claim 4.2. For every positive integer n, we have InsertionSortSteps(n) =

3n2 + 9n − 10
.
2

We don’t have the techniques to prove this just yet. If you would like
to see a proof, though, skip ahead to the induction section of our proofs
chapter! We’ll tackle this problem there (along with another recurrence
relation), in Section 7.8.

4.4 Runtime and Algorithms

In the above few sections, we studied InsertionSortSteps(n) and ana-
lyzed the maximum number of operations it needs to sort a list of length
n. In general, this sort of run-time analysis of an algorithm — i.e.
the number of elementary operations needed for that algorithm to run
— is a very useful thing to be able to do!
To give a brief example of why this is useful, consider another, less well-
known algorithm that we can use to sort a list:

Algorithm 4.6. The following algorithm, BogoSort(L), takes in a list

L = (l1 , l2 , . . . ln ) of n numbers and orders it from least to greatest. It
does this by using the following algorithm:
1. One by one, starting with the first entry in our list and working
our way through our list, compare the values stored in consecutive
elements in our list.
If these elements never decrease — i.e. if when we perform these
comparisons, we see that l1 ≤ l2 ≤ . . . ≤ ln — then our list is already
sorted! Stop.
2. Otherwise, our list is not already sorted. In this case, randomly
shuffle the elements of our list around, and loop back to step 1.

Here’s a sample run of this algorithm on the list (1, 7, 1, 0), where I’ve
used random.org to shuffle our list when needed by step 3:

63

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

list iteration count step sorted list iteration count step sorted list iteration count step sorted
(1,7,1,0) 1 1 no (7,0,1,1) 6 1 no (1,7,0,1) 11 1 no
(1,1,0,7) 2 (1,1,0,7) 2 (1,7,1,0) 2
(1,1,0,7) 2 1 no (1,1,0,7) 7 1 no (1,7,1,0) 12 1 no
(7,0,1,1) 2 (1,1,7,0) 2 (1,7,1,0) 2
(7,0,1,1) 3 1 no (1,1,7,0) 8 1 no (1,7,1,0) 13 1 no
(1,0,1,7) 2 (0,1,7,1) 2 (0,1,7,1) 2
(1,0,1,7) 4 1 no (0,1,7,1) 9 1 no (0,1,7,1) 14 1 no
(7,1,1,0) 2 (7,1,0,1) 2 (0,1,1,7) 2
(7,1,1,0) 5 1 no (7,1,0,1) 10 1 no (0,1,1,7) 15 1 yes!
(7,0,1,1) 2 (1,7,0,1) 2
This . . . is not great. If you used BogoSort to sort a deck of cards, your
process would look like the following:
❼ One-by-one, go through your deck of cards and see if they’re or-
dered.
❼ If during this process you spot any cards that are out of order,
throw the whole deck in the air, collect the cards together, and
start again.
By studying the running time of this algorithm, we can make “not great”
into something rigorous:
Claim 4.3. The worst-case running time for BogoSort to sort any list
containing more than one element is ∞.

Proof. If L is a list containing n different elements, there are n! many

different ways to order L’s elements (this is ordered choice without rep-
etition, where we think of ordering our list as “choosing” elements one-
by-one to put in order.) If all of these elements are different, then there
is exactly one way for us to put these elements in order.
1
Therefore, on each iteration BogoSort has a n! chance of successfully
n!−1
sorting our list, and therefore has a n! chance of failing to sort our
list. For any n > 1, n!−1 is nonzero, and so the chance that our algorithm
fails on any given iteration is nonzero.
Therefore, in the worst-case scenario, it is possible for our algorithm to
just fail on each iteration, and thereby this algorithm could have infinite
run-time.

In terms of running time, then, we’ve shown that BogoSort has a strictly
2
3n2 + 9n2 − 10
worse runtime than InsertionSort, as ∞ > . Success!
2
This sort of comparison was particularly easy to perform, as one of the
two things we were comparing was ∞. However, this comparison process
can get trickier if we examine more interesting algorithms. Let’s consider
a third sorting algorithm:
Algorithm 4.7. The following algorithm, MergeSort(L), takes in a list
L = (l1 , l2 , . . . ln ) of n numbers and orders it from least to greatest. It
does this by using the following algorithm:
1. If L contains at most one number, L is trivially sorted! In this
situation, stop.
2. Otherwise, L contains at least two numbers. In this case,
(a) Split L in half into two lists L1 , L2 .
(b) Apply MergeSort to each of L1 , L2 to sort them.
3. Now, we “merge” these two sorted lists:
(a) Create a big list with n entries in it, all of which are initially
blank.
(b) Compare the first element in L1 to the first element in L2 .
If L1 , L2 are both sorted, these first elements are the smallest
elements in L1 , L2 .

64

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

(c) Therefore, the smaller of those two first elements is the small-
est element in our entire list. Take it, remove it from the list
it came from, and put it in the first blank location in our big
list.
(d) Repeat (b)+(c) until our big list is full!

As before, to better understand this algorithm, let’s run it on an example

list like (1, 7, 1, 0, 2):
original L step L1 L2 new list
(1,7,1,0,2) 1
2(a) (1, 7) (1, 0, 2)
2(b) (1, 7) (0, 1, 2)
3(a) ( , , , , )
3(b+c) (1, 7) (1, 2) (0, , , , )
3(b+c) (7) (1, 2) (0,1, , , )
3(b+c) (7) (2) (0,1,1, , )
3(b+c) (7) () (0,1,1,2, )
3(b+c),(d) () () (0,1,1,2,7)

original step L1 L2 new original step L1 L2 new

(1,7) 1 (1,0,2) 1
2(a) (1) (7) 2(a) (1) (0, 2)
2(b) (1) (7) 2(b) (1) (0, 2)
3(a) ( , ) 3(a) ( , , )
3(b+c) () (7) (1, ) 3(b+c) (1) (2) (0, , )
3(b+c),(d) () () (1,7) 3(b+c) () (2) (0,1, )
3(b+c),(d) () () (0,1,2)

original step L1 L2 new original step L1 L2 new original step L1 L2 new

(1) 1 (1) 1 (0,2) 1
2(a) (0) (2)
original step L1 L2 new
(7) 1 2(b) (0) (2)
3(a) ( , )
3(b+c) () (2) (0, )
3(b+c),(d) () () (0,2)

original step L1 L2 new original step L1 L2 new

(0) 1 (2) 1

Here, we use the colored boxes to help us visualize the recursive appli-
cations of MergeSort to smaller and smaller lists.
As before, it is worth taking a moment to explain why this algorithm
works:

Claim 4.4. MergeSort (i.e. algorithm 4.7) works.

Proof. We proceed in the same way as when we studied InsertionSort:

❼ Does the algorithm have any bugs? Nope! The only things we
do in our algorithm are compare elements, split our lists, and copy
elements over. Those are all well-defined things that can be done
without dividing by zero or other sorts of disallowed operations!
❼ Does the algorithm run forever? Nope!

65

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

To see why, let’s make a recurrence relation like we did before. To

be precise:
Claim 4.5. If we let MergeSortSteps(n) denote the maximum
number of steps needed by MergeSort to sort a list of n elements,
then MergeSortSteps(n) = 1 + 4n + 2MergeSortSteps(n/2) if n is
even.

Proof. By definition, MergeSort performs the following operations:

– In step 1, it performs one operation (it looks up the size of
the list.)
– In step 2, it splits the list in half. We can do this with at
most n write operations by just making two blank lists of size
n/2 and copying elements over one-by-one to these new lists.
It then runs MergeSort on each half. Doing this gives us an
If n is odd, this gets a bit extra 2 ⋅ MergeSortSteps(n/2) steps.
more annoying and we’d have
– In step 3, we repeatedly compare the first element in L1 to
MergeSortSteps(“n/2 rounded up”)
plus MergeSortSteps(“n/2 rounded the first element in L2 , remove the smaller of the two ele-
down”) here instead. This doesn’t ments from that list, and put it into our big list. This is one
materially change things, but it is comparison and two write operations, which we did as many
annoying enough to warrant just
looking at the even cases.
times as we had elements in our original list; so we have at
most 3 ⋅ n operations here in total.
In total, then, we have
MergeSortSteps(n) = 1 + n + 2 ⋅ MergeSortSteps(n/2) + 3 ⋅ n
= 1 + 4n + 2 ⋅ MergeSortSteps(n/2) .

If we just round any odd-length list up to an even-length list by

adding a blank cell, this gives us a recurrence relation that lets
us reduce the task of calculating MergeSortSteps(n) for any n
to the task of calculating smaller values of MergeSortSteps(n).
Therefore this process cannot run forever, by the same reasoning
as with InsertionSortSteps(n)!
❼ Does the algorithm sort our list? Yes!
To see why, make the following observations:
– Our algorithm trivially succeeds at sorting any list with one
element.
– Our algorithm also succeeds at sorting any list with 2 ele-
ments. To see why, note that by definition, it takes those two
elements and splits them into two one-element lists. From
there, it puts the smaller element from those two lists into
the first position, and puts the larger in the second position,
according to our rules. That’s a sorted list!
– Now, consider any list on 3 or 4 elements. By definition, our
algorithm will do the following:
✯ It takes that list, and splits it into two lists of size 1 or 2.
✯ It sorts those lists (and succeeds, by our statements above!)
✯ It then repeatedly takes the smaller of the first elements
in either of those two lists, removes it from that list, and
puts it into our larger list. Because those small lists are
sorted, each time we do this we’re removing the smallest
unsorted element (as the first element in a sorted list is its
smallest element.) Therefore, this process puts elements
into our largest list in order by size, and thus generates
a sorted list.

66

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

So our process works on lists of 3 or 4 elements!

– The same logic will tell us that our process works for lists on
5-8 elements: because any list on 5-8 elements will split into
two lists of size at most 4, and our process works on lists of
size at most 4, it will continue to succeed!
– In general, our process will always work! This is because our
process works by splitting n in half, and thus it reduces the
task of sorting a list of size n into the task of sorting two lists
of size ≈ n/2. Repeatedly performing this reduction eventually
reduces our task to just sorting a bunch of small lists, which
we’ve shown here that we can do.

As well, by using the same sort of “guess-a-pattern” idea from before,

we can refine this to a closed-form solution!
Note that if we use our definition, we can calculate the following values
for MergeSortSteps(2k ). By definition, MergeSortSteps(1) = 1, as this
algorithm stops when given any list of length 1. Therefore, by repeatedly
using Claim 4.5, we can get the following:

MergeSortSteps(21 ) = 1 + 4 ⋅ 2 + 2 ⋅ MergeSortSteps(2/2) = 1 + 8 + 2 ⋅ MergeSortSteps(1) = 11,

MergeSortSteps(22 ) = 1 + 4 ⋅ 4 + 2 ⋅ MergeSortSteps(4/2) = 1 + 16 + 2 ⋅ MergeSortSteps(2) = 39,
MergeSortSteps(23 ) = 1 + 4 ⋅ 8 + 2 ⋅ MergeSortSteps(8/2) = 1 + 32 + 2 ⋅ MergeSortSteps(4) = 111,
MergeSortSteps(24 ) = 1 + 4 ⋅ 16 + 2 ⋅ MergeSortSteps(16/2) = 1 + 64 + 2 ⋅ MergeSortSteps(8) = 287,
MergeSortSteps(25 ) = 1 + 4 ⋅ 32 + 2 ⋅ MergeSortSteps(32/2) = 1 + 128 + 2 ⋅ MergeSortSteps(16) = 703,
MergeSortSteps(26 ) = 1 + 4 ⋅ 64 + 2 ⋅ MergeSortSteps(64/2) = 1 + 256 + 2 ⋅ MergeSortSteps(32) = 1663,
MergeSortSteps(27 ) = 1 + 4 ⋅ 128 + 2 ⋅ MergeSortSteps(128/2) = 1 + 512 + 2 ⋅ MergeSortSteps(64) = 3839,

which in table form is the following:

k 1 2 3 4 5 6 7
MergeSortSteps(2k ) 11 39 111 287 703 1663 3839

(Note that we’re using 2k as the length of our lists. This is because our
result only works for even numbers, so we want something that stays
even when we keep dividing it by 2.)
Spotting the pattern here is a pain without more advanced mathematics;
even the Online Encyclopedia of Integer Sequences doesn’t recognize it.
WolframAlpha, however, does!

Again, this is a claim whose proof will have to wait for Section 7.8. For
now, though, let’s take the following as given:
Claim 4.6. MergeSortSteps(2k ) = k ⋅ 2k+2 − 2k+1 − 1, for every natural
number k.

67

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

We can easily extend this to a list whose length is not a power of two
by just “rounding its length up to the nearest power of 2” by adding in
some blank cells: i.e. if we had a list of length 28, we’d add in 4 blank
cells to get a list of length 32.
If we do this, then Claim 4.6 becomes the following:
Observation 4.10. MergeSortSteps(n) ≤ k ⋅ 2k+2 − 2k+1 − 1, where 2k
is n rounded up to the nearest power of 2.
Recall that ⌈x⌉ is “x rounded up.”
To simplify this a bit to just write things in terms of n, notice that if 2k
is n rounded up to the nearest power of 2, then k = ⌈log2 (n)⌉. Plug this
into Observation 4.10, and you’ll get the following:

Two useful tricks we’re using in this MergeSortSteps(n) ≤ ⌈log2 (n)⌉ ⋅ 2⌈log2 (n)⌉+2 − 2⌈log2 (n)⌉+1 − 1
≤ (log2 (n) + 1)2(log2 (n)+1)+2 − 2(log2 (n)+1)+1 − 1
calculation:
❼ x ≤ ⌈x⌉. That is: rounding a number
= (log2 (n) + 1)2log2 (n) ⋅ 23 − 2log2 (n) ⋅ 22 − 1
up only makes it larger.
❼ ⌈x⌉ < x + 1. That is: rounding up
never increases a number by more = 8n log2 (n) + 8n − 4n − 1.
than 1.
= 8n log2 (n) + 4n − 1 .

Nice! This is worth making into its own observation:

Observation 4.11. MergeSortSteps(n) ≤ 8n log2 (n) + 4n − 1, for every
positive integer n > 1.

4.5 Comparing Runtimes: Limits

Now, we have an interesting problem on our hands: between MergeSort
and InsertionSort, which of these two algorithms is the most efficient?
That is: in terms of functions, which of the following is better?

3n2 + 9n − 10
8n log2 (n) + 4n − 1 vs.
2

To start, we can make a table to compare values:

n 1 2 3 4 5 6 7
3n2 + 9n − 10
1 10 22 37 55 76 100
2
8n log2 (n) + 8 log2 (n) − 2n − 1 3 23 49.0 79 111.9 147.1 184.2

It looks like the MergeSort algorithm needs more steps than InsertionSort
so far. However, this table only calculated the first few values of n; in
real life, however, we often find ourselves sorting huge lists! So: in the
long run, which should we prefer?
If you would like a more rigorous defi- To answer this, we need the idea of a limit:
nition here than “gets close,” look into
classes like Maths 130 and Maths 250! Definition 4.4. Given any function f that is defined on the natural
numbers N, we say that lim f (n) = L if “ as n goes to infinity, f (n)
gets closer and closer to L.” In particular, we say that lim f (n) = +∞
n→∞

if “as n goes to infinity, f (n) also grows without bound.”

n→∞

This is a tricky concept! To understand it, let’s look at a handful of

examples:

68

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

1
Problem 4.1. What is lim ?
n→∞ 2 − n1

Answer. First, notice that as n goes to infinity, n1 goes to 0, because

its denominator is going to infinity while the numerator stays fixed.
1 1
Therefore, 2 − n1 goes to 2, and so lim = .
n→∞ 2 − 1 2
n

Problem 4.2. What is lim log2 (n)?

n→∞

Answer. lim log2 (n) = ∞. This is because for any positive integer k,
n→∞
we can make log2 (n) ≥ k by setting n to be any value ≥ 2k .
In other words, as n grows, it eventually gets larger than 2k for any fixed
k, and thus log2 (n) itself grows without bound.
1
Problem 4.3. What is lim ?
n→∞ log2 ( n1 )

Answer. To find this limit, we simply break our function down into
smaller pieces:
❼ First, notice that as n takes on increasingly large positive values,
1
n
goes to 0 and stays positive.
❼ Therefore, log2 ( n1 ) goes to negative infinity, as log2 (tiny positive
numbers) yields increasingly huge negative numbers.
❼ Therefore, log 1( 1 ) goes to 0, as 1
huge negative numbers
yields tiny neg-
2 n
ative numbers.
1
In total, then, lim = 0.
n→∞ log ( 1 )
2 n

n2 + 3n − 4
Problem 4.4. What is lim ?
n→∞ n3 + 2
Answer. It is tempting to just “plug in infinity” into the fraction above,
and say that
∞2 +3∞−4 ∞
“because ∞3 +2
= ∞
= 1, our limit is 1”

However, you can’t do manipulations like this with infinity! For example,
because n1 = nn2 , we have
n 1
lim = lim = 0,
n→∞ n2 n→∞ n
even though the method above would say that
∞ ∞
“because ∞2
= ∞
= 1, our limit is 1.”

The issue here is that there are different growth rates at which various
expressions approach infinity: i.e. in our example above, n2 approaches
infinity considerably faster than n, and so the ratio nn2 approaches 0 even
though the numerator and denominator individually approach infinity.
Instead, if we ever have both the numerator and denominator approach-
ing +∞, we need to first simplify our fraction to proceed further! In
this problem, notice that if we divide both the numerator and the de-
nominator by n3 , the highest power present in either the numerator or
denominator, we get the following:
1 3 4
n2 + 3n − 4 1/n3 n
+ n2
− n3
⋅ = 2
.
n3 + 2 1/n3 1+ n3

69

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

As noted above, each of n1 , n32 , − n43 , n23 go to 0 as n goes to infinity,

because their denominators are going off to infinity while the numerator
is staying fixed.
Therefore, we have
1 3 4
n2 + 3n − 4 n
+ n2
− n3 0+0+0
lim = lim 2
= lim = 0.
n→∞ n3 + 2 n→∞ 1+ n3
n→∞ 1 + 0

With the idea of limits in mind, we can now talk about how to compare
functions:
Definition 4.5. Let f (n) and g(n) be two functions which depend on
n. We say that the function f (n) grows faster than the function g(n)
∣f (n)∣
n→∞ ∣g(n)∣
if lim = +∞.

Intuitively, this definition is saying that for huge values of n, the ratio of
f (n) to g(n) goes to infinity: that is, f (n) is eventually as many times
larger than g(n) as we could want.
We work an example of this idea here, to see how it works in practice:
Example 4.14. We claim that the function f (n) = n2 + 2n + 1 grows
faster than g(n) = 5n + 5. To see this, by our definition above, we want
n2 + 2n + 1
to look at the limit lim .
n→∞ 5n + 5
Notice that we can factor the numerator into (n + 1)2 . Plugging this
(n + 1)2
into our fraction leaves us with , which we can simplify (as in
5(n + 1)
n+1
Problem 4.4,) to .
5
n+1
As n goes to infinity, goes to infinity.
5
1 1
This grows without bound: as n goes to infinity, so does n + ! There-
5 5
fore we’ve shown that n2 + 2n + 1 grows faster than 5n + 5.

To deal with a comparison like the one we’re trying to do between

MergeSort and InsertionSort, however, we need some more tricks!

4.6 Limit Techniques and Heuristics

Without calculus, our techniques for limits are a little hamstrung. As
those of you who have seen NCEA L3 calculus, things like L’Hôpital’s
rule are extremely useful for quickly evaluating limits!
With that said, though, we do have a handful of useful techniques and
heuristics that we can use to get by. Here’s an easy (if not very rigorous)
technique:
Observation 4.12. Plugging In Values. Probably the simplest thing
you can do, when given a limit, is just physically plug in numbers and
figure out where the function is going.

For instance, suppose that we wanted to compare the runtime of our

two functions MergeSortSteps and InsertionSortSteps. By defini-
tion, this means that we want to find the limit
3n2 +9n−10
2
8n log2 (n) + 4n − 1
lim .
n→∞

70

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

To do this, we could just plug in various values of n that are close to but
greater than zero, and see what happens!

x f(x)
1 0.33
10 0.63
100 2.70
1000 17.97
10000 136.03

It certainly looks like f (x) is growing arbitrarily large, so we could

quite reasonably believe that the number of steps required to calculate
InsertionSort is growing faster than the number of steps needed to
calculate MergeSort (and thus MergeSort is the preferable algorithm.)
However, this method has its limitations:
❼ One issue with the above is that it is fairly prone to human error if
you’re manually doing this by plugging numbers into a calculator.
x3 − 3x2 + 3x
That is: if you have a limit like lim , and you’re
x→∞ x2 − x
trying to plug in something like x = 1000000 into that fraction, it’s
going to be really easy to forget a zero in one of those x expressions.
❼ Another is that it can be pretty hard to tell whether or not you’re
actually plugging in enough values to figure out the pattern! For
example, when we made our table to compare the runtimes of these
two functions by listing values from 1 through 7, we thought that
MergeSortSteps was growing faster. This larger table seems to
be telling the opposite story: but maybe the situation will reverse
itself again if we zoom out further!
To give a second example, suppose that someone asked you to
calculate
lim log2 (log2 (log2 (log2 (log2 (n)))))
n→∞

If you plugged in n = 100, 1000, 10000,, you’d get ≈ −0.9, −0.34, −0.11.
This looks like it’s slowly increasing to 0, so you’d be tempted to
guess that
lim log2 (log2 (log2 (log2 (log2 (n)))) = 0.
n→∞

This is very false! As we saw before, as n goes to infinity, log2 (n)

goes to infinity. Therefore, log2 (log2 (n)) also goes to infinity, and
in general any composition of logs will eventually go off to infinity
as well: i.e. +∞ is the correct limit here. So plugging things in
can lead us to make mistakes!
A second technique, that we used in several of our worked problems
earlier, is the following:
Observation 4.13. Simplifying Fractions. In the special case where
f (x)
your limit has the form lim , a useful technique you can try is
x→∞ g(x)
f (x)
simpliflication! Basically, take your fraction , and try to simplify
g(x)
it by factoring the top and bottom and canceling terms.

This often gives you a new function where you no longer have the top
and bottom going to zero, which is often much easier to work with.

71

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

x2 − x
Example 4.15. If we had the limit lim , we could factor an x
x→+∞ x
out of the top and bottom to get lim x − 1. This is +∞.
x→∞
x2 − x
For a second example, if we had the limit lim , we could factor
x→∞ x2
1 − x1
an x2 out of the top and bottom to get the simplified limit lim .
x→+∞ 1
The numerator goes to 1 and the denominator just is 1, so this is 1.

Sometimes, however, we don’t have something that we can easily express

as a fraction:
Problem 4.5. What is the limit
1
lim
x→+∞ 1 − 21/x

To approach this limit, we need another technique:

Observation 4.14. Break It Down. Given a limit lim f (x), we can
x→+∞
often figure out what’s going on with it by breaking our functions down
into small pieces, looking at what those individual pieces do as x goes to
+∞, and then slowly “zooming back out” to see what the whole function
does.

To understand this method, let’s apply it to Problem 4.5.

Answer to Problem 4.5. Understanding the function in our limit all
at once is hard! But, notice that for very large values of x, we know that
❼ 1/x is a very small positive number, therefore
❼ 21/x is basically 2basically 0, positive , which is slightly larger than 1,
therefore
❼ 1 − 21/x is basically 1 − (slightly larger than 1), and so in turn is a
tiny negative number, therefore
1
❼ is 1 over a tiny negative number, and thus is an increas-
1 − 21/x
ingly huge negative number.
1
Therefore, as x goes to positive infinity goes to −∞, and we’ve
1 − 21/x
found our limit by breaking our function down into smaller pieces!

Observation 4.15. Heuristics. Our last limit idea is the following:

with some calculus (take Maths 130!), you can prove that
Constants << Logarithms << Polynomials << Exponentials << Factorials,

where “<<” means “grows slower than.”

Within those groups, we sort these expressions by degrees and bases: i.e.

n << n2 << n3 << n4 << . . .,

and

2n << 3n << 4n << 5n << . . .,

and so on / so forth.
Finally, any sum of expressions grows as fast as its largest expression:
i.e. a factorial plus a log plus a constant grows at a factorial rate, a n4
plus a log plus a square root grows as fast as n4 grows, etc.
As a brief justification for this, let’s look at a table:

72

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Runtime vs. Input 10 20 30 40 50

1 1 ⋅ 10−6 sec. 1 ⋅ 10−6 sec. 1 ⋅ 10−6 sec. 1 ⋅ 10−6 sec. 1 ⋅ 10−6 sec.
log2 (n) 3.3 ⋅ 10−6 sec. 4.3 ⋅ 10−6 sec. 4.9 ⋅ 10−6 sec. 5.3 ⋅ 10−6 sec. 5.6 ⋅ 10−6 sec.
n 1 ⋅ 10−5 sec. 2 ⋅ 10−5 sec. 3 ⋅ 10−5 sec. 4 ⋅ 10−5 sec. 5 ⋅ 10−5 sec.
n2 1 ⋅ 10−4 sec. 4 ⋅ 10−4 sec. 9 ⋅ 10−4 sec. 1.6 ⋅ 10−3 sec. 2.5 ⋅ 10−3 sec.
n3 1 ⋅ 10−3 sec. 8 ⋅ 10−3 sec. .027 sec. .064 sec. .125 sec.
n5 .1 sec. 3.2 sec. 24.3 sec. 1.7 min. 5.2 min.
2n 1 ⋅ 10−3 sec. 1 sec. 17.9 min. 12.7 days. 35.7 years
n! 3.63 sec. 77146 years 8.4 ⋅ 1018 years 2.6 ⋅ 1034 years 9.6 ⋅ 1050 years

Above, we’ve plotted five algorithms with runtimes 1, log2 (n), n, n2 , n3 , n5 , 2n , n!

versus input sizes for n ranging from 10 to 50, with the assumption that
we can perform one step every 10−6 seconds.
In this table, you can make the following observations:
❼ The constant-runtime algorithm is great!
❼ The logarithmic-runtime algorithm is also pretty great!
❼ The polynomial-runtime algorithms all take longer than the con-
stant or log algorithm, but are all at least reasonable.
❼ The exponential runtime algorithm starts off OK, but gets horrible
fast. . .
❼ . . . but is somehow not as bad as the factorial-runtime algorithm,
which is almost immediately unusable for any value of n.
We can use this heuristic to quickly answer our question about which of
MergeSortSteps and InsertionSortSteps is preferable, without hav-
ing to plug in any numbers at all!
Claim 4.7. InsertionSortSteps grows faster than MergeSortSteps.

Proof. As noted before, by definition, we want to find the limit

3n2 +9n−10
2
8n log2 (n) + 4n − 1
lim .
n→∞

By dividing through both sides by n, this means that we’re looking at

the ratio
5
1.5n + 4.5 − n
8 log2 (n) + 4 −
lim 1
.
n→∞
n

The top is a linear expression (i.e. polynomial of degree 1), as the

largest-growing object in the numerator is the 1.5n term. The bottom,
conversely, is a logarithmic expression, as the fastest-growing object in
the denominator is the 8 log2 (n) term.
Linear expressions grow much faster than logs, so our “plug things in”
step didn’t lie: InsertionSortSteps is growing faster than MergeSortSteps
(and thus MergeSort is the preferable algorithm, as in the long run it
will need to perform less operations to get to the same answer.)

To study one more example of this idea and finish our chapter, let’s use
this principle to answer our last exercise:
Answer to 4.1. In this problem, we had two multiplication algorithms
and wanted to determine which is best.
³¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ·¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ µ
a times

Algorithm 4.1 calculated a ⋅ b by basically just calculating b + b + . . . + b.

As such, its runtime is pretty much just some constant times a; we need
a iterations of this process to calculate the answer here, and we can let
our constant be the number of steps we perform in each iteration.
Conversely, Algorithm 4.2 ran by taking a and repeatedly subtracting 1
from a and dividing a by 2 until it was 0 (while doing some other stuff.)

73

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

For the purpose of this problem, we don’t have to think about why
this process works, though look at practice problem 6 if you’re curious!
Instead, if we simply take it on faith that this algorithm does work, it’s
easy to calculate how long it takes to complete: it will need as many
iterations as it takes to reduce a to 0 by these operations.
In the worst-case scenario for the number of iterations, we only ever
divide by 2; i.e. a is a power of 2. In this case, it takes k iterations if
a = 2k ; i.e. we need log2 (a) iterations, and thus need a constant times
log2 (a) many operations.
As we saw before, logarithms grow much slower than linear functions!
Therefore, Algorithm 4.2 is likely the better algorithm to work with.

4.7 Practice Problems

1. (-) Your two friends, Jiawei and Francis, have written a pair of
algorithms to solve a problem. On an input of size n, Jiawei’s
algorithm needs to perform n2 + 2n calculations to find the answer,
while Francis’s algorithm needs to perform log2 (n)+2n calculations
to find the same answer.
Jiawei says that their algorithm is faster, because their runtime
is polynomial while Francis’s algorithm has an exponential in it.
Francis says that their algorithm is faster, because their algorithm
has a logarithm in it while Jiawei’s is polynomial.
Who is right, and why?

√
2. (-) Let f (x) = 10x + 1, g(x) = x − 3, h(x) = log10 (x3 + 3) and j(x) =
3
x. Calculate the following compositions:

❼ f ○ h(x). ❼ g ○ f (x).
❼ h ○ j(x). ❼ f ○ g(x).

3. We say that a function f ∶ A → B has an inverse if there is some

function f −1 ∶ B → A such that f ○ f −1 (x) = x = f −1 ○ f for all x:
vice-versa. For example, log2 (x) and
that is, f −1 “undoes” f , and √
2x are inverses, as is x5 and 5 x.
❼ Suppose that f (x) = 3x + 2. Find f (x)’s inverse.
❼ Now, suppose that g(x) = 3. Does this constant function have
an inverse?
4. Let A be the collection of all students enrolled at the University of
Auckland and B be the collection of classes currently running at
the University of Auckland. Which of the following are functions?
❼ The rule f ∶ A → B, that given any student outputs the classes
they’re enrolled in.
❼ The rule f ∶ B → A, that given any class outputs the student
with the top mark in that class.
❼ The rule f ∶ A → A, that given any student outputs that
student.
❼ The rule f ∶ B → B, that given any class outputs its prereq-
uisites.
5. You’re a programmer! You’ve found yourself dealing with a pro-
gram puzzle(n) that has no comments in its code, and you want to
know what it does. After some experimentation, you’ve found that
puzzle(n) takes in as input an integer n, and does the following:

74

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

(i) Take n and square it.

(ii) If n is 1, 2, 3 or 6, output n and stop.
(iii) Otherwise, replace n with (n2 ) % 10, i.e. the last digit of n2 ,
and go to (ii).

❼ Is puzzle(n) a function, if we think of its domain and codomain

as Z?
❼ What is the range of puzzle(n)?
6. (+) Show that after each step in Algorithm 4.2, the quantity
a ⋅ b + prod is always the same. Using this, prove that Algorithm
4.2 works!
7. (+) In our answer to 4.1, we used a sort of handwave-y “this takes
about log2 (a) many operations” argument. Let’s make this more
rigorous!
That is: let clevermultsteps(a) denote the number of steps that
this algorithm needs to multiply a given number b by a.
(a) Find a recurrence relation clevermultsteps(a) in terms of
clevermultsteps(a/2), that holds whenever a is even.
(b) Use this relation to calculate clevermultsteps(a) for a =
2, 4, 8, 16, 32, and plug these values into either the OEIS or
WolframAlpha. What pattern do you see?
8. Consider the following sorting algorithm, called BubbleSort:
Algorithm 4.8. The following algorithm, BubbleSort(L), takes
in a list L = (l1 , l2 , . . . ln ) of n numbers and orders it from least to
greatest. It does this by using the following algorithm:
(i) Compare l1 to l2 . If l1 > l2 , swap these two values; otherwise,
leave them where they are.
(ii) Now, move on, and compare the “next” pair of adjacent values
l2 , l3 . Again, if these elements are out of order (i.e. l2 > l3 )
swap them.
(iii) Keep doing this through the entire list!
(iv) At the end of this process, if you made no swaps, stop: your
list is in order, by definition. Stop!
(v) Otherwise, the list might be out of order! Return to (i).
(a) Use this process to sort the list (6, 1, 4, 2, 3, 2).
(b) (+) Let BubbleSortSteps(n) denote the maximum number
of steps that BubbleSort needs to sort a list of length n.
Find an expression for BubbleSortSteps(n), and calculate
its values for n = 1, 5, 10 and 100.
(c) What is the smallest number of steps that BubbleSort would
need to sort a list of length n?
9. Find the following limits:

2n + 2n+1 sin(2n + n) + n
n→∞ n − log (n)
(a) lim (d) lim
n→∞ 2n − 2n−1 2

(e) lim log2 ( )

n3 + 3n2 + 3n + 1 1
(b) lim
n→∞ n3 − 3n2 + 3n − 1 n→∞ log2 (n)
nn
(c) lim 2log2 (n)−6 (f) lim
n→∞ n→∞ n!

10. Which of the three algorithms BubbleSort, MergeSort and InsertionSort

is the fastest (i.e. needs the fewest operations) to sort a list of 5
elements? Which needs the fewest to sort a list of 10 elements?
How about 100? How about for huge values of n?

75

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Graphs
2019 Chapter 5

Exercise 5.1. In the 1700’s, the city of Königsberg was divided by the
river Pregel into four parts: a northern region, a southern region, and
two islands. These regions were connected by seven bridges, drawn in
red in the map at right.
These bridges were particularly beautiful pieces of architecture, and so
when people would offer tours of the city they would try to take tourists
across each bridge to show it off. However, tour guides at the time
noticed that no matter how they could construct their walks around the
city, they’d always either miss a bridge or have to double a few bridges
up: they could never find a “perfect” tour that crossed each bridge exactly
once.
A map of Königsberg, c. 1730. Can you find such a tour? That is: can you come up with a walk through
Source: Wikipedia. the city that starts and ends at the same place, and walks over each bridge
exactly once?
Exercise 5.2. You’re a civil engineer! Suppose that you have a set of
buildings, and you want to hook up each of these buildings to the city’s
utilities, namely water, gas, and electricity.
However, your city is very earthquake-prone. As a safety precaution,
they’ve asked that no two utilities are allowed to vertically “cross over”
each other: that is, you can’t have water mains running beneath the gas
mains, or gas pipes running above the electricity wires. So, for example,
while the configuration drawn at left wouldn’t work for two buildings, the
right one would!
Can you connect three buildings to your utilities? Or is this impossible?

5.1 Graph Theory

The field of graph theory originally started as a branch of “recreational
mathematics,” and specifically as the puzzle in Exercise 5.1. Until the
1900’s, people viewed graph theory as a branch of mathematics used
to solve riddles similar to the Königsburg bridge problem, such as the
following:
❼ Can you take a dodecahedron, start from one corner, and walk
A solution to the dodecahedron puzzle!
Source: Wikipedia
along the edges in a way that visits every other vertex exactly
once and returns to where you start?
❼ Take a knight, and put it in the top-right-hand corner of an 8 × 8
chessboard. Can you move it around so that it visits every other
square on the board exactly once, and returns to where it started?
❼ Take a box, divided into five rectangles. Can you draw a single
continuous line that crosses each wall exactly once?
In recent years, however, mathematicians and computer scientists have
transformed graph theory into one of the most applicable fields of math-
ematics in existence. Its power to describe networks has made it the
perfect tool for studying many problems in the modern world; graphs
can be used to model the internet, social networks, the spread of dis-
eases through a population, travel, computer chip design, and countless
other phenomena. Graphs are everywhere in the modern world, and

76

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

their analysis is key to solving many of the most important problems of

the 21st century.
In these notes, we’ll start by building up some definitions that will let
us talk about graphs and their properties; from there, we’ll describe how
to model various real-life objects with graphs, and then transition to
solving a few real-life problems with graphs. This is (as with everything
in this class) just the tip of the iceberg; if you want to see more, take
papers like Compsci 225 and Maths 326!
We start with an intuitive definition of a graph:
Definition 5.1. Intuitively, a graph is just a way of modeling a collec-
tion of objects and the connections between them.
Example 5.1.
For example, all of the following objects can be thought of as graphs:
❼ Consider Facebook, or any other social network. You can think of
any of these objects as consisting of two things: (1) the people,
and (2) the connections (i.e. friendship, or following) between
people.
❼ Think about the internet. You can model this as a collection of
webpages, connected by links.
❼ Alternately, you can model the internet as a collection of comput-
ers, connected by cables/wireless signals.
❼ Look at all of New Zealand! You can model the country as a
collection of cities, connected by roads.
❼ Alternately, you can choose to model New Zealand as a collection
of regions, with two regions linked up if there’s a direct flight from
one to the other.

If Definition 5.1 above feels a bit too informal for you, here’s a more
precise way to describe a graph: For simplicty’s sake, we will use the
word “graph” to refer to a simple undi-
Definition 5.2. A simple undirected loopless graph G consists of rected loopless graph, and assume that
two things: a collection V of vertices, and another collection E of all graphs are simple undirected loop-
edges, which we think of as distinct unordered pairs of distinct elements less graphs unless we explicitly say oth-
erwise.
in V . We think of the vertices in a graph as the objects we’re study-
ing, and the edges in a graph as a way to describe how those objects are
connected.
To describe an edge connecting a pair of vertices a, b in our graph G, we
use our set language from earlier and write this as {a, b}. We say that
a and b are the endpoints of the edge {a, b} when this happens.
Example 5.2. The following describes a graph G:
❼ V = {a, b, c, d, e}
❼ E = {{a,b}, {b,c}, {c,d}, {d,e}, {e,a}}
a a
Given a graph G = (V, E), we can visualize G by drawing its vertices as e b d c
points on a piece of paper, and its edges as connections between those
points. d c b e
Several ways of drawing the graph G defined above are drawn at right.
Notice that there are many ways to draw the same graph! Also note that
edges don’t have to be straight lines, and that we allow them to cross. a b c d e
In general, it is often much easier to describe a graph by drawing it
rather than listing its vertices and edges one-by-one. Wherever we can,
we’ll describe graphs by drawing pictures; we encourage you to do the
same!
Here’s a few examples of graphs, described with this vertex+edge lan-
guage:

77

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Example 5.3.
❼ We can represent a maze as a graph! Take the collection of “rooms”
in our maze, and think of them as our vertices. Now, connect any
two room-vertices with an edge if they are linked by a doorway.
With this done, finding a way out of our maze is the same as
finding a walk through our graph! This can often simplify things
for us, as the graph visualization lets us ignore some of the more
irrelevant information (i.e. right angles, walls without doors, etc)
in the maze.
❼ You and your three flatmates (Aang, Korra, and Zuko) have just
moved to a new house. You have a list of tasks to do over the
weekend: you all want to paint the walls, move in your furniture,
cook dinner, and do some gardening. Suppose that you have the
following skills:
– You really like to paint and garden.
– Aang is a great cook who enjoys gardening.
– Zuko is also a great cook who’s strong (furniture.)
– Korra is strong (furniture) and likes to paint.
You Paint You can visualize this with a graph! Make a vertex for each flat-
mate, and a vertex for each task. Then, connect flatmates to their
Aang Garden preferred tasks with edges. With this, a “good” way to divide
up tasks is to find a “matching:” that is, a way to pick out four
Zuko Furniture
edges in our graph, such that every person and every task is in
exactly one edge (so that the work is divided!) Such a matching is
Korra Cook
highlighted in the graph at right.

Just as there are many different ways to model the connections between a
set of objects, there are other notions of graphs beyond that of a simple
graph. Here are some such definitions:
Definition 5.3. A simple graph with loops is just like a simple graph
G, except we no longer require the pairs of elements in E to be distinct;
that is, we can have things like {v, v} ∈ E.
A multigraph is a simple graph, except we allow ourselves to repeat
A simple graph with loops A multigraph
on five vertices. edges in E multiple times; i.e. we could have three distinct edges e1 , e2 , e3 ∈
E with each equal to the same pair {x, y}.
on four vertices.

A directed graph is a simple graph, except we think of our edges as

ordered pairs: i.e. we have things like x → y ∈ E, not {x, y}.
A directed graph on seven vertices.

You can mix-and-match these definitions: you can have a directed graph
with loops, or a multigraph with loops but not directions, or pretty much
anything else you’d want!
Remark 5.1. In this course, we’ll assume that the word graph means
simple undirected graph without loops unless explicitly stated oth-
erwise.

There are a handful of graphs that come up frequently in computer sci-

ence and mathematics, are are worth giving specific names. We describe
a few of these graphs here:
Definition 5.4. The complete graph on n vertices Kn is defined for
any positive integer n as follows: take n vertices. Now, take every pos-
sible pair of distinct vertices, and connect them with an edge! We draw
K3 K4 K5 K6 several examples at right.
In this sense, a complete graph is a graph in which we have as many
edges as is possible for a graph on n vertices.

78

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Definition 5.5. The complete bipartite graph on m, n vertices Km,n

is defined for any two positive integers m, n as follows: take m + n ver-
tices. Split them into two groups, one of size m and one of size n. Then,
for every pair of vertices {a, b} where a is from the m group and b is from
the n group, connect a to b. These graphs have no edges connecting any
two members from the “same” group, but have all of the possible “cross-
ing” edges going from one group to the other.

Definition 5.6. The cycle graph on n vertices Cn is defined for any

integer n ≥ 3 as follows: take n vertices, and label them 1, 2, . . . n. Now,
draw edges {1, 2},{2, 3}, . . . until you get to the last edge {n − 1, n}; then
connect this up into a closed cycle by drawing {n, 1} as an edge as well. C3 C4 C5 C6

To practice all of this graph theory language, let’s try solving one of our
exercises!
Answer to Exercise 5.2. You cannot solve the utility problem!
To see why, let’s use the language of graph theory. Make each of our
three buildings into vertices; as well, make each of the three utilities into
vertices. If we connect each building to each utility, notice that we’ve
created the graph K3,3 !
b
We claim that K3,3 cannot be drawn without any edges crossing, and a b c
thus that this puzzle cannot be solved. To see why this is true, let’s use d f
the “contradiction” technique we used above.
Think about what a solution would look like. In particular, notice that a c
K3,3 contains a “hexagon,” i.e. a C6 , as drawn at right. Therefore, in d e f
e
any drawing of K3,3 , we will have to draw this cycle C6 first.
Because this cycle is a closed loop, it separates space into an “inside” and
an “outside.” Therefore, if we are drawing K3,3 without edges crossing,
after drawing the C6 part, any remaining edge will have to either be
drawn entirely “inside” the C6 or “outside” the C6 . That is, we can’t
have an edge cross from inside to outside or vice-versa, because that b b
would involve us crossing over pre-existing edges. d f d f
Therefore, if we have a crossing-free drawing of K3,3 , after we draw the or
C6 part of this graph, when we go to draw the {d, c} edge we have two a c a c
options: either draw this edge entirely on the inside of our C6 , or entirely
on the outside. e e
If we draw this edge on the inside, then on the inside the vertices f and
a are separated by this edge; therefore, to draw the edge {a, f } we must
go around the outside. Similarly, if we draw this edge on the outside,
then a, f are separated from each other on any outside walk, and the
edge {a, f } must be drawn inside of the hexagon. b b
In either of these cases, notice that there is no walk that can be drawn d f d f
from b to e on either the inside or the outside! Therefore we cannot or
draw our last edge {b, e} without having a crossing, and thus have a a c a c
contradiction to our claim that such a crossing-free drawing of K3,3 was
possible. e e

5.2 Graphs: Useful Concepts

In the above section, we saw how some of the language of graph theory
could help us solve problems. In this section, we explore a number of
other definitions and concepts that can come in handy!
Definition 5.7. Take a graph G. We say that two vertices a, b in G are
adjacent if the edge {a, b} is in G. We say that a and b are neighbors
if they are adjacent.

79

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Example 5.4. If we look at the flatmate graph from Example 5.3 above,
we can see that the vertices “You” and “Paint” are adjacent, but that
“You” and “Furniture” are not adjacent. Similarly, “Aang” and “Cook”
are adjacent, but “Aang” and “Zuko” are not adjacent.
Notice that adjacency is just a property of individual edges! Even though
we drew the Aang and Zuko vertices next to each other on our graph,
we did not connect them with an edge, so they are not connected. As
well, even though you can go from “You” to “Furniture” by using the
multiple-edge walk

You → Paint → Korra → Furniture,

you cannot go from “You” to “Furniture” with a single edge, and so

these vertices are not adjacent.
It is often useful to be able to refer to all of the neighbors of a ver-
tex: in this graph, for instance, “Korra” is neighbors with “Paint” and
“Furniture.”
Definition 5.8. Take a graph G, and a vertex v in G. We say that the
degree of v, written deg(v), is the number of edges that contain v as an
endpoint.
Example 5.5. In the graph with red vertices drawn in the margins, the
vertex a has degree 0, the vertex b has degree 1, the vertices d and e
have degree 2, and the vertex c has degree 3.
a b
Notice that a vertex can have degree 0: i.e. it is possible to have a
c vertex with no edges! This will often represent an “isolated” object in
d e our graph: i.e. something without any connections to our wider graph.
Exercise 5.3. A bit more practice with the concept of degree: convince
yourself of the following claims!
❼ The degree of every vertex in Kn is n − 1.
❼ The only possible degrees in Km,n are m and n.
❼ The degree of every vertex in a cycle graph Cn is 2.

To practice these concepts, let’s try writing a few arguments:

Claim 5.1. (The “degree-sum formula,” or “handshaking theorem.”)
Take any graph G. Then, the sum of the degrees of all of the vertices in
G is always two times the number of edges in G.

Proof. We can prove this by “counting” the number of times vertices in

our graph show up as endpoints of edges in our graph. Notice that we
could do this in two different ways:
1. On one hand, every edge has two endpoints. Therefore, the total
number of times vertices are used as endpoints in our graph is
simply twice the number of edges.
2. On the other hand, every endpoint is counted once when we calcu-
late the degree of the corresponding vertex. Therefore, if we sum
up the degrees of all of the vertices in our graph, this counts the
total number of times vertices are used as endpoints in our graph.
However, both of these ways were counting the same thing! Therefore,
we know that the answer we got from (1) must be the same as the answer
we got in (2). In other words, the sum of degrees of vertices in our graph
must be twice the number of edges, as claimed.
Example 5.6. In the graph at right, we have four vertices of degree 2,
five vertices of degree 3, and one vertex of degree 5. Therefore, the sum
of degrees in this graph is 4 ⋅ 2 + 5 ⋅ 3 + 1 ⋅ 5 = 28.

80

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Claim 5.1 says that this should be twice the number of edges; i.e. that
this graph should have 14 edges. This is true: count them by hand!

This formula comes up often in graph theory, especially when trying

to show that certain types of graphs are “impossible!” Consider the
following result:

Claim 5.2. You cannot have a graph on seven vertices in which the
degree of every vertex is 3.

Proof. Before reading this proof, try to show this yourself: that is, take
pen and paper, and try to draw a graph on seven vertices where all of
the degrees are 3!
In doing so, we make two claims:
❼ You won’t succeed: you’ll always wind up with one vertex forced
to have degree 2 or 4 (or you’ll accidentally make something that’s
not a simple graph by drawing multiple edges / loops.)
❼ It won’t be obvious why this keeps failing: i.e. there will be a lot of
different things you could have tried, and writing down a “proof”
for why this is impossible may feel like it would involve a ton of
tedious casework.
However, if we use the degree-sum formula, we claim that this problem
is quite simple! We start by using a principle that’s served us well in
many prior problems: let’s suppose that our claim is somehow false, and
it is possible to have a graph G on seven vertices in which all degrees
are 3.
The degree-sum formula, when applied to G, tells us that the sum of
the degrees in G must be twice the number of edges. Because all of the
degrees in G are 3 and there are seven vertices, we know that this degree
sum is just 3 ⋅ 7 = 21.
Therefore, we must have that two times the number of edges in our graph
is 21. That is, we have 10.5 edges . . . which is impossible, because edges
come in whole-number quantities! (I.e. you either have an edge {a, b}
or you don’t: there is no “half” of an edge {a in a simple graph.)
Therefore, such a graph G cannot exist, as the number of edges required
for such a graph G is impossible to construct. In other words, no graphs
exist on seven vertices in which all degrees are 3!

5.3 Walks and connectedness

In several of the examples we looked at earlier, the idea of a walk through
a graph was an incredibly useful one:
❼ In the “internet” graph, if we model this graph as a collection of
computers connected by cables/wireless signals, in order for two
people to communicate (i.e. send an email,) we have to find a way
through this graph that links these two computers together.
❼ In the Königsberg bridge problem from Example 5.1, we can turn
the city of Königsberg into a multigraph by making the four regions
of the city into vertices, and connecting these vertices with edges.
In this setting, our goal is to find a walk through this graph that
uses every edge exactly once, and starts + ends at the same place.
❼ In the “maze” graph, we wanted to find a walk through our graph
that starts at the entrance and leaves through the exit.

81

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ In the New Zealand graph, where we had a bunch of cities as our

vertices connected by roads, we often want to navigate from one
city to the next. Doing so is finding a route through our graph!
As such, we should come up with a definition for what a “walk” is in a
graph:

Definition 5.9. In a graph G, we define a walk of length n as any

sequence of n edges from G of the form

{v0 , v1 }, {v1 , v2 }, {v2 , v3 }, . . . , {vn−1 , vn }.

We say that this walk starts at v0 and ends at vn .

We say that a walk is a circuit or closed walk if it starts where it ends;
i.e. if v0 = vn .
a b c d
We say that a walk is a path if it does not repeat any vertices, with the
following exception: if the first and last vertex of path are the same and
all of the others are distinct, we allow this to be a path as well. In this
e f g last case, we call our walk a cycle.
a b c d We often describe a walk by just listing its vertices in order: i.e.

v0 → v1 → v2 → . . . → vn−1 → vn
e f g
is a valid way to describe a walk.

Example 5.7. In the graph in the margins, the following sequences are
walks from e to g:
❼ e→a→f →b→e→d→f →b→g
❼ e→c→g
Notice that walks can repeat edges and vertices!

A useful property that graphs can have, related to this concept of walks,
a is being connected:

Definition 5.10. Given a graph G, we say that G is connected if for

b every pair of vertices a, b in G, there is a path from a to b in G.

Example 5.8. The boxed graph at left is not connected, as there is no

path from a to b in this graph. The one at right, however, is connected,
as there is a path from any vertex to any other vertex.

In many applications, we’ll want to find the shortest path between

two objects: i.e. in a transit graph you’ll want to find the path with the
shortest possible length, or in the internet graph you’ll want to find the
path with the lowest total ping. To capture this idea, we’ll often want
to attach weights to the edges of our graph, to represent paths that are
physically longer / more expensive to use / etc. We do this as follows:

Definition 5.11. An edge-weighted graph G is a graph G along with

a function f that assigns every edge a number.

Example 5.9. A map of the South Island of New Zealand is drawn

below. We can turn this into a graph by replacing each region with a
vertex, and connecting two regions if they border each other.
With this done, we can turn this graph into a weighted graph by
labelling each edge with the total amount of time it would take you to
drive from the largest city in one region to the largest city in the next
region. This gives us the graph below:

82

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Tasman Tasman
3 4
Buller Buller Canterbury
7 5 2
2
West Coast Canterbury Mid-Canterbury
West Coast 3 1
8 4 South Canterbury
6 2
Mid-Canterbury
South Canterbury Southland North Otago
4
North Otago 3 4

Otago
Otago
Southland
In graphs like this, we’ll often solve problems like the following:
Example 5.10. Suppose that you’re a traveling salesman. In partic-
ular, you’re traveling the South Island, and trying to sell rugby tickets
for nine rugby teams there (one for each region in the map above.)
You want to start and finish in Mid-Canterbury, and visit each other
region exactly once to sell tickets in it. What circuit can you take through
these cities that minimizes your total travel time, while still visiting
each city exactly once?
Without knowing any mathematics, you’d probably guess that the short-
est route is to just go around the perimeter of the island. Intuitively,
at the least, this makes sense: avoiding the southern alps is probably a
good way to save time!
In real life, however, maps can get a lot messier than this. Consider a
map of all of the airports in the world, or even just in New Zealand (at
right.) If you were an Air New Zealand representative and wanted to
visit each airport, how would you do so in the shortest amount of time
and still return home to Auckland?
Publicly-available map sourced from
In general: suppose you have n cities C1 , . . . Cn that you need to visit for http://www.airlineroutemaps.com’s
work, and you’re trying to come up with an order to visit them in that’s Air New Zealand page.
the fastest. For each pair of cities {Ci , Cj }, assume that you know the
time it takes to travel from Ci to Cj . How can you find the cheapest way
to visit each city exactly once, so that you start and end at the same
place?
These sorts of tasks are known as traveling salesman problems, and
companies all over the world solve them daily to move pilots, cargo, and
people to where they need to be. Given that it’s a remarkably practical
problem, you’d think that we’d have a good solution to this problem by
now, right?
. . . not so much. Finding a “quick” solution (i.e. one with non-exponential
runtime) to the traveling salesman problem is an open problem in the-
oretical computer science; if you could do this, you would solve a problem
that’s stumped mathematicians for nearly a century, advance mathemat-
ics and computer science into a new golden age, and quite likely go down
in history as one of the greatest minds of the millenium . So, uh, extra-credit problem.
This is a fancy way of saying “this problem is really hard.” So: why
mention it here? Well: in computer science in general, and graph theory
in particular, we often find ourselves having to solve problems that don’t
have known good or efficient algorithms. Despite this, people expect
us to find answers anyways: so it’s useful to know how to find “good
enough” solutions in cases like this!
For the traveling salesman problem, one brute-force approach you could
use to find the answer could be coded like this:

83

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Algorithm 5.9. Init: Take our graph G, containing n vertices. Let s

be the vertex we start and end at. Let c be a cost function, that
given any edge {x, y} in G outputs the cost of traveling along that
edge.
1. Write down every possible walk in which we can list the n vertices
of G, starting and ending at s.
2. For each walk {s, v1 }, {v1 , v2 }, . . . , {vn−1 , vn }, {vn , s}, calculate

c({s, v1 }) + c({v1 , v2 }) + . . . + c({vn , s}).

Assume that c({x, y}) is infinite if the edge doesn’t exist (i.e. that
it would take “forever” to travel along a walk that is impossible to
travel along.)
3. Output the smallest number/walk you find.

Points in favor of this algorithm: it works! Also, it’s not too hard to
code (try it!)
Points against this algorithm: if you were trying to visit 25 cities in a
week, it would take the world’s fastest supercomputer over ten thousand
years to answer your problem. (If you were trying to visit 75 cities,
the heat death of the universe occurs before this algorithm is likely to
terminate.)
This is because the algorithm needs us to consider every possible order
of the n vertices in G to complete. There are (n − 1)! = (n − 1) ⋅ (n − 2) ⋅
To see why, think about how you’d (n − 3) ⋅ . . . ⋅ 3 ⋅ 2 ⋅ 1 many ways in which we can order our n cities, and
make an ordering of the cities. You’d the factorial function grows incredibly quickly, as we saw before!
start by choosing a city to travel to
from s: there are n − 1 choices here, Another approach (which, as authors who would like to book their travel
as we can possibly go anywhere other before the heat death of the universe, we are in favor of) is to use ran-
than s. From there, we have n − 2 domness to solve this problem! Consider the following algorithm:
choices for our second city, and then
n − 3 for our third city, and so on/so Algorithm 5.10. Init: Take our graph G, starting vertex s, and cost
forth!
function c just like before.
1. Start from s and randomly choose a city we haven’t visited, and
then go to that city.
2. Keep randomly picking new cities until we’ve ran out of new choices,
and then return to s.
3. Calculate the total cost of that path.
4. Run this process, say, ten thousand times (which, while large, will
be much smaller than n! for almost all values of n that you’ll run
into.)
5. Output the smallest number/path you find.

Points against this algorithm: strictly speaking, it probably won’t work.

That is: we’re just repeatedly randomly picking path and measuring
their length. There’s no guarantee that we’ll ever pick the “shortest”
path!
Points in favor of this algorithm: it’s easy to code, it’s really fast, and
if you only care about just getting close-ish to the right answer it’s
actually1 not too bad in many situations!
In the long run, it’s probably better if your phone gives you slightly
suboptimal directions in a second rather than taking two years to find
the absolute best walk to the Countdown, so in general this is probably
a better way to go. But in certain small situations (or times when you
randomly have a supercomputer at hand) brute-force can also be the
way to go: it really depends on what you’re trying to solve!
To close this section, let’s use our language of walks to answer our last
remaining exercise:
1 In particular there are lots of tweaks you can apply here to make this pretty

decent in most cases, while still keeping it fast.

84

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Answer to Exercise 5.1. It is impossible to construct such a walk! To

see why, let’s use graph theory and reduce our city of Königsberg to a
multigraph, as done earlier.
With this done, we can see that each vertex in this graph has odd degree:
the three outer ones have degree 3, and the inner one has degree 5.
We claim that this means that we cannot have a circuit (i.e. a walk that
starts and ends at the same place) that only uses every edge once!
To see why, we use the “assume not” / contradiction technique that
has often worked for us in the past. Suppose that this graph did have a
circuit that used every edge exactly once. Describe this circuit in general
as follows:

{v0 , v1 }, {v1 , v2 }, {v2 , v3 }, . . . , {vn−1 , vn }, {vn , v0 }.

Pick any vertex x in our graph. Notice that each time x comes up in the
above circuit, it does so twice: if x = vi for some i, it shows up in both
{vi−1 , vi } and {vi , vi+1 }. You can think of this as saying that each time
our circuit “enters” a vertex along some edge, it must “leave” it along
another edge!
As a result, any vertex x shows up an even number of times in the circuit
we’ve came up with here. As well, we assumed that this circuit contains
every edge exactly once. Therefore, every vertex x shows up in an even
number of edges in our graph!
That is, deg(x) is even for every vertex x.
However, we know that our graph has odd-degree vertices. This is a
contradiction! Therefore, we have shown that it is impossible to solve
this puzzle.

5.4 Practice Problems

1. (-) Can you find a graph on 9 vertices in which all vertices have
degree 1? How about 10?
2. Prove or disprove: if G is a graph on an odd number of vertices,
then the degree of every vertex in G must be even.
3. Show that if G is a graph on n vertices, then G contains at most
n(n−1)
2
edges.
4. (+) Show that if G is a graph on 10 vertices with at least 37 edges,
then G must be connected.
5. (-) Find the shortest possible length of a circuit in the South Island
map from Example 5.9.
6. Suppose you used Algorithm 5.9 to solve problem 1. How many
iterations would it take to find the answer?
7. Draw a map for where you grew up. Label your home, school,
local grocery store, and a couple of your favorite places to visit
outside of home. Use Google Maps or something similar to find
the distances between these things. What’s the shortest walk that
visits all of them, starting and ending from home?
8. (+) Can you draw K5 without having any edges cross? Either do
so, or explain why this is not possible.

85

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Trees
2019 Chapter 6

Exercise 6.1. A graph G on 2n vertices is said to have “doubled degrees”

if it has exactly 2 vertices with degree k, for every k ∈ {1, 2, . . . n}. For
example, the graph drawn at right has doubled degrees.
How many trees exist with “doubled degrees?”

Exercise 6.2. You’re a curator of a large modern art museum! Because

A gallery guarded your museum is particularly “edgy,” the room in which you’re displaying
by 3 cameras your artwork is a very strange-looking polygon (see margins.)
You want to install 360○ cameras in the corners of your gallery, in such a
way that your cameras see the entire room. What is the smallest number
of cameras you need to install?
In general, what’s the largest number of cameras you could need for an
Your gallery! art gallery that is a n-sided polygon?

6.1 Trees
In our last section, we saw that the language of graph theory could be
used to describe tons of real-life objects: the internet, transportation,
social networks, tasks, and many other things! Even your computer (a
particularly relevant thing to consider in a computer science class) can
be described as a graph:
❼ Vertices: all of the files and folders in your computer.
❼ Edges: Draw an edge from a file or folder to every object it con-
tains.
If you draw this out, you’d get something similar to the drawing below:

C:

Windows Users Program

Files

notepad.exe Admin GHopper Steam

Documents Desktop Path

of Exile

res1.doc res2.doc meme.jpg

This graph represents the file system for your computer, and is ex-
tremely useful for organizing files: imagine trying to find a document
if literally every file on your computer had to live on your desktop, for
instance!
This graph has a particularly useful structure: starting from C: , there’s
always exactly one way to get to any other file or folder if you don’t allow
backtracking. That is: there are no files you can’t get to by starting from
your root and working your way down, and also there are no files that
you can get to in multiple different ways! This is a very nice property
for a file system to have: you want to be able to navigate to every file

86

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

in some way, and it’s very nice to know that files in different places are
different (imagine deleting a file from your desktop and having all copies
of it disappear in other places!)
We call graphs with the structural property described above trees. Trees
come up all the time in real life:
❼ PDF documents (like the one you’re reading right now!) are tree-
based formats. Every PDF has a root note, followed by various
sections, each of which contains various subsections.
❼ In genealogy and genetics, people study family trees: i.e. take
your great-grandmother, all of her children, all of her children’s
children, and so on/so forth until you’re out of relatives. This is
a tree, as starting from your great-grandmother there should only
be one way to get to any relative.
❼ Given any game (e.g. chess, or tic-tac-toe, or Starcraft), you can
build a decision tree to model possible outcomes as the game
progresses. To do so, make a vertex for the starting state. Then
make a vertex for every possible move player 1 could make, and
connect the starting state to all of these. For each of those states,
make a vertex for every response player 2 could make, and connect
those states up as well; doing this for all possible moves generates
a decision tree, which you can use to win!
In short: they’re useful!
To define what a tree is, we first need to introduce a useful concept from
graph theory that we didn’t have time to discuss last chapter:

Definition 6.1. A graph G has another graph H as a subgraph if H

is “contained within” G. In other words, if you can take G and remove
vertices and/or edges from it until you get the graph H, then H is a
subgraph of G.

For example, the graph at right has C5 as a subgraph, because we can

delete the “inside” vertices and edges to have just a C5 left over.
Note that any graph G is “trivially” a subgraph of itself, as we can just
delete “nothing” from G and have G left over.
With this stated, we can define a tree as follows:

Definition 6.2. A tree is a graph T that is connected and has no cycle

graph Cn as a subgraph.

For example, the three graphs in the margins are not trees, as each of
them has a cycle graph of some length as a subgraph.
However, the three graphs below are all trees:

Happy little trees.

We call vertices of degree 1 in a tree the leaves of the tree. For example,
the leaves of the trees above are colored green.
To get a bit of practice with these ideas, let’s prove a straightforward
claim about trees:

Theorem 6.1. If T is a tree containing at least one edge, then T has

at least two leaves.

Proof. Consider the following process for generating a path in T :

87

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Algorithm 6.11.
1. Choose any edge e = {x0 , y0 } in G.
2. Starting from i = 0, repeatedly do the following: if xi has degree
≥ 2, then pick a new edge {xi , xi+1 } leaving xi . Because T is a
tree, xi+1 is not equal to any of our previously-chosen vertices (if
it was, then we’d have created a cycle.) Stop when xi eventually
has degree 1.
3. Starting from i = 0, do the same thing for yi .

Notice that this process must eventually stop: on a tree with n vertices,
we can only put n vertices in our path because the “no cycle” property
stops us from repeating vertices. When it stops, the endpoints of the
path generated are both leaves because this is the only way we stop this
process. Therefore, this process eventually finds two leaves in any tree!

As a bit of extra practice, let’s try to use our tree language to sketch a
solution to our second exercise:
Answer to Exercise 6.2. It turns out that you can guard any n-sided
polygon (without any holes, and where all sides are straight) with at
most ⌊ n3 ⌋ cameras! To do so, use the following process:
❼ Take your n-sided polygon. By connecting opposite vertices, divide
it up into triangles.
❼ Turn this into a graph: think of each triangle as a vertex, and
connect two triangles with an edge when they share a side.
❼ This graph is a tree! (Why? Justify this to yourself.)
❼ Use this tree structure to do the following:
– Take any triangle. Color its 3 vertices red, blue, and green.
– Now, go to any triangle that shares a boundary with that
colored triangle. It will have 2 of its three vertices given
colors. Give its third vertex the color it’s currently missing.
– Repeat this process! It never runs into conflicts, because our
graph is a tree (and so we don’t have cycles.)
❼ Result of the above: every triangle has one red vertex, one blue
vertex, and one green vertex.
❼ Put a camera on the least-used color! This needs at most n/3
rounded down cameras, as we’re using the least popular of three
colors. It also guards everything, as a camera sees everything in
each triangle it’s in!

6.2 Useful Results on Trees

One particularly useful thing about trees is that they can be defined in
many different ways! In the section above, we define a tree as a con-
nected graph with no cycles. However, we have two other properties
Recall from Claim that that two state- that also characterize when a graph is a tree:
ments are equivalent if they hold
in precisely the same situations: i.e. Theorem 6.2.
whenever one is true, the other is true, there is exactly one
and vice-versa.
T is a tree is equivalent to path between any two
vertices in T .
Theorem 6.3.
T is connected
T is a tree is equivalent to
and has n − 1 edges

88

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

We prove the first of these theorems here:

Proof of Theorem 6.2. Because we’re proving that these two statements
are equivalent, we need to show that if either of them is true, then the
other statement follows. That is, if you wanted to show that “attending
office hours” and “getting an A+ in Compsci 120” were equivalent things,
you wouldn’t be satisfied if I said “everyone who got an A+ in Compsci
120 attended office hours:” you’d also want to know whether “everyone
who attended office hours got an A+”!
As such, this proof needs to go in 2 steps:
1. First, we need to show that if T is a tree, then there’s a unique
path between any two vertices in T .
2. Then, we need to show that if there’s a unique path between any
two vertices in T , then T is a tree.
We do each of these one-by-one:
1. Because T is a tree, by definition we know that T is connected.
By the definition of connected, we know that for any two vertices
x, y there is at least one path that goes from x to y. To complete
our proof, then, we just need to show that there aren’t multiple
paths between any two vertices.
To see why two distinct paths is impossible, we proceed by contra-
diction: i.e. we suppose that we’re wrong, and that it is somehow
possible for us to have two different paths linking a pair of vertices.
Let’s give those vertices and paths names: that is, let’s assume
that there are vertices x, y linked by two different paths P1 = {v1 =
x, v2 }, {v2 , v3 }, . . . {vn−1 , y} and P2 = {w1 = x, w2 }, {w2 , w3 }, . . . {wm−1 , y}.
Because these two paths are different, there must be some value i
w2 wi wi+1 wl-1
such that vi ≠ wi . Let i be the smallest such value, so that these w1 w wi-1 wl
paths agree at vi = wi and diverge immediately afterwards. v2 3
v1 v3 vi-1 vk
vi v vk-1
These two paths must eventually meet back up, as they end at the i+1

same vertex y. Let k, l be the two smallest values greater than i

such that vk = wl . Notice that this means that all of the vertices
vi−1 , vi , vi+1 , . . . vk−1 , vk , wl−1 , wl−2 . . . , wi are distinct (as otherwise
we could have picked even smaller values at which these paths met
back up.)
Now, look at the walk formed by starting P1 at vi−1 , proceeding
until vk , and then taking P2 backwards from wl until wi−1 . This
walk repeats no vertices other than the starting and ending one,
by construction. Therefore it is a cycle!
But we are in a tree, and trees do not contain cycles. Therefore
this is a contradiction to our assumption that we had two distinct
paths. In other words, our assumption that two paths could exist
was false, and we must have exactly one path, as claimed!
2. More-or-less, we can just reverse the argument above!
That is: if T has our unique path property, then every two vertices
in T are connected by a path, and so T is connected.
v w
To see why T cannot have any cycles: simply notice if T did contain
a cycle, then it would give us two different paths between two
vertices: you could go one way or the other around the cycle!
Therefore, our unique path property stops us from having cycles,
and thus means that T is a tree.

89

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This result should help us understand how our “trees don’t have cycle”
property connects to the “there’s exactly one path from C: to any other
file” property that made our filesystem example so useful!
The second of these results requires induction, and so we’ll delay its proof
until Section 7.9. It’s quite useful, though, and worth knowing even if
we can’t prove it yet! For example, it lets us solve one of our exercises:
Answer to Exercise 6.1. Let T be a tree on 2n vertices with the “dou-
bled degrees” property. Notice that if a graph T on 2n vertices has
the “doubled degree” property, then the sum of the degrees in T is
n2 + n
1 + 1 + . . . + n + n = 2(1 + 2 + . . . + n) = 2 = n2 + n.
2
As well, the degree-sum formula from our graph theory chapter tells us
that the sum of the degrees in T is twice the number of edges. Therefore,
we have that n2 + n = 2E.
Finally, because T is a tree, we know that it has one less edge than it
has vertices (i.e. it has 2n − 1 edges, because G is on 2n vertices.)
Combining this all together tells us that n2 + n = 2(2n − 1) = 4n − 2; i.e.
n2 − 3n + 2 = 0; i.e. (n − 1)(n − 2) = 0, i.e. n = 1 or n = 2. In other words,
if T is a tree with the doubled degrees property, then T is either a tree
on 2 or 4 vertices.
For n = 1, the “doubled degree” property would tell us that T should be
a two-vertex graph with two vertices of degree 1. There is exactly one
tree of this form, namely .
For n = 2, the “doubled degree” property would tell us that T should be
a four-vertex graph with two vertices of degree 1 and two of degree 2.
There is also exactly one tree of this form, namely .

6.3 Rooted Trees

In many of the examples we looked at earlier (file systems, genealogy)
it is natural to think of one vertex in our tree as the start, or root
of our tree, and of all of our other vertices as “descending” from that
root vertex. We formalize this idea with the concept of a rooted tree,
defined here:
Definition 6.3. We say that a rooted tree is any tree in which we
designate one vertex r to be the “root” of the tree.

Given a rooted tree, we can draw it as follows:

Algorithm 6.12.
level 0 0: At the top of the page, draw the root vertex r. We think of this as
level 1
“level 0” of the tree.
1: Below this vertex, draw all vertices adjacent to r along with those
level 2 edges. Call this “level 1.”
2: Below these vertices, draw all vertices that are adjacent to vertices
level 3
in level 1 that have not already been drawn. Call this “level 2.”
level 4 ...
k: In general, if level k − 1 has been drawn, draw all vertices that are
adjacent to vertices in level k − 1 that have not already been drawn.
Call this “level k.”
B
Keep doing this until you run out of vertices to draw!
B
A
A Example 6.1. Three trees are drawn in the margins. Below, we draw
B
A each of them with vertex A chosen as the root:

90

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

A A A

B B B

We do this again,but now with vertex B chosen as the root:

B B B

A A A

The following terms are useful when discussing rooted trees:

Definition 6.4. We say that the children of a vertex v are all of the
neighbors of v at the level directly below v, and the parent of v is the
neighbor of v at the level directly above v. The height of a rooted tree is
the largest level index created when drawing the graph as above.

A particularly useful tree in computer science is a binary tree:

Definition 6.5. A binary tree is a rooted tree in which every vertex

has either no children, one child, or 2 children. A binary tree is called You can generalize this to m-ary trees
full if every vertex only ever has no children or two children. for any m, by changing the restriction
here to ask that every vertex has at
most m children.
Example 6.2. Three binary trees are drawn below.

To finish out our chapter and practice working with this concept, let’s
study a quick result about binary trees:

Exercise 6.3. Suppose that T is a full binary (i.e. 2-ary) tree with 100
leaves. How many vertices does T have in total?

Answer to Exercise 6.3. There are three kinds of vertices in T :

❼ The root vertex r. Because this is a full binary tree that contains
more than one vertex, r must have exactly two children, and thus
has degree 2.
❼ All of the other parent vertices. Each of these have two children
because we’re a full binary tree, and exactly one parent by (c) +
the fact that they’re not the root. So these all have degree 3.
❼ All of the leaves, which have degree 1.
Suppose that there are p parent vertices in our graph; then the sum
of degrees in this graph is 2 + 3p + 100. As shown in class, the sum of
degrees in any graph is twice the number of edges. Because this is a
tree on 1 + p + 100 vertices (100 leaves, one root, and p parents), this
is therefore equal to 2p + 200, as any tree has one less edge than it has
vertices.
So: 2 + 3p + 100 = 2p + 200 implies that p = 98, and therefore that we have
199 vertices in total!

91

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

6.4 Practice Problems

1. (-) In a rooted tree on n vertices, what is the maximum number of
children a vertex can have?
2. Show that in a rooted tree, no vertex has two distinct parents.
3. Suppose that T is a full ternary (i.e. 3-ary) tree with exactly 99
leaves. How many vertices in total does T have?
4. Suppose that G is a graph with the following two properties:
❼ G is connected.
❼ If we delete any edge from G, G is no longer connected.
Show that G is a tree.
5. Suppose that T is a rooted full binary tree on 99 vertices. What
is the maximum height of T ? What is the minimum height of T ?
Justify your claims.
6. (+) A graceful labeling of a graph with E edges is a labeling l(v)
of its vertices with distinct integers from the set {0, 1, 2, . . . E},
such that each edge {u, v} is uniquely determined by the difference
∣l(u) − l(v)∣.
Let Pn denote the n-vertex path graph, formed by drawing n
vertices v1 , . . . vn in a row and connecting each vi to vi+1 (drawn
P3 P4 P5 P6 in the margins.)
Show that each Pn tree is graceful.
7. (+) A caterpillar tree is a tree such that deleting all of its leaves
leaves us with a single path (i.e. they kinda look like caterpillars.)
Show that all caterpillar trees are graceful.
8. (++) Show that all trees are graceful.

92

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

93

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Compsci 120 University of Auckland

Proofs
2019 Chapter 7

Exercise 7.1. You’re about to leave on holiday, but you forgot to pack
socks! You’ve ran back to your room, but the light’s burnt out, so you
can’t see the colours of your socks.
You know that in your sock drawer that there are ten pairs of green socks,
ten pairs of black socks, and eleven pairs of blue socks (all mixed up.)
How many of your socks do you need to take before you can be sure
you’ve grabbed at least one matching pair?
Exercise 7.2. You’re a mad scientist! You’ve conducted an experiment
on yourself to get superpowers. It worked, but to keep the powers you
need to take two different tablets each day; if you forget one, or take
more than one of either type, you’ll, um, explode.
Unfortunately, they look completely identical, and you’ve just dropped
your last two days of supply (four tablets) on the floor.
What can you do?

7.1 Proofs: Motivation and Fundamentals

Throughout this coursebook, we’ve stressed the importance of clear, log-
ical explanations for why things are true. In our previous chapters, we
motivated the need for these arguments in a number of ways:
❼ In our answer to Exercise 1.1, we said that making a good, logical
argument is often a useful skill in the workplace! In real life, you
will often have to work for people who aren’t particularly “tech-y”
and will expect you to be able to do literally impossible tasks. Be-
ing able to give a clear and patient explanation for why something
can never happen is a good skill to have!
❼ In our argument for Claim 1.1, we explained why an argument
that just checks a few hundred or even a few thousand cases is
often not enough in mathematics and computer science. As we
saw then, there are tons of examples of claims that are true for all
of the small values you’d like to check, but that suddenly become
completely false when you get to large values (i.e. the ones that
you could encounter when running code!)
❼ Finally, in chapter 4 we studied algorithms, step-by-step pro-
cesses that we could easily turn into computer programs. When
doing so, we often added in discussions about why these algorithms
were “guaranteed” to work or why these algorithms would never
“crash;” these arguments made it so that we could trust these
processes to always work.
This is something that you do in real life as a computer scientist
all the time! Good code involves carefully thinking about all of the
possible inputs you could be given, and for each case ensuring that
your code is well-behaved. Writing such code and documenting
it in a way that others can understand is a key part of being a
professional programmer!
These motivations aren’t just hypotheticals! Here’s a pair of stories that
illustrate why knowing how to make logical arguments is a useful skill:

94

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Story 7.1. Janelle Shane is a researcher in optics who works with neural
networks and machine learning. Roughly speaking, the way that a neural
network works is the following:
❼ Take a bunch of examples of the thing you want the neural network
to recognize, as well as a bunch of nonexamples. “Show” is hard to define in words, but
there are some great YouTube videos:
❼ “Show” the neural network these examples and nonexamples. see SethBling’s Mar I/O videos for a
❼ The neural network will then come up with a set of rules that it fun and accessible introduction!
believes describes what it means for
People are often tempted to just use the results of a neural network
directly, without checking whether its discovered rules make sense. Doing
so, as Janelle notes, leads to some fascinatingly weird behaviour:
❼ “There was an algorithm that was supposed to sort a list of num-
bers. Instead, it learned to delete the list, so that it was no longer
technically unsorted.”
❼ “ In 1997, some programmers built algorithms that could play tic-
tac-toe remotely against each other on an infinitely large board.
One programmer, rather than designing their algorithms strategy,
let it evolve its own approach. Surprisingly, the algorithm suddenly
began winning all its games. It turned out that the algorithms
strategy was to place its move very, very far away, so that when its
opponents computer tried to simulate the new greatly-expanded
board, the huge gameboard would cause it to run out of memory
and crash, forfeiting the game.”
❼ “An algorithm that was supposed to figure out how to apply a
minimum force to a plane landing on an aircraft carrier. Instead,
it discovered that if it applied a *huge* force, it would overflow
the programs memory and would register instead as a very *small*
force. The pilot would die but, hey, perfect score.”
In short: just because something works for a bunch of examples doesn’t
mean it’s good!
Story 7.2. A somewhat darker story on the importance of being able
to read and understand proofs comes from the NSA, and something
called a Dual Elliptic Curve Deterministic Random Bit Generator. This
was an algorithm, designed by the NSA (a USA security agency,) that
they claimed was a cryptographically secure way to generate random
numbers. See the New York Times for an article
summarizing the scandal.
However, this algorithm was one that the NSA had built a “backdoor”
Also, check out Kleptography: Using
into. That is, they designed the algorithm around certain secret values so Cryptography Against Cryptography
that anyone with knowledge of those values (i.e. the NSA) could predict and Cryptanalysis of the Dual Ellip-
the randomly-generated numbers with a higher-than-normal degree of tic Curve Pseudorandom Generator ,
if you’d like to read through some re-
accuracy and thereby defeat cryptographic systems using this algorithm.
search papers describing the NSA’s al-
The NSA managed to get their algorithm used as a “standard” for over gorithm/its weaknesses.
seven years. However, many mathematicians and computer scientists
were suspicious of the NSA’s algorithm from the very start, in large
part because it was not something that was proven to work!
Their research led to the eventual revocation of the NSA’s algorithm as
a standard.

So: we have some motivation for why we would want to write clear,
logical arguments. The next question for us, then, is what counts as a
valid argument?
Every major field of study in academia, roughly speaking, has a way of
“showing” that something is true. In English, if you wanted to argue
that the whale in Melville’s Moby Dick was intrinsically tied up with

95

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

mortality, you would write an essay that quoted Melville’s story alongside
some of of his other writings and perhaps some contemporary literature,
and logically argue (using these quotations as “evidence”) that your
claim holds. Similarly, if you were a physicist and you wanted to show
that the speed of light is roughly 3.0⋅108 meters per second, you’d set up
a series of experiments, collect data, and see if it supports your claim.
In mathematics, a proof is an argument that mathematicians use to
show that something is true. However, the concepts of “argument” and
“truth” aren’t quite as precise as you might like; certainly, you’ve had
lots of “arguments” with siblings or classmates that haven’t proven some-
thing is true!
In mathematics, the same sort of thing happens: there are many ar-
guments that (to an outsider) look like a convincing reason for why
something is true, but fail to live up to the standards of a mathemati-
cian. In Chapter 1, we already studied a pair of “failed” proofs: namely,
our first attempts at proving Claim 1.1 and Exercise 1.1. We said that
these arguments failed because they did not work in general: that is,
they only considered a few cases, and did not consider all of the possible
ways to put dominoes on a chessboard, or to pick a pair of integers.
This, however, is not the only way in which a proof might fail us! Here’s
another dodgy proof:

√
x+y
Claim 7.1. Given any two nonnegative real numbers x, y, we have 2
≥
xy.

“Bad” proof: √ x+y

xy ≤
2
(x + y)2
xy ≤
4
4xy ≤ (x + y)2
4xy ≤ x2 + 2xy + y 2
0 ≤ x2 − 2xy + y 2
0 ≤ (x − y)2 .
2
A defense of the “bad” proof: We’re not using examples; we’re working
in general! Also, we totally showed that this claim is true: after all, we
started with our claim and turned it into a true thing!
Why this proof is not acceptable in mathematics:
❼ We have no idea what x and y are! In particular, by plugging in
√
some sample values of x and y, we can see that this is sometimes
√ √
true and sometimes false: for x = 1, y = 4 we do indeed have xy =
4 = 2 ≤ 1+4 = 2.5, but for x = −1, y = −1 the claim (−1) ⋅ (−1) ≤
as −1 ≤/ 1! So, to do anything here, we first need
2
−1−1
2
is very false,
to know what x and y are. That is: we need to define what set
x, y come from!
❼ This proof is “backwards:” that is, it starts by assuming our claim
is true, and from there gets to a true statement. This is not a log-
ically sound way to make an argument! For example, if we assume
that 1=2, we can easily deduce a true statement by multiplying
both sides by 0:
1=2
⇒0 ⋅ 1 = 0 ⋅ 2
⇒0 = 0.

96

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This doesn’t prove that 1=2, though! As we said above, proofs

need to start with true things, and then through argument get to
what you’re trying to show.
❼ Finally, this proof has no words! This flaw in some sense is why the
other two flaws could exist: if you had to write out in words what
x and y were, and how you went from one line to the next, it would
probably become clear that this proof was written backwards and
also that we have to be careful with what x, y are allowed to be.
This sort of thing is often easy to fix, though! If your proof is “back-
wards,” simply try starting from the end and reasoning your way back-
wards to the start. If your logic was flawed, somewhere along the way
you’ll encounter a nonreversible step.
For example, if we tried to reverse our proof that 1 = 2, we could go from
0 = 0 to 0 ⋅ 1 = 0 ⋅ 2, but would see that we can’t “divide by 0” to get to
the desired conclusion (and thus that this doesn’t work.)
With this in mind, let’s try a “fixed” version of this proof:

Theorem 7.1. (The arithmetic mean-geometric mean inequality.) For

any two nonnegative real numbers x, y, we have that the geometric mean
of x and y is less than or equal to the arithmetic mean of x and y: in
other words, we have that
√ x+y
xy ≤ .
2

Proof. Take any pair of nonnegative real numbers x, y. We know that

any squared real number is nonnegative: so, in specific, we have that
the square of x − y, (x − y)2 is nonnegative. If we take the equation
0 ≤ (x − y)2 and perform some algebraic manipulations, we can deduce
that

0 ≤ (x − y)2
⇒ 0 ≤ x2 − 2xy + y 2
⇒ 4xy ≤ x2 + 2xy + y 2
⇒ 4xy ≤ (x + y)2
(x + y)2
⇒ xy ≤ .
4

Because x and y are both nonnegative, we can take square roots of both
sides to get

√ ∣x + y∣
xy ≤ .
2

Again, because both x and y are nonnegative, we can also remove the
absolute-value signs on the sum x + y, which gives us
√ x+y
xy ≤ ,
2
which is what we wanted to prove.

Much better! This proof doesn’t have logical flaws, it’s easier to read,
and we’ve justified all of our steps so that even a skeptical reader would
believe us.

97

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

7.2 Direct Proofs

In our past chapters, we’ve seen a
number of useful techniques for writ- The proof we just wrote serves as a nice example of the first proof tech-
ing such arguments while studying new nique we’ll study in this class: the idea of a direct proof. To prove
mathematical concepts. In this chap-
ter, we’re reversing this structure a bit: that a given claim is true, the most straightforward path we’ve used in
instead of focusing on new mathemat- this class has been the following:
ics, we’re going to focus our sections
on new argument techniques. By
❼ Write down things that you know are true that relate to your claim.
doing this, we’re hoping that the argu- This typically includes the definitions of any terms referred to in
ments you’ve been reading in the past the definition, any results from class or the tutorials/assignments
few chapters will become ones that that look related, and maybe some fundamental facts you know
you’re comfortable with writing and
reading on your own!
entering this class about numbers.
❼ Combine those things by using logic or algebra to create more
things you know are true.
❼ Keep doing this until you get to the claim!
A particularly common form of direct proof comes up when people want
to prove a statement of the form “if A holds, then B must follow” for two
propositions A and B (or equivalently, “A implies B,” which we write in
symbols as A ⇒ B.)
To write a direct proof of such a statement, we proceed as before, but
also throw in the assumption that A holds! That is, to prove “A implies
B,” we assume that A is true, and try to combine this assumption with
other known true things to deduce that B is true. (Logically speaking,
this is because A ⇒ B holds as long as we’re never in the situation where
A is true and B is false. Therefore, if we can show that A being true
forces B to also be true, then we know that our claim must hold!)
We illustrate this with a pair of examples here:

Claim 7.2. If n is an odd integer, then n2 can be written as a multiple

of 4 plus one.

Proof. We start by “assuming” the part by the “if:” that is, we assume
that n is an odd integer. By definition, this means that we can write
n = 2k + 1 for some other integer k.
We seek to study n2 . By our observation above, this is just (2k + 1)2 =
4k 2 + 4k + 1 = 4(k 2 + k) + 1. This is a multiple of 4 plus 1, as claimed!
Therefore we have completed our proof.

Claim 7.3. If G is a graph, then G must have an even number of vertices

with odd degrees; that is, it is impossible to have a graph G with an odd
number of vertices with odd degrees.

Proof. We start this proof by thinking about all of the facts that we know
about graphs and degrees. There’s one result that should immediately
jump to mind, namely the degree-sum formula: for any graph G,

The sum of the degrees Twice the number

=
of the vertices in G of the edges in G

Let’s use this result! Specifically: in this problem, we’re studying vertices
with odd degree. How can we turn this result into something that talks
about odd-degree vertices? Well: from our work in our first chapter, we
know that every integer is either even or odd. If we apply this idea to
our degree-sum formula, we get the following:

The sum of the The sum of the Two times the

odd degrees of + even degrees of = number of the
vertices in G vertices in G edges in G

98

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

We wanted to study the odd-degree vertices, so let’s get them isolated

on one side of our equation:

The sum of the Two times the The sum of the

odd degrees of = number of the − even degrees of
vertices in G edges in G vertices in G

On the right-hand side, notice that we have an even number (twice the
number of edges) minus a bunch of even numbers (the degrees of all
even-degree vertices in G); therefore, the right-hand-side is even!
As a result, the left-hand-side is also even. But this means that the sum
of all odd-degree vertices is an even number.
We know that summing an odd number of odd numbers is always odd,
and that summing an even number of odd numbers is always even. Be-
cause the left-hand side is even, we know we must be in the second case;
that is, that we have an even number of vertices of odd degree, as
claimed!

7.3 Proof by Cases

Our second proof technique is best illustrated by an example:

Theorem 7.2. For every natural number n, if n is a square number , An integer n is said to be a square
then n ≡/ 2 mod 3. number if we can write n = k2 for
some other integer k. For example,
0, 1, 4, 9, 16, 25 . . . are all square num-
Proof. As always, we start by expanding our definitions. If n is a square bers!
number, then by definition we know that n = k 2 for some integer k.
From here, we use the particularly clever trick that this section is devoted
to: we consider cases. That is: we want to look at what n is congruent
to modulo 3.
We don’t have any information about what n or k theirselves are modulo
3, so it would seem hard to introduce this information into our proof!
However, by the definition of the modulus operator % , we know that
every number is congruent to one of 0, 1 or 2 modulo 3. By definition,
then, this means that we most always be in one of the following three
cases:k ≡ 0 mod 3, k ≡ 1 mod 3 or k ≡ 2 mod 3.
In each of these cases, we can now expand our definitions and use our
knowledge of modular arithmetic to proceed further:
1. Assume that we’re in the k ≡ 0 mod 3 case. In this situation, we
have that k ≡ 3m for some m, which means that k 2 = 9m2 = 3(3m2 )
is also a multiple of 3. Thus, k 2 ≡ 0 mod 3.
2. Now, assume instead that we’re in the k ≡ 1 mod 3 case. In this
situation, we have that k ≡ 3m + 1 for some m, which means that
k 2 = 9m2 + 6m + 1 = 3(3m2 + 2m) + 1. Thus, k 2 ≡ 1 mod 3.
3. Finally, consider the last remaining case, where k ≡ 2 mod 3. In
this situation, we have that k ≡ 3m + 2 for some m, which means
that k 2 = 9m2 + 12m + 4 = 3(3m2 + 4m + 1) + 1. Thus, k 2 ≡ 1 mod 3.
In all three of these cases, we’ve seen that n = k 2 is not congruent to 2
modulo 3. These cases cover all of the possibilities! Therefore, we know
that n is simply never congruent to 2 modulo 3 in any situation, and
have therefore proven our claim.

The trick to the proof above was that we were able to introduce addi-
tional information about k (namely, its remainder on division by 3) by
simply considering all possible remainders as separate cases! This

99

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

technique, called proof by cases, is a powerful technique in several

situations:
❼ Whenever you’re dealing with integers and answering a question
about whether some expression f (n) is even or odd, try considering
the two cases “n is even” and “n is odd.”
❼ If you’re dealing with a claim about modular arithmetic, or with
claims like “is a multiple of”, considering the different possible
remainders that a number could have (i.e. the three cases where
k % 3 = 0, 1 or 2 we considered above) is often a great approach.
❼ If you’re dealing with claims about rational and irrational numbers,
separating the cases “x is rational” and “x is irrational” can be
handy.
We practice this in the examples below:
Proof by cases is an excellent technique Claim 7.4. For any real number x, we claim that ∣x + 7∣ − x ≥ 7.
whenever you see an absolute value, as
it lets you get rid of the absolute value
Proof. We proceed by considering cases:
in each case.
❼ We first consider the case where x ≥ −7. In this case, x + 7 ≥ 0, and
so ∣x + 7∣ − x is just x + 7 − x = 7. In this case, our inequality holds!
❼ Now, we consider the case where x < −7. In this case, x + 7 < 0,
and so we have ∣x + 7∣ = −(x + 7) (as the negative of a negative is a
positive!)
Therefore, we have ∣x + 7∣ − x = −(x + 7) − x = −7 − 2x. If x < −7¡
then −2x > 14, and so −2x − 7 > 14 − 7 = 7. Therefore our claim
holds in this case as well!
Because every number is either greater than or equal to 7 or less than
7, we’ve considered all possible cases. As our claim was true in each
possible case, this completes our proof!

Claim 7.5. For every two numbers x, y, we always have that max(x, y)+
min(x, y) = x + y.

Proof. We consider two possible cases:

❼ x > y. In this case, we have max(x, y) = x and min(x, y) = y;
therefore, max(x, y) + min(x, y) = x + y as claimed.
❼ x ≤ y. In this case, we have max(x, y) = y and min(x, y) = x;
therefore, max(x, y) + min(x, y) = y + x = x + y, also as claimed.
This covers all possible cases, as for any two numbers x, y either x > y
or x ≤ y! Therefore, we’ve proven our claim.

In the next example, we return to the tricks we used to calculate the last
digit of a number in Claim 1.10:
Claim 7.6. For any integer k, we have that (k 4 ) % 10 is always either
Proof by cases is also usually a good 0, 1, 5, or 6.
idea if you see the modulus operator!
Proof. We saw before in our chapter on integers that if d0 is the last
digit of an integer n, then nm % 10 is equal to dm
0 % 10 for any positive
integer power m.
Therefore, in our claim, we don’t have to actually consider every possible
integer k; we can just consider the ten different possible last digits k could
have, and calculate the cubes of each of those! We do so here:

❼ 04 % 10 = 0 % 10 = 0. ❼ 34 % 10 = 81 % 10 = 1. ❼ 64 % 10 = (36)2 % 10 = 62 % 10 = 6.
❼ 14 % 10 = 1 % 10 = 1. ❼ 4
4 % 10 = 256 % 10 = 6. ❼ 74 % 10 = (49)2 % 10 = 92 % 10 = 6.
❼ 24 % 10 = 16 % 10 = 6. ❼ 4
5 % 10 = 625 % 10 = 5. ❼ 84 % 10 = (64)2 % 10 = 42 % 10 = 6.
❼ 94 % 10 = (81)2 % 10 = 12 % 10 = 1.

100

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

In all ten cases, our remainders are always 0, 1, 5, or 6, as claimed!

Therefore, we’ve proven our claim.

Finally, we can use a proof by cases to prove one of our exercises:

Answer to Exercise 7.1. Even though there are lots of socks in the
drawer, there are only 3 colours. Therefore, we can just take 4 socks to
make sure that at least 2 of them are the same colour. To understand
why, let’s look at the colours of three first socks.
There are two possible cases here:
❼ If we were lucky enough to pick two matched socks from those first
three, then we’ve succeeded!
❼ However, in the worst-case scenario the first three socks we took
were all different colours, and we do not yet have a pair. In this
situation, we have one sock of each colour.
In this case, however, our fourth sock is guaranteed to match at
one of our three previously chosen socks!
In any case, we’ve grabbed a pair of socks, as desired.

7.4 Proof by Contradiction

Contradiction — i.e. the “if we’re stuck on a problem, suppose we’re
wrong and see what happens” proof technique — is a method we’ve
already used to considerable success throughout this coursebook! In this
section, we study several more examples of proof by contradiction, and
talk a bit about the trickier aspects of this proof method.
To start, let’s examine one of the most famous proofs by contradiction!
In this proof, we’re going to really pick apart the structure of a proof by
contradiction, so that we can see why this method works:
√
Claim 7.7. The number 2 is not rational.

√
Proof. As always, let’s start by unpacking our definitions:
❼ 2 is the unique positive real number such that when we square
it, we get 2.
❼ A number x is rational if we can write x = m
n
, where m and n are
integers and n is nonzero.
With this done, our claim can be unpacked to the following:
√
“For a real number x, if x = 2, then there are no values of m, n ∈ Z
with n ≠ 0 such that x = m
n
.”

So: how do we do √ this? Because the problem wants us to show that

we cannot write 2 = m n
for any integers m, n with n ≠ 0, we can’t just
check a few examples: we’d have to look at all of them, and this could
be quite difficult! We’d have to find some useful property that makes all
examples of this form fail, and this could be quite hard to find.
Instead, consider the following way to “side-step” these difficulties. In-
√each one fails,
stead of looking at all pairs m, n and trying to show that
let’s assume that we have one such pair m, n such that 2 = m n
!
With this assumption in hand, let’s now show that this assumption
“breaks mathematics” in some way: that starting from this assump-
tion, we can get to something we know is impossible, like 1 + 1 = 0. If we
can do this, then we know that our original assumption that there was
such a fraction mn
must have been nonsense (i.e. false), and therefore
that our claim that no such fraction exists is true!

101

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

We do this√here. Suppose that we can find two integers m, n with n ≠ 0

such that 2 = m n
. If m and n have common factors, divide through
by those factors to write mn
in its simplest possible form: that is, don’t
12
write something like 36 or 24 , write 21
2
Then if we square both sides, we get 2 = mn2
. Multiplying both sides
by n2 gives us 2n2 = m2 , which means that m2 is even (because it is a
multiple of 2)!
This means that m is even (see the tutorials from earlier in this course!),
and therefore that we can write m = 2k for some integer k. If we plug this
into our equation 2n2 = m2 , we get 2n2 = (2k)2 = 4k 2 , and by dividing
by 2 we have n2 = 2k 2 .
This means that n2 is even, and therefore that n is even as well (same
logic as before.)
But this means that both n and m are multiples of 2; that is, that they
have a common factor! We said earlier that we’d divided through by
any common factors to get rid of them, so this is a contradiction: from
√ and false.
our initial assumption we got to something that is both true
√
As a result, our original assumption (that we could write 2 = m n
) must
be false; that is, we have shown that 2 ≠ m n
for any integers m, n with
n ≠ 0, as desired. Yay!

We can generalize the form of the argument we just made above as

follows:
❼ We have a claim we’re trying to prove; let’s denote it P , for short-
hand.
❼ Instead of proving P is true directly, we want to prove that “not
P ” is impossible.
❼ To do this, we can simply do the following:
1. Assume, for the moment, that “not-P ” is actually true!
2. Working from this assumption, find a pair of contradictory
statements that are implied by “not P .” That is, find a pair
of statements Q and “not-Q” such that if P was false, both Q
and “not-Q” would both hold. Common examples are “1=1”
and “1=0”, or “n is even” and “n is false”, or “x is positive”
and “x is negative”: stuff like that.
3. This proof demonstrates that “not-P ” must be impossible,
because it implies two contradictory things (like the two si-
multaneous claims “n is even”and “n is odd.”) Mathematics
is free from false statements and contradictions; therefore, we
know that this must be impossible. In other words, “not-P ”
must be false and P must be true!
A beautiful quote about proofs by con- This is how a proof by contradiction works. You take your claim P ,
tradiction, by the mathematician G. assume it’s false, and use “not P ” to deduce contradictory statements,
H. Hardy: “[Proof by contradiction],
which Euclid loved so much, is one of
which you know mathematics cannot contain.
a mathematician’s finest weapons. It is We consider another example here:
a far finer gambit than any chess gam-
bit: a chess player may offer the sac- Claim 7.8. There are two irrational numbers a and b such that ab is
rifice of a pawn or even a piece, but a
mathematician offers the game.”
rational.

Proof. In the example we’re studying here, we want to show that it’s
impossible for ab to be irrational for every pair of irrational numbers
a, b. To do this via a proof by contradiction, we do the following: first,
assume that ab is irrational for every pair of irrational √
numbers a, b! If
we apply this knowledge to one of the few numbers ( 2) we know is
irrational, our assumption tells us that in specific

102

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

√ √2
2 is irrational.

What do we do from here? Well: pretty√much the only thing we have

is our assumption, our knowledge that 2 is irrational, and our new
√ √2
belief that 2 is also irrational. The only thing really left to do,
√ √2 √
then, is to let a = 2 , b = 2, and apply our hypothesis again. But
this is excellent! On one hand, our we have that ab is irrational by our
hypothesis. On the other hand, we have that ab is equal to

√ √2 √ √2⋅√2 √ 2
√
2
( 2 ) = 2 = 2 = 2,

which is clearly rational. This is a contradiction! Therefore, we know

that our hypothesis must be false: there must be a pair of irrational
numbers a, b such that ab is rational.

An interesting quirk of the above proof is that it didn’t actually give us

a pair of irrational numbers a, b such that ab is rational! It simply told
us that either
√ √2 √
❼ 2 is rational, in which case a = b = 2 is an example, or
√ √2 √ √2 √
❼ 2 irrational, in which case a = 2 , b = 2 is an example,
but it never actually tells us which pair satisfies our claim! This is a weird
property of proofs by contradiction: they are often nonconstructive
proofs, in that they will tell you that a statement is true or false without
necessarily giving you an example that demonstrates the truth of that
statement.
To stick with the classical route, let’s study another one of the first proofs
by induction, that we considered all the way back in our first chapter:

Claim 7.9. (Euclid) There are infinitely many prime numbers.

Proof. As we did with our argument that no number can be both even
and odd at the same time, let’s approach this with a bit of a thought
experiment: what would happen if there were not infinitely many prime
numbers?
Well: if this were to happen, then there would be some fixed number of
primes in existence. Let’s give that number a name, and say that there
were n primes in existence. Then, if we had a piece of paper with n
lines on it, we could in theory write down all of the prime numbers that
existed!
If we labeled those lines 1, 2, . . . n, we could then refer to those prime
numbers by their labels: that is, we could refer to our prime numbers
by calling them p1 , p2 , p3 , . . . pn . (Giving things names: a very useful
technique!)
In this world where we have all of these prime numbers, what can we
do with them? Well: as we saw before, a particularly useful property
about prime numbers is that they form the building blocks out of which
we can make all integers. Therefore, we’re motivated to take our primes
and stick them together, and see what happens!
After a lot of effort, you might eventually hit on the clever combination
of our prime numbers that Euclid discovered: think about what happens
if we multiply all of our prime numbers together, and then add 1 to that
entire sum. That is: look at the number

M = 1 + (p1 ⋅ p2 ⋅ p3 ⋅ . . . ⋅ pn )

103

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

On one hand: take any of the prime numbers on our list. To indicate
that we’re taking a general prime number from our list, let’s refer to
that prime number as pi , where i could be any index. Look at Mpi
. By
definition, this is equal to

1 ⎛ ³¹¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ · ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ ¹ µ ⎞
all of the primes except for pi

+⎜ p1 ⋅ p 2 ⋅ . . . ⋅ p n ⎟.
pi ⎝ ⎠

In particular, notice that this is not an integer! p1i is some fraction

strictly between 0 and 1, because pi is a prime and therefore at least
2, while the right-hand-bit is a product of integers and therefore is an
integer itself.
Therefore, we’ve shown that if we multiply pi by a number to get M ,
that number cannot be an integer; in other words, we have shown that
pi is not a factor of M . This holds for any of our primes, because pi was
an arbitrary prime; therefore M is not a multiple of any of our prime
numbers!
On the other hand, though, we know that M is an integer. Therefore,
we know that we can factor M into prime numbers! Do so, and write M
as a product of prime numbers.
Our argument above tells us that that none of those prime numbers can
be from our list p1 , . . . pn . But this list was supposed to contain all of
the prime numbers! Therefore, we know that our original assumption
that we could write down all of the prime numbers must have been false:
that is, there must have been infinitely many prime numbers.

7.5 Common Contradiction Mistakes: Not

Understanding Negation
In a proof by contradiction, we’re trying to prove that a claim P is true
by showing that “not-P ” cannot be false. Perhaps surprisingly, the most
common mistake people make when using a proof by contradiction is in
their very first step: specifically, in writing just what “not-P ” is for a
given claim!
For example, consider the following claim:

Claim 7.10. For every pair of integers x, y such that x and y are both
odd, we have that x ⋅ y is also odd.

Here are a number of incorrect ways that people will try to negate this
claim:
1. “For every pair of integers x, y such that x, y are both even, we
have that x ⋅ y is also even.”
The first mistake made here is in the first two words, where we
wrote “for every!” That is: Claim 7.10 is a claim about all pairs
of integers. As such, if someone were to say that P was false,
they’d just have to have one counterexample to prove us wrong!
That is: if someone made a claim that every UoA student was
enrolled in Compsci 120, you wouldn’t prove them wrong by trying
to show that every UoA student is not enrolled in Compsci 120;
you’d just have to find at least one student not in Compsci 120.
This tells us the first part of how we should write the negation of
this claim: it should go “There is a pair of integers x, y. . . ”

104

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

2. This, however, is not enough! That is: “There is a pair of integers

x, y such that x, y are both even and x ⋅ y is also even” is also not
the negation of our claim.
To see why this fails, note that Claim 7.10 is a claim about all pairs
of odd integers. As such, if we’re trying to say that this claim fails,
we still need to work in the same universe as our normal claim, and
we need to find a counterexample that consists of a pair of odd
integers.
That is: if someone told you “every Compsci 120 student who’s
never been to Antartica currently has an A,” you don’t disprove
them by trying to find a student who’s been to Antarctica! Their
claim was about people who haven’t been to Antarctica; to disprove
it, you need to work within the same bounds!
As such, the correct negation of Claim 7.10 is the following:

“There is a pair of integers x, y such that both x, y are odd, and yet
x ⋅ y is even.”

Much more reasonable! Using similar logic, the claim

Let’s consider another claim: Claim 7.11. “There is an even prime
number.”
Claim 7.12. If G is a graph containing ≥ 2 vertices, then G contains
should negate to the following:
two vertices whose degrees are equal.
“Every prime number is odd.”
There are many tempting and incorrect ways to negate this claim:
This is because the opposite of a claim
1. ‘If G is a claim containing < 2 vertices, then G does not contain 2 about something existing is that there
are no counterexamples (i.e. a claim
vertices whose degrees are different.” This is the same “negating
about everything). As well, the
the universe” error from before! universe of numbers we’re studying
(primes) should remain the same, leav-
That is: our claim is about graphs on 2 or more vertices. Its nega-
ing only the conclusion (even) to flip to
tion should still talk about graphs on 2 or more vertices! As well, “odd.”
our claim was about vertices whose degrees agreed. Its negation
should still talk about vertices with equal degree!
2. This suggests the following as a fix: “If G is a graph containing
≥ 2 vertices, then G does not contains two vertices whose degrees
are equal.”
This has the right universe, but it fails for a second, more subtle
reason: the opposite of “if A then B” is not “if A then not-B.”
That is: suppose that someone claimed to you “if you attend tu-
torials in Compsci 120, you’ll pass the class.” The above strategy
would say that you could disprove their claim by saying “if you
attend tutorials in Compsci 120, then you won’t pass the class.”
This doesn’t really make sense, though! Their claim here is a really
strong guarantee: it says that everyone who attends tutorials in
Compsci 120 will pass. To disprove this, you don’t need to show
that everyone who attends tutorials won’t pass; that’s way too
hard! Instead, you’d just need to find at least one person who
(1) attended all of the tutorials but (2) didn’t pass the class.
That is: the opposite of “if A then B” is “there is a situation where
A holds and B fails.”
By using this, the correct negation of Claim 7.12 is the following:

“There is a graph G on n ≥ 2 vertices, such that G does not contain

two vertices with the same degree.”

We can summarize the observations we made above as follows:

105

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Observation 7.16. The phrases “For every” and “There exist” get
switched around when writing a proof by negation. This is because we
disprove a claim about everything by finding a single counterexample,
and we prove that no example of a thing can exist by showing that
everything is not a counterexample!
Observation 7.17. The “universe” of a claim remains the same: i.e.
we don’t disprove a claim about all even numbers by studying odd num-
bers.
Observation 7.18. The opposite of an “if A then B” statement is “there
is a situation where A holds and B fails.” That is: if someone tells you
that when it rains outside the sidewalk gets wet, you just need to find a
situation where (1) it’s raining and (2) some bit of sidewalk is still dry
to disprove their claim!

To finish this section and put this to use, let’s prove Claim 7.12!

Proof of Claim 7.12. As noted above, the contradictive assumption here

would be that G is a graph on n ≥ 2 vertices in which all of the vertices
have different degrees.
We know that the maximum degree of any vertex in G is n − 1, because
any vertex is at most adjacent to every other vertex. As well, the
minimum degree of any vertex is trivially 0. Therefore, there are in
theory n different possible degrees for the n vertices in G, namely the
values 0, 1, . . . n − 1.
If no degrees are repeated in G, then (because there are n vertices and n
different possible degrees) there is exactly one vertex with degree i, for
every i ∈ {0, 1, . . . n − 1}. If n ≥ 2, note that in particular n − 1 ≥ 1, and
so the degree-0 and degree-(n − 1) vertices are different.
Now, notice that if there is a vertex with degree n − 1, it is connected
to every other vertex in our graph. In particular it must be connected
to the vertex that has degree 0, which contradicts the property that this
vertex is supposed to have degree 0.
Therefore we have a contradiction, and can conclude that our original
claim must hold.

7.6 Proof by Construction

In many of the proofs above, we’ve been focused on proving claims about
“all” numbers x, y, or “all” odd integers n, or other sorts of “universal”
claims about things. When we’re proving claims of these forms, we need
to use techniques and arguments like the ones above where we work in
general / don’t get to use examples to prove our claim!
Sometimes, however, we’ll find ourselves with claims of the form “There
exists a number n such that...” or “There is a value x with the prop-
erty...” In this sort of situation, we’re not being asked to show that
something is true for all values: instead, we’re just asked to find a single
example!
In situations like this, a common technique is proof by construction,
where we simply create an object with the desired properties. We illus-
trate this with an example:
Claim 7.13. There is an odd integer that is a power of two.

Proof. Notice that 20 = 1. Therefore, 1 is a power of 2. 1 is also odd, as

we can write it in the form 1 + 2k for some integer k (specifically, k = 0.)
Therefore we’ve constructed the claimed integer, as desired.

106

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Notice the following two aspects of this proof:

❼ We didn’t have to work with a general integer n; instead, we got
to give a specific example! This is because our claim was of the
form “There is. . . ”, which means that we’re just asked for a sin-
gle example. If our proof had started “For all. . . ”, this would be
different, and this proof would be invalid (just like how examples
weren’t enough for a proof in our earlier “the sum of any two odd
numbers is even” claim.)
❼ Also notice that we didn’t just say “1 is the answer” and ended our
proof; we actually took the time to explain why 1 has the desired
properties. You should expect to always do this!
We give a second example, to illustrate how these sorts of things come
up in combinatorics and/or “puzzle” mathematics:
Claim 7.14. Take the aces and face cards from a standard 52-card deck.
Can you arrange them in a 4 × 4 grid so that no suits or symbols are
repeated in any row or column?

Proof. Behold!

In this proof, we don’t have much to really explain: the solution pre-
sented self-evidently has the desired property (just check every row and
column.) If it was unclear, though, we’d have to have some explanation
along with our answer!
We close by giving a pair of slightly trickier examples for how construc-
tion can work, by using processes and algorithms:
Definition 7.1. Given a graph G, a vertex coloring of G with k colors G H
is any way to assign each vertex of G one of k different colors, so that
no two adjacent vertices get the same color.

Claim 7.15. We can vertex-color any tree T with at most 2 colors.

Proof. Consider the following algorithm to paint any connected graph

G’s vertices with the colors red and blue:
Algorithm 7.13.

Init: Choose any vertex v in G, and paint it red.

107

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

1. Take all currently uncolored vertices that are connected to any red
vertices by an edge, and color them blue.
2. Take all currently uncolored vertices that are connected to any blue
v
vertices by an edge, and color them red.
3. If there are any uncolored vertices left, go back to (i) and repeat.

We claim that this algorithm will always succeed at coming up with a

valid vertex coloring of any tree G; indeed, more generally, we claim
that this algorithm will always succeed at making a valid 2-coloring of
any graph G that doesn’t contain an odd-length circuit as a subgraph!
(bipartite!)
Because any tree does not have an odd-length circuit as a subgraph
(indeed, it doesn’t contain a cycle subgraph of any length), this would
prove our claim.
To see why, we use a second proof technique: contradiction! Think about
what would happen if this algorithm would fail, given a connected graph
G with no odd-length circuit subgraphs.
Because G is connected, the above process will eventually color every
vertex of G; we first color v, then its neighbors, then its neighbor’s
neighbors, and so on/so forth, coloring every vertex within a walk of k
edges by the k-th pass. So if the algorithm fails, it does so because in
its coloring there must be an edge {x, y} in which x, y both get the same
color.
Notice that if a vertex w is colored blue, it is because we can walk to
to w from our starting v in either one step, or three steps, or five steps
. . . or in general an odd number of steps. This is because we alternated
between red and blue in our algorithm. Similarly, if a vertex w is colored
red, it is because we can walk to w in either 0 or 2 or 4 or 6 or . . . or an
even number of edges.
Take any walk Px from v to x, and any other walk Py from v to y.
We have proven that either Px , Py both have an even number of edges,
or that they both have an odd number of edges. Therefore, the circuit
formed by starting at v, walking along Px to x, using the {x, y} edge to
go to y, and then reversing Py to return to v has either (even + 1 + even)
or (odd + 1 + odd) length. Both of the quantities, in particular, are odd!
This contradicts our assumption that G had no odd-length circuits.
As a result, our original claim (that G is bipartite) must be true!

Claim 7.16. A Hamiltonian circuit in a graph G is a walk that starts

and ends at the same vertex, and along the way visits every other ver-
tex exactly once. For example, the cube graph Q3 drawn at right has
a Hamiltonian circuit (highlighted.) has a Hamiltonian circuit (high-
lighted.)
The n × n grid graphs are defined by drawing a n × n grid of vertices
and connecting adjacent vertices, as drawn below:

G2,2 G3,3 G4,4

Prove for all n ∈ N that G2n,2n has a Hamiltonian circuit.

Proof. Consider the following constructive process for generating such a

circuit:

108

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ Label the (2n)2 vertices in the grid graph G2n,2n with coordinates
(i, j), where vertex (1, 1) is the vertex in the bottom-left-hand
corner and (2n, 2n) is the vertex in the upper-right-hand corner.
❼ Start at (1, 1),
❼ From this vertex, walk to the right until you’re at the bottom-right
corner (1, 2n).
❼ Go up one step to (2, 2n), and then walk back to the left until
you’re at (2, 2).
❼ Go up one step to (3, 2), then walk back to the right until you’re
at (3, 2n).
❼ Go up one step to (4, 2n), and then walk back to the left until
you’re at (4, 2).
❼ Go up one step to (5, 2), then walk back to the right until you’re
at (5, 2n).
❼ . . . Keep doing this! Eventually, you will find yourself at (2n, 2),
having walked on all of the vertices whose second coordinates are
not equal to 1, and not having visited any vertices whose second
coordinate is 1 other than (1,1). (This is where the “2n” part
comes in: because we go right on odd rows and left on even rows,
if our grid has even height then we’ll be going left on our top row
and thus wind up at (2n, 2) as claimed.)
❼ Walk from (2n, 2) to (2n, 1), and then go down to (1, 1).
By construction we have visited all vertices in our graph exactly once,
and thus created a Hamiltonian circuit, as desired.

Finally, we can answer one our earlier exercises by using construction:

Answer to Exercise 7.2. This has a nice constructive answer: take

half of each of the four tablets today, and the other half tomorrow.
Because this is constructive, we don’t have to explain how we came up
with this clever idea: we can just present it as an answer! (Though if you
did something clever to come up with an idea, it is good to mention how
you did this.) We just have to explain why this works, which is pretty
simple: half of each tablet gives us half of the tablets of each type, and
thus exactly one dose for each kind.

7.7 Proof by Induction: First Examples

Sometimes, in mathematics, we will want to study a statement P (n)
that depends on some variable n. For example:
1. P (n) = “The sum of the first n natural numbers is n(n+1)
2
.”
2. P (n) = “If q ≥ 2, we have n ≤ q n .
3. P (n) = “Every polynomial of degree n has at most n roots.”
4. P (n) = Take a 2n × 2n grid of unit squares, and remove one square
from the top-right-hand corner of your grid. The resulting shape
can be tiled by - shapes.
For any fixed n, we can usually use our earlier proof methods to prove
that the claim holds! For instance, let P (n) be the fourth example
above, and consider P (3), which is the claim that if we take a 8 × 8 grid
of squares and delete the top-right-hand corner square, we can tile the
rest of the shape with tiles. We can prove this by construction by
just giving an explicit way to do it: see the drawing at right!

109

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

However, sometimes we will want to prove that one of these statements

holds for every value n ∈ N. How can we do this?
The answer here is mathematical induction! Mathematical induction
is just a formal way of writing up our “building-block plus preserved
property” process, in a way that will hopefully let us avoid everyone
having the same shoe size. We describe it here:
❼ To start, take a claim P (n) that we want to prove holds for every
n ∈ N.
❼ The first step in an inductive proof is the base step: in this step,
we explicitly prove that the statement P holds for a few small
cases using normal proof methods (typically construction or just
calculation.)
Usually you just prove that your claim holds when n = 0, but
sometimes you start with n = 1 if your claim is one where 0 is a
“dumb” case, or prove a handful of cases like n = 0, 1, 2, 3 to get
the hang of things before moving on. You can think of this as the
“building block” step from before!
❼ With this done, we move to the induction step! Here, we prove
the following statement:
If our claim P is true for all values up to some n,
then it will continue to be true at the next value n + 1.
Because this is an implication, i.e. an if-then proof, we usually
prove it directly by assuming that our claim holds for all values up
to some n, and then use this assumption to prove that our claim
holds when we have n + 1 in our claim.
Just doing these two steps shows that your claim P is true for every
natural number n! To see why, just examine what these two steps tell
you:
❼ By our “base case” reasoning, we know that our claim is true at
n=0.
❼ By our “inductive step” reasoning, we know that if our claim is
true at 0, it is true at the “next” value n + 1 = 1.
❼ By our “inductive step” reasoning, we know that if our claim is
true up to 1, it is true at the “next” value 2.
❼ By our “inductive step” reasoning, we know that if our claim is
true up to 2, it is true at the “next” value 3.
❼ ...
❼ By continuing this process, we eventually get to any n! Therefore
our claim is true for every n ∈ N, as desired.
The way we usually think of inductive proofs is to think of toppling
dominoes. Specifically, think of each of your P (n) propositions as indi-
vidual dominoes – one labeled P (0), one labeled P (1), one labeled P (2),
and so on/so forth. With our inductive step, we are insuring that all of
our dominoes are lined up – in other words, that if we’ve knocked over
some of them, the “next one” will also be knocked over. Then, we can
think of the base step as “knocking over” the first domino. Once we do
that, the inductive step makes it so that all of the later dominoes also
have to fall, and therefore that our proposition must be true for all n
(because all the dominoes fell!)
To illustrate how these kinds of proofs go, let’s go back to our tiling
problem, and prove that we can tile this grids for every n ∈ N! (As an
added bonus, let’s prove it for grids where we remove one square from
anywhere, not just the top-right-hand corner!)

110

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Claim 7.17. For any n ∈ N, take a 2n × 2n grid of unit squares, and

remove one square from somewhere in your grid. The resulting grid can
be tiled by - shapes.

Proof. As suggested by the section title, we proceed by induction, where

our proposition P (n) is “we can tile a 2n × 2n grid of 1 × 1 squares with
one square deleted by using - shapes.”
Base case: we want to prove P (0). So: what *is* P (0)?
Well: for n = 0, we have a 20 ×20 = 1×1 grid, which we’ve removed a 1×1
square from. In other words, we have nothing. If you want, you can
think of “nothing” as being something we can trivially cover by placing
no three-square shapes!
Alternately, you can decide that 0 is a stupid case and look at n = 1
instead. For n = 1, we simply have a 2 × 2 grid with one square punched
out. As this *is* one of our three-square shapes, we are done here; just
place a tile on top of our grid!
Either starting place is fine. In general, we recommend doing as many
base cases as you need to do in order to feel comfortable with the pattern
and believe that you’ve done something concrete! Most of the time,
though, the base case will feel kinda silly; don’t worry about this! The
inductive step will do all of the heavy lifting for us.
Inductive step: We want to prove that if we know that our claim holds
up to n, then it holds for n + 1 as well; formally, this means that we want
to show that if P (0) and P (1) and . . . and P (n) all hold, then P (n + 1)
must follow.
In this problem in particular, this means that we’re assuming that we
can tile a 2k × 2k -grid with a square deleted for any k ≤ n, and want to
use this assumption to tile a 2n+1 × 2n+1 grid with a square deleted.
To do this, take any 2n+1 × 2n+1 grid with a square deleted. Divide it
into four 2n × 2n squares by cutting it in half horizontally and vertically.
Finally, by rotating our grid if needed, make it so that the one missing
square is in the upper-right hand corner.
Take this grid, and carefully cut out one three-square shape in the center
as drawn at right.
Now, look at each of the four 2n × 2n squares in this picture. They all
are missing exactly one square: the upper-right hand one because of our
original setup, and the other three because of our placed three-square-
shape.
By our inductive hypothesis P (n) we know that all of these smaller
squares can be tiled! Doing so then gives us a tiling of the whole shape;
in other words, we’ve shown how to use our P (n) results to get a tiling
of the 2n+1 × 2n+1 grid.
As this completes our inductive step, we are thus done with our proof
by induction.

The claim we proved above — one where we were some sense “growing”
or “extending” a result on small values of n to get to larger values of n
— is precisely the kind of question that induction is set up to solve! The
Fibonacci numbers, which we introduce in the next question, is another
object where this sort of “extension” approach is useful to consider.

Definition 7.2. The Fibonacci numbers fn are defined by a recur-

rence relation as follows:
❼ f0 = 0, f1 = 1.
❼ For any n ≥ 2, fn = fn−2 + fn−1 .

111

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

To illustrate how it works, let’s use it to calculate the first few values of
the Fibonacci sequence! We know that f0 = 0, f1 = 1 by definition.
To find f2 , we can use the fact that for any n ≥ 2, fn = fn−2 + fn−1 to
calculate that.

f2 = f0 + f1 = 0 + 1 = 1.

We can calculate further values of fn similarly (see right!)

f3 = f1 + f2 = 1 + 1 = 2, When doing this, you’ll likely notice a number of interesting properties
f4 = f2 + f3 = 1 + 2 = 3, about the Fibonacci sequence: see
f5 = f3 + f4 = 2 + 3 = 5, https://en.wikipedia.org/wiki/Fibonacci number
f6 = f4 + f5 = 3 + 5 = 8,
for a ton of weird/beautiful properties these numbers have! We prove
f7 = f5 + f6 = 5 + 8 = 13,
f8 = f6 + f7 = 8 + 13 = 21,
one of these properties here:
f9 = f7 + f8 = 13 + 21 = 34,
f10 = f8 + f9 = 21 + 34 = 55, Claim: For any n ∈ N, the n-th Fibonacci number is even if and
f11 = f9 + f10 = 34 + 55 = 89, only if n is a multiple of 3.
f12 = f10 + f11 = 55 + 89 = 144,
⋮ Proof. Let P (n) denote the claim “(the n-th Fibonacci number is even)
⇔ (n is a multiple of 3).” We want to prove that P (n) holds for all
n ∈ N, and proceed to prove this claim by induction.
Our base cases are pretty easy to check! We calculated the Fibonacci
numbers from f0 to f12 above, and we can see that the only ones that are
even are f0 , f3 , f6 , f9 and f12 ; so we know that P (0), P (3), P (6), P (9),
and P (12) all hold.
We now move to the inductive step: here, we want to prove P (0)
and P (1) and P (2) and . . . and P (n), when all combined together, im-
ply P (n + 1). We start with what we’re assuming, namely that all of
P (0), P (1), . . . P (n) are all true: that is, we’re assuming that the k-th
Fibonacci number is even if and only if it is a multiple of 3, for every
k ∈ {0, 1, . . . n}.
We want to prove P (n + 1), i.e. that the n + 1-th Fibonacci number is
even if and only if it is a multiple of 3.
So: let’s consider cases! There are two possible cases for the value n + 1:
either it is a multiple of 3, or it’s not.
❼ If n + 1 is a multiple of 3, we can write n + 1 = 3k for some k ∈ Z.
Notice that this means that n = 3k − 1 and n − 1 = 3k − 2, and in
particular that both of the values n, n − 1 are not multiples of 3!
As a result, our inductive assumption tells us that fn , fn−1 are both
not even, because they’re not multiples of 3! But being not-even
just means that these numbers are both odd. As a result, because
fn+1 = fn + fn−1 =odd+ odd= even, we have shown that fn+1 is
even in this case.
❼ If n + 1 is not a multiple of 3, then n + 1 either has remainder 1 or 2
when we divide it by 3; this is because any number has remainder
0, 1 or 2 when divided by 3. This means we can write n + 1 = 3k + 1
or 3k + 2, for some k ∈ Z.
As a result, we can see that of the two numbers n, n−1, exactly one
of them is a multiple of 3; if n + 1 = 3k + 1 then n, n − 1 = 3k, 3k − 1,
and if n + 1 = 3k + 2 then n, n − 1 = 3k + 1, 3k. As a result, our
inductive hypothesis tells us that exactly one of fn , fn−1 are odd,
and the other is even.
Therefore, because fn+1 = fn + fn−1 =(one odd number plus one
even number)= odd, we have shown that fn+1 is odd in this case.
So, by using strong induction, we have proven that fn is even if and only
if it is a multiple of 3!

112

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

7.8 Induction: Two Recurrence Relations

As we just saw in the section above, induction is a useful tool to study
recurrence relations! In this section, we continue with this theme, and
use induction to prove Claim 4.2 and Claim 4.6 from our algorithms
chapter.
Claim 7.18 (Claim 4.2). For every positive integer n, we have InsertionSortSteps(n) =
3n2 + 9n − 10
.
2
Proof. We proceed by induction. First, we can notice that the table of
values we calculated earlier validates our claim for the first few values of
n:
n 1 2 3 4 5 6 7
InsertionSortSteps(n) 1 10 22 37 55 76 100
3n2 + 9n − 10
1 10 22 37 55 76 100
2

This gives us our base case.

For the inductive step, we proceed as always: we assume that our
claim holds up to some value n, and seek to prove it for n + 1.
In particular, if our claim holds up to some value n, we have

3n2 + 9n − 10
InsertionSortSteps(n) = .
2

As well, by Claim ??, we know that

InsertionSortSteps(n + 1) = 3(n + 2) + InsertionSortSteps(n).

By combining these together, we get

3n2 + 9n − 10 6(n + 2) 3n2 + 9n − 10 3n2 + 15n + 2

InsertionSortSteps(n + 1) = 3(n + 2) + = + = .
2 2 2 2

But we know that

3(n + 1)2 + 9(n + 1) − 10 3n2 + 6n + 3 + 9n + 9 − 10 3n2 + 15n + 2
= =
2 2 2
.
In other words, we’ve shown that our claim holds at n + 1, and have thus
proven our claim by induction!

We can study MergeSort in the same way:

Claim 7.19 (Claim 4.6). MergeSortSteps(2k ) = k ⋅ 2k+2 − 2k+1 − 1, for
every natural number k.

Proof. We again proceed by induction. Again, as before, our previously-

calculated table of values suffices for a base case:
k 1 2 3 4 5 6 7
MergeSortSteps(2k ) 11 39 111 287 703 1663 3839
k ⋅ 2k+2 − 2k+1 − 1 11 39 111 287 703 1663 3839

With this established, we turn to the inductive step. Here, we again

assume that our claim holds up to some value k, and seek to prove it for
k + 1.

113

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

In particular, if our claim holds up to some value k, we have

MergeSortSteps(2k ) = k ⋅ 2k+2 − 2k+1 − 1

As well, by Claim 4.5, we know that

MergeSortSteps(2k+1 ) = 1 + 4 ⋅ 2k+1 + 2 ⋅ MergeSortSteps(2k ).

Again, by combining these together, we get

MergeSortSteps(2k+1 ) = 1 + 4 ⋅ 2k+1 + 2(k ⋅ 2k+2 − 2k+1 − 1)

= 1 + 2k+3 + 2k2k+2 − 2 ⋅ 2k+1 − 2
= k ⋅ 2k+3 + 2k+3 − 2k+2 − 1,

But we know that

(k + 1) ⋅ 2(k+1)+2 − 2(k+1)+1 − 1 = k ⋅ 2k+3 + 2k+3 − 2k+2 − 1

.
In other words, we’ve shown that our claim holds at k + 1, and have thus
proven our claim by induction!

7.9 Induction, Graphs, and Trees

Induction is a particularly useful technique to use when studying graphs
and trees! We prove three claims here, two of which you may recall from
our section on trees:
Claim 7.20. If G is a connected multigraph with loops (i.e. we allow
multiple edges, and also allow an edge to have both of its endpoints be
equal) on n vertices, then G contains at least n − 1 edges.

Proof. We proceed by induction on n. For n = 0, 1, this claim is trivially

true, as we always have that E is a nonnegative number.
This establishes our base cases, so we now turn to the inductive step:
here, we assume that our claim holds for all connected graphs on at most
n vertices, and seek to use that assumption to prove that our claim holds
for connected graphs on n + 1 vertices.
1 1 1
To do this, consider the following operation, called edge contraction.
6 2 6 2 6 2

3
We define this as follows: take any graph G and any edge e in G with two
5 4 5 4 5 distinct endpoints. We define Ge , the graph that this edge, as follows:
1 take G, delete e, and then combine e’s two endpoints together into a
6 2 6 2 2
single vertex, preserving all of the other edges that the graph has along
the way.
We draw examples of this process at right: here, we have started with
a graph on six vertices, and then contracted one by one the edges high-
lighted in red at each step.
Notice that contracting an edge decreases the number of vertices by 1
at each step, as it “squishes together” two adjacent vertices into one
vertex. It also decreases the number of edges by 1 at each step, as we
are contracting an edge to a point!
Finally, notice that contracting an edge preserves the property that our
graph is connected. To see why, take any walk

{v0 , v1 }, {v1 , v2 }, . . . {vi−1 , vi }, {vi , vi+1 }, {vi+1 , vi+2 }, . . . {vn−1 , vn }

in our graph. Notice that if we contracted an edge {vi , vi+1 } in this walk,
this would collapse the vertices vi , vi+1 into some new vertex vi⊕i+1 and

114

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

preserve all of the edges other than {vi , vi+1 }. As a result, our walk
would just become

{v0 , v1 }, {v1 , v2 }, . . . {vi−1 , vi⊕i+1 }, {vi⊕i+1 , vi+2 }, . . . {vn−1 , vn },

and thus still connects the vertices v0 , vn . Therefore, edge contraction

cannot “break” any pre-existing walks, and so preserves the property
that our graph is connected.
We can use this process to prove our claim via induction:
❼ Take any connected multigraph graph G on n + 1 vertices.
❼ Take any edge e in G with two distinct endpoints (such an edge
exists, because G contains at least two different vertices and G is
connected) and contract that edge. This gives us a new graph Ge ,
which is connected and contains n vertices.
❼ Therefore, by induction, we know that in Ge , the number of edges
is at least n − 1.
❼ We also know that G has exactly one more edge than Ge .
❼ Therefore, in G, we know that we have at least n − 1 + 1 = (n + 1) − 1
edges. In other words, we’ve proven that our claim holds for graphs
on n + 1 vertices, as desired!

Notice that this result applies to simple graphs as well, as any simple
graph is certainly a multigraph!
We can also use induction to prove Theorem 6.2! We split this result
into two parts, as it’s a longish equivalence proof:

Theorem 7.3 (Half of Theorem 6.2). If T is a tree on n vertices, then

T contains exactly n − 1 edges.

Proof. We proceed by induction. Our base case is straightforward: any

tree on 1 vertex clearly has no edges (as it’s a simple graph.) If you
want, you can also consider 2-vertex graphs as well; the only connected
two-vertex graph is , which has one edge as desired.
For the inductive step, let’s assume that our property holds for all trees
on up to n vertices. Let T be any tree on n + 1 vertices; we want to use
our assumption to prove that T contains exactly (n + 1) − 1 = n edges.
To do this, let l be a leaf vertex in T (we know that l exists by our earlier
theorem.) Delete l and the edge connecting l to the rest of T from T ;
call the resulting graph T − l.
T − l contains n vertices, because we started with n + 1 vertices and
deleted one vertex. It is also still connected (because l was degree 1, the
only walk that would need to use the edge to l is a walk going directly to
l, and we deleted l.) Finally T − l contains no cycle subgraphs, because
T contained no cycle subgraphs and deleting things from T cannot have
somehow caused a cycle to exist.
Therefore T − l is a tree! By induction, T − l contains n − 1 edges.
Therefore T itself contains (n − 1) + 1 = n edges, because T is just T − l
plus the vertex l and the single edge connecting l to the rest of T . In
other words, we’ve proven our inductive claim!

Theorem 7.4 (The other half of Theorem 6.2). If G is a connected

graph on n vertices containing exactly n − 1 edges, then G is a tree.

Proof. We proceed by contradiction; suppose that G is a connected graph

on n vertices containing n − 1 edges that is somehow not a tree.

115

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Because G is connected, the only way that G can fail to be a tree is if

it contains a cycle subgraph. Let {v1 , v2 }, {v2 , v3 }, . . . {vk , v1 } be such a
cycle subgraph.
Take G and delete the edge {v1 , v2 } from G. We claim that G is still
connected.
To see why, take any walk in G that uses the edge {v1 , v2 }, and replace
each use of {v1 , v2 } with the sequence of edges {v1 , vk }, {vk , vk−1 }, . . . {v3 , v2 }.
In other words, every time you’d go directly from v1 to v2 along that
edge, instead use the cycle to go the “other” way around!
As a result, if two vertices x, y used to be connected by a walk in G, they
are still connected after deleting {v1 , v2 }; in other words, G − {v1 , v2 } is
still connected.
But G − {v1 , v2 } is a graph on n vertices containing n − 2 edges, as we
had n − 1 edges and deleted one. But in Claim 7.20, we proved that a
connected graph on n vertices must contain at least n − 1 edges! In other
words, we have a contradiction, and so our claim that G was a tree
must have been correct.

7.10 Proof Methods: How to Choose

With all of these proof methods at our fingertips, a natural question is
this: how do you choose a method? One answer is the following:

Just try methods one-by-one until something works!

Paper is cheap, and it’s usually just a lot faster to try stuff and see which
things break than to predict ahead of time which method is “best.” Also,
most problems in maths can be solved by a number of different meth-
ods: there’s rarely a single “correct” approach to a problem! Instead,
many problems can be solved with many different techniques, and each
different proof can help illustrate a new way of thinking about the task
at hand.
With that said, though, there are clues or signs in a problem statement
that can indicate that certain techniques might be useful. There are no
hard-and-fast rules here, but the following observations often come in
handy:
❼ Are you proving a claim of the form “if (some claim A is true),
then (some other claim B is true)?” If so, a direct proof is maybe
a good idea! Write down what it would mean for A to be true, and
try to use that assumption to prove that B is also true.
❼ Are you dealing with modular arithmetic, even versus odd num-
bers, claims about “is a multiple of,” or absolute values? Cases
are often useful here. (More generally: if you have any problem
where the inputs or outputs can be split into cases, do so! Proofs
by cases often combine with other proof methods.)
❼ Are you being asked a claim of the form “Show that blah exists?”
Construction’s a good way to go here! (This is opposed to claims of
the form “Show that every x has property f oo,” which you usually
do not do by construction, as it’s hard to construct every x!)
❼ Are you proving a claim where it seems like your previous results
stick together to give you a later result? (Tiling problems that
involve a general integer n, anything defined recursively like the
Fibonacci sequence, processes that have recursion in them, . . . )
When you’re writing your proof, do you find yourself wanting to

116

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

use “. . . ” to show how a pattern you’ve found continues? Then

this is probably a good candidate for induction!
Induction is often especially useful for studying the runtime of an
algorithm, or for proving that a given algorithm is “guaranteed”
to give a specific output.
❼ Are you totally stuck? Try contradiction! Contradiction often
gives you something to start from: i.e. it turns claims like “show
that every object X has property blah” into “what would happen if
an object X failed to have property blah?” This is often an easier
place to start from! It can be a lot easier to think about how to
“break” things and find contradictions of any kind, than to try to
proceed directly and argue why some very specific property must
hold.
Contradiction is a particularly nice technique if you’re trying to
show that some task is impossible: the opening line of “suppose
that this is possible” often makes proofs a lot easier to start.
❼ Are you still totally stuck? Maybe it’s false: try a disproof! Best-
case scenario: you disprove it and can move on. Worst-case sce-
nario: even if you fail at disproving it, if you think about why you
weren’t able to disprove your claim, you might be able to turn that
back into a good proof.
To get some practice with this, we solve a few problems below, and in
each proof explain why we picked the methods that we did!
Claim 7.21. For every positive integer n, 16n − 1 is a multiple of 15.

Proof. Let’s think about which of our proof methods we want to try:
❼ Direct proof: we could try this. This would involve expanding out
what it means to be a multiple of 15, and trying to use logic/known
results to get to the conclusion.
❼ Cases: even though cases is often a good technique when working
with mods / multiple problems, this is likely not a great idea here.
This is because there isn’t really a clear set of cases you’d want to
divide n into: even versus odd doesn’t seem relevant, and consider-
ing all fifteen possible remainders of n % 15 seems painful enough
to not do unless absolutely necessary.
❼ Contradiction: could do, if we’re stuck!
❼ Construction: not relevant. We’re proving something for every
integer, not building examples for some values.
❼ Induction: This doesn’t obviously look like induction, in that it’s
not clear how you’d relate 16n − 1 to the “next” value 16n+1 − 1.
With some algebraic trickery, though, this is possible! Notice that
16(16n − 1) = 16n+1 − 16 = (16n+1 − 1) − 15, and thus that we’ve
related one step to the next (in a way that involves a 15, which n 16n − 1
seems promising.) So if you saw this trick, then this is promising! 0 1-1=0
1 16 − 1 = 15
❼ Disproof: If you were suspicious of this claim, you could start by 2 162 − 1 = 255 = 15 ⋅ 17
calculating a handful of values of 16n − 1, and see if any failed to 3 163 − 1 = 4095 = 15 ⋅ 273
be a multiple of 15. 4 164 − 1 = 65535 = 15 ⋅ 4369
No obvious counterexamples immediately showed up in our table
in the margins, so let’s not try to disprove this just yet.
So, amongst our proof methods, a direct proof and induction look
promising. Let’s try induction first!
Base case: we saw in our table in the margins that our case holds for
n = 0, 1, 2, 3 and 4. So we’ve established our claim for a number of base
cases.

117

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Inductive step: For the inductive step, we assume that we’ve proven
our claim for n: i.e. that 16n − 1 is a multiple of 15. We seek to use this
claim to prove that our claim holds for the “next” value n + 1: i.e. that
16n+1 − 1 is also a multiple of 15.
This is not too hard to do! Notice that as we observed above,

16n+1 − 1 = 16n+1 − 16 + 15 = 16(16n − 1) + 15.

If 16n −1 is a multiple of 15, then by definition we can write 16n −1 = 15k

for some integer k. Doing so tells us that the right-hand-side is

16(15k) + 15 = 15(16k) + 15 = 15(16k + 1) = a multiple of 15.

Therefore, 16n+1 − 1 is also a multiple of 15! As a result, we’ve proven

our claim by induction: we showed that it holds for the first few values
of n, and then showed that it will stay true, as if it is true for some
value of n it must stay true for the “next” value n + 1.

This is not the only way you could prove this result! We could also use
a direct proof:

Proof. We want to show that 16n − 1 is always a multiple of 15, for any
positive integer n.
By definition, this holds true if and only if 16n ≡ 1 mod 15.
So: we know that 16 ≡ 1 mod 15, because 16 − 1 is itself a multiple of 15.
We also know from Claim ?? that for any positive integers a, b, c, n that
if a ≡ b mod c, then an ≡ bn mod c as well.
Combining these facts tells us that for any positive integer n, we have
16n ≡ 1n mod 15. Because 1n = 1 for all n, this gives us 16n ≡ 1 mod 15.
By definition, this means that for every positive integer n we’ve shown
that 16n − 1 is a multiple of 15, as desired!

Claim 7.22. Consider the following program puzzle(n), which is a

slightly modified version of the algorithm you studied in Practice Problem
5 on page 74. It takes in as input a nonnegative integer n, and does the
following:

(i) If n is either 0, 1, or 6, output n and stop. Otherwise, go to

(ii).
(ii) If n has two or more digits, replace n with its last digit and go
to (i). Otherwise, go to (iii).
(iii) Replace n with n2 and go to (i).

Prove that for every nonnegative odd number n, if puzzle(n) stops, it

outputs 1.

Proof. We consider proof methods:

❼ Direct proof: we could try this, in that really any proof can be
written in a direct method. (In this sense, “direct” often just
means “not worrying about a specific technique.”)
❼ Cases: This seems promising! Our program does different things
based on different inputs. As such, cases is a natural technique to
use!
❼ Contradiction: could do, if we’re stuck!
❼ Construction: not relevant. We’re proving something for every
integer, not building examples for some values.

118

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

❼ Induction: Not a great technique here. It doesn’t look like knowing

what our program does on input n would tell us much about what
it does on input n + 1.
❼ Disproof: this seems true (run the program on a bunch of odd
values if you’re skeptical!), so disproving it doesn’t seem like a
good idea.
Cases looked like the strongest approach: so let’s try that!
Take any nonnegative odd number n. Because n is a nonnegative integer,
this means that n = 1, 3, 5, 7, 9 or n is a two-digit number (whose last digit
is 1,3,5,7 or 9.)
After one iteration of our program, if n was a two-digit number it will
be replaced with one of 1,3,5,7,9; so it suffices to just understand those
five cases:
❼ If n = 1, then the program immediately stops and outputs 1, as
desired.
❼ If n = 3, then on our first iteration we square n to get 9; on our
second iteration we square again to get 81; on our third iteration we
replace this two-digit number with 1; and on our fourth iteration
we stop and output 1.
❼ If n = 5, we saw earlier that this case enters an infinite loop.
❼ If n = 7, then on our first iteration we square n and get 49; on
our second iteration we replace this two-digit number with 9; on
our third iteration we square to get 81; on our fourth iteration we
replace this two-digit number with 1, which we then output and
halt on our fifth iteration.
❼ If n = 9, we go to 81 and then 1 and then halt (as described above.)
In all of these cases, we either run forever or output 1, as claimed!

Again, we can use a more direct approach if we see it:

Proof. Take any nonnegative odd number n. Notice that if n is odd,

then no matter what our program does n will stay odd! This is because
the square of an odd number is odd, and the last digit of an odd number
is odd.
Therefore, n will never be reduced to either of 0 or 6. As a result, the
only possibilities that remain are “the program halts when n = 1” or
“the program runs forever,” as there are no other conditions that cause
our program to halt.

We close this section with a third problem about graphs:

Claim 7.23. Given a graph G, an edge coloring of G with k colors is

any way to assign each edge of G one of k different colors, so that no
two edges of the same color share an endpoint in common. An example
of an edge coloring is given at right. a 3-edge-coloring
Show that there is a graph G in which all vertices have degree 3, and yet
at least four colors are needed to create an edge-coloring of G.

Proof. While we could go through all of the proof methods again, we’ll
shortcut the process and explain why we know this is a constructive
proof: it’s asking us to show that there is a graph with some property!
This isn’t a “show all graphs have property f oo” problem or a “take any
graph G, show that it cannot be blah” task: this is just asking us to find
some single graph with a given property.
So, uh: behold the graph at right!

119

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

This is the Petersen graph P , a particularly useful counterexample to

many claims in graph theory. We claim that P is a graph that needs
four colors to properly color its edges; i.e. you cannot edge-color P with
three colors. Note that to complete our proof, we need to explain why
using just 3 colors to edge-color this grap is impossible: that is, it’s not
enough to just give our object, we also need to show that it has the
desired property!
To do this, we now need a new proof technique. We claim that if you
went through your list of proof techniques, none would stand out and
you’d get stuck for a while. In this situation, contradiction is what
we’d go to!
Here, a proof by contradiction would start as follows: suppose that we
could use only three colors to color the edges of P . Call them red, green
y and blue.
Make the following observations:
1. Notice that because every vertex of P is degree 3, every vertex of
b P has one red, one green, and one blue edge leaving it.
x
a 2. Notice that on the outer pentagon of P , we need to use all three
colors: if we tried to use just two colors to edge-color the pentagon,
then we would have two edges with the same color touching each
other.
3. Take a red edge on the outer pentagon of P . Call its two endpoints
x, y, and let the inner vertices adjacent to x, y be called a, b. (Look
at the picture in the margins to make sense of this!)
Because {x, y} is red, the edge {a, x} is not red. Therefore, the red
edge that (1) told us must be connected to a is on this inner star.
Similarly, because {x, y} is red, {b, y} is not red. Therefore, the
red edge that (1) told us must be connected to b is also on this
inner star.
Finally, because a and b are not adjacent, these red edges are
different: that is, there are two red edges in this inner star.
4. Take a blue edge on the outer pentagon of P . The same logic as
in (3) guarantees two blue edges in the inner star.
5. Take a green edge on the outer pentagon of P . The same logic as
in (3) guarantees two green edges in the inner star.
6. Conclusion: the inner start has two blue edges, two red edges, and
two green edges.
. . . but it only has five edges! Therefore this is impossible; i.e.
we’ve reached a contradiction, and our original claim (that at
least four colors are required) has been proven.

7.11 Practice Problems

1. You’re a programmer! You’ve found yourself dealing with a pro-
gram mystery(n) that has no comments in its code, and you want
to know what it does. After some experimentation, you’ve found
that mystery(n) takes in as input a natural number n, and does
the following:

(i) If n is either 0, 1, 2, or 3, output n and stop. Otherwise,

go to (ii).

120

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

(ii) If n is even, replace n with n/2 and go back to (i). Oth-

erwise, go to (iii).
(iii) Replace n with n + 5 and go to (i).

Come up with the following proofs about mystery(n):

(a) Use contradiction to prove the claim “if this program outputs
3 on input n, then n is not a power of 2.”
(b) Disprove the claim “Given any natural number n as input, this
program will eventually stop” by finding a counterexample.
(c) Prove by construction the claim “There is some input to this
program that causes it to output 0.”
(d) Write a direct proof that if “the output of this program is 1”
then “the input to this program was 1.”
2. (-) Let a, b, c be three integers, such that a divides b and a divides
c. Write a direct proof that a also divides b − c.
3. (-) Write a proof by cases that for any integer n, the number 3n2 +
n − 16 is even.
√
4. (-) Prove by contradiction that the number 19 is irrational.
5. Suppose that a, b are a pair of real numbers with the following
property: if x is any number greater than b, then x must also be
greater than a. Prove by contradiction that a ≤ b.
6. (+) The game of generalized n-tic-tac-toe is played as follows:
on a n × n grid, two players X and O take turns placing their re-
spective symbols x, o into cells of the grid. No cell can be repeated.
The game ends whenever any player gets n consecutive copies of
their symbol on the same row /column / diagonal, or when the
grid is completely filled in without any player having any such n
consecutive symbols. (Normal tic-tac-toe is where n = 3.)
Prove that there is no strategy in generalized tic-tac-toe where the
second player to move is guaranteed to win.
7. A queen in the game of chess is a piece, shaped like . In
the game of chess, when moved, a queen (when placed in a given
cell in a chessboard) can go to any cell within the same row, any
cell within the same column, or any cell along the two diagonals
through the cell that it starts from. We illustrate this in the mar-
gins.
The n-queens problem is the following task: Take a n × n chess-
board. Can you place n distinct queens on this chessboard, so
that no queen can capture any other (i.e. so that there is no way
to move any one queen into a cell currently occupied by another
queen?)
(a) Prove by cases that there are no solutions to the 3-queens
problem.
(b) Prove by construction that there is a solution to the 4-queens
problem.
(c) Prove by construction that there is a solution to the 8-queens
problem.
8. Prove or disprove the following claim: if G is a graph in which the
degree of every vertex is 3, then G cannot be bipartite.
9. Prove or disprove the following claim: if G is a graph in which the
degree of every vertex is at least 2, then G is connected.

121

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

10. Consider the following two-player game: starting with the single
number 123, two players alternately subtract numbers from the set
{1, 2, 3} from this value. The player who first gets this sum to 0
wins.
If you want to win this game, should you go first or second? Prove
that your chosen player has a winning strategy. (Hint: try induc-
tion!)
11. Take an equilateral triangle with side length 2n . Divide it up into
side-length 1 equilateral triangles, and delete the top triangle. Call
this shape Tn :

Take three side-length 1 equilateral triangles. Join them together

to form the following tile: Prove that you can tile Tn with
tiles, for every n ∈ N.
12. Consider the following inductive “proof:”
Claim: If G is a graph containing at least 3 vertices, and every
vertex in G has degree at least 2, then G contains a C3 subgraph
.
Proof. We proceed by induction on the number of vertices in G.
To start: we assume that our claim holds for all graphs on up to
n vertices, and seek to prove that it holds for all graphs on n + 1
vertices as well.
To do this: for any n > 3, take any graph G on n vertices in
which every vertex has degree at least 2. Add a new vertex v to
this graph, and connect v to at least two other vertices in G; this
gives us a new graph G′ on n + 1 vertices.
We know by our inductive assumption that G itself contains a
C3 subgraph. Therefore this new graph G′ on n + 1 vertices also
contains a C3 subgraph! This is what we wanted to prove, and
thus finishes our inductive proof.
Find every logical flaw in this proof. Explain why the flaws you
have found are indeed mistakes. (Hint: there are at least two flaws
here!)
13. Consider the following solitaire game:

The picture above contains three circles drawn in the plane. In

each of the bounded regions formed by the intersections of these
circles, we’ve placed a coin, which is white on one side and black
on the other. All of the coins start with their black side up.
The moves you’re allowed to perform in this game are the following:
❼ You can at any time flip all of the coins within any circle.
❼ Alternately, you can at any time take any circle and flip all
of its white coins over to black.

122

Downloaded by Raven Ebony (ravenebony133@gmail.com)

 lOMoARcPSD|54546769

Can you ever reach the following configuration? Prove your claim.

123

Downloaded by Raven Ebony (ravenebony133@gmail.com)