Professional Documents
Culture Documents
“To keep your secret is wisdom; but to expect others to keep it is folly.”
Samuel Johnson
Codes have been used by the military to keep secrets from the enemy for thousands
of years. In the information age, when many people use the internet for banking and
shopping, they do not want the information they enter to be available to unauthorised C Student could
people so such information is encrypted on secure websites.
have related this
to personal interest.
This report looks at some of the codes which have been used over time and how
mathematics is used in making and breaking them.
A No aim or
The Caesar Shift Cipher rationale in the
introduction
The first documented use of codes for military purposes was by Julius Caesar (100-
44BC). The type of code he used is commonly known as a Caesar shift. In the
following table the middle row gives plain text and the bottom row gives the
corresponding cipher text for a Caesar shift of 2 places. It is conventional to write
uncoded plain text using lower case letters and coded text using upper case.
Position 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Plain 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
Code C D E F G H I J K L M N O P Q R S T U V W X Y Z A B
This code can be represented using the mapping x → x + 2 where x stands for the
position of the letter in the alphabet. So k, with postion number 10, becomes M with
position number 12. Towards the end of the alphabet, y would go to the letter with
position number 24 + 2 = 26 but 25 is the maximum position number as there are
only 26 letters in the alphabet and we used 0 for a. 26 is subtracted to give 0. This is
called addition modulo 26.
Positions 0 to 25 have been used (rather than 1 to 26) because another way to think
of modular arithmetic is as a remainder from division. 26 is equivalent to 0 modulo 26
because 26 ÷ 26 = 1 rem 0 .
Using this code, the second quote at the start of this report would be encoded as
follows:
C Student's
secretsaremadetobefoundoutwithtime own example
UGETGVUCTGOCFGVQDGHQWPFQWVYKVJVKOG with connection
to the topic
Spaces between words have been removed to make it more difficult to break the
code but if someone found this message and knew that it was a Caesar shift code it
would not take long to break. There are only 25 possible Caesar shifts; 26 if we
include the one that does not move the letters at all but that one is not very secret!
It is possible to write out all 25 possibilities using nothing more complicated than pen
and paper and once you have a message that makes sense you know you have
decoded it. Using a computer to do all 25 possible decodings would make the
process of decoding very quick.
Substitution ciphers
In a substitution cipher each letter is replaced by another letter. A Caesar shift is one
type of substitution cipher but the pattern of substitutions in a Caesar shift makes it
easy to break. A general subsitution cipher can be made by rearranging the 26 letters
of the alphabet and writing this random permutation below the ordinary alphabet. B Good use of
There would be 26! ≈ 4 ×1026 possible codes. It would take a long time to go through approximation notation
all the possible decodings to find the right one! However, some of the possible
substitution codes are too easy to break; the one where each letter is encoded by the
same letter is no use but nor are those where a large number of letters do not
change.
Permutations where nothing stays in its original place are called derangements. The
number of derangements of 26 letters is the subfactorial of 26, denoted !26. To see
how to work this out, consider a much shorter alphabet of 4 letters. The total number
of permutations is 4! = 24. These are shown below.
The 9 derangements are shown in bold, with a box round them. To calculate the
number of derangements, start with the total number of permutations and take away
the ones which have at least one letter in its original position.
There are 6 arrangements with a in its original position, these are on a grey
background. This is because, with a in its original position, there are 3 other letters
which can be arranged in 3! = 6 ways. Likewise, there are 3! permutations with b in its
original place and the same for c and d. However, subtracting 4 × 3! will be too much
as arrangements with both a and b in their original places have been counted twice.
B Mathematical
So 4!− 4 × 3! (1.1) is too small. explanation is
well presented.
If a and b are both in their places, there are 2 letters left which can be arranged in 2!
ways. The same is true for other pairs of letters. There are 4C2 pairs of letters and so
4
we need to add on C2 × 2! to (1.1) to give
However, arrangements which have 3 letters in the same places have been counted
more than once. In fact, it is not possible for 3 letters to be in the same place and the
4th letter to be in a different place so it is only the one arrangement with all four letters
in the same place that has been counted too many times. abcd has been counted 4
times but it should only be counted once.
C Good demonstration
4! 4! 4! 4! 1 1 1 1
!4 =4!− + − + =4!1 − + − + (1.4) of learning and describing
1! 2! 3! 4! 1! 2! 3! 4! unfamiliar mathematics
In general
B Good mathematical
1 1 1 1 definition of terms
n n !1 − + − + ....... + (−1) n (1.5)
!=
1! 2! 3! n ! E Work here is beyond
syllabus requirements.
The Maclaurin series for ex is
x 2 x3 xr
e x =1 + x + + + ......... + + .........
2! 3! r!
1 1 1 n 1 E Good mathematical
so 1 − + − + ...... + (−1) ≈ e −1
1! 2! 3! n! understanding
demonstrated here
n!
!n ≈
e
n!
In fact, rounding gives !n so the number of derangements of the alphabet is
e
26!
!26 ≈ ≈ 1.5 ×1026 . This is still far more possibilities than could be run through in a
e
reasonable time but substitution codes can be broken quickly and easily.
In English, some letters are used more often than others. The following graph shows
how often the different letters are used.
% frequency in English
14
12
10
8
6
4
2
0
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
If a substitution code is used then the letter that has been used to encode e will be
the most frequent letter. Using computers, counting frequencies of letters is quick and
accurate so substitution codes are not secure. The earliest known description of
frequency analysis for breaking codes dates from the 9th century and is by the Arabic
scholar Al-Kindi. In addition to looking at frequencies of single letters, the frequencies C Opportunity for
of pairs of letters can be compared with commonly occuring pairs such as th, er, on
and so on.
student to demonstrate
personal engagement
with an example
The Vigenère Cipher
In the 16th century Blaise de Vigenère built on the work of others to develop a cipher
which was not vulnerable to frequency analysis but which was simple to use. It
makes use of Caesar shifts but does not use the same shift all the time. To use the
cipher, a keyword is needed. The person doing the coding and the person receiving
the message need to know the keyword but that is all they need to remember. By
contrast, for a substitiution cipher which is not a Caesar shift, the whole table for
coding needs to be either remembered or written down.
The method is illustrated using the keyword STELLA. Caesar shift alphabets which
start with S, T, E, L and A will be used.
Plain 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
Code S T U V W X Y Z A B C D E F G H I J K L M N O P Q R
Code T U V W X Y Z A B C D E F G H I J K L M N O P Q R S
Code E F G H I J K L M N O P Q R S T U V W X Y Z A B C D
Code L M N O P Q R S T U V W X Y Z A B C D E F G H I J K
Code 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
To code the initial quote from this report, the keyword is written repeatedly above the
text to be coded:
LHOPPPQHYCDEUKIETSOBWOZMTNXEZEPIINEOLAICDTGDIPAILBWQZLDR
Charles Babbage, who developed an early computer, realised that to break the
Vigenère code the first thing to do was to decide the length of the keyword.
Repeated sets of letters in the encoded message are shown in red. They are likely to
come from the same letters in the original message being encoded in the same way
and, in this case, it is likely that the keyword is 3 letters long because the distance
between repeats is a multiple of 3.
123123123123123123123123123123123
JERFIWFLXFIWFLXVHXEAMCNWVHXHIWFLX
If the code word has 3 letters then all the letters numbered 1 above have been coded
with the same Caesar shift. The same applies for letters numbered 2 and letters
numbered 3. Frequency analysis can now be used separately for each set of letters
with the same number to break the code. There is a useful computerised tool for
breaking Vigenère codes at
http://www.simonsingh.net/The_Black_Chamber/cracking_tool.html
To make adaptations of the Vigenère cipher more secure, a longer keyword needs to
be used and it should not be a real word so that guesswork cannot be used in working
it out. This makes it much harder to use the Vigenère cipher unless the process can
be automated. In 1918, a German inventor, Arthur Scherbius, invented the Enigma
machine. It was further developed and used by the German army during the second
world war.
The Enigma machine had a keyboard. Operators typed in a message and the
machine encoded it. The basic idea is that a rotor is used to move on and encode the
next letter using a different substitution alphabet. With one rotor and 26 positions, this
would give 26 possible initial starting postions and so 26 possible codes. However 3
rotors were used and each was set to a different starting position every day giving
26 × 26 × 26 = 17 576 possible codes.
Operators also chose 3 rotors from 5 possible rotors, ie 5P3 = 60 ways of positioning
the rotors so this now gives 60×17 576 = 1 054 560 possible codes.
The number of ways of choosing the first pair then the second pair and so on is
However, the order of selecting the pairs does not matter so each set of 10 pairs has
been counted too many times; the answer in (1.6) needs to be divided by 10!
26!
≈ 1.5 ×1014
6!× 210 ×10!
Taken together with the options for the rotors, this gives a total of
1.5 ×1014 ×1 054 560 ≈ 1.59 ×1020 possible codes. With no possibility of using
frequency analysis, the Enigma code was considered to be unbreakable.
Operators of Enigma machines were issued with books which gave them the settings
for the day. They were told which letters to connect on the keyboard, which rotors to
use and the initial setting for the rotors. If the code breakers had an Enigma machine
and could find out the settings for the day, they would be able to decipher all the
messages for that day.
Messages would start with the Enigma machine in this position and then the setting of
the rotors for the message would be transmitted; the rotors were then adjusted before
transmitting the rest of the message. This meant that even if the code breakers had
all the messages for the day, each one would use a different code (apart from the first
3 letters of each message). This did not give the code breakers very much to go on.
To break the Enigma code, the code breakers used a copy of an Enigma machine and
constructed machines, called bombes, which could work through a large number of
possible codes. The machines could not work through anything like all possible codes
in one day so further information and deduction was necessary.
The Enigma machine never encoded a letter as itself. This helped with the decoding
of some messages. One day a message came through which had no letter L in it.
The operator had been sending a test message and had just pressed the letter L
every time so it was the only letter that did not feature in the encoded message.
Knowing that messages which were sent at particular times of day were about the
weather enabled the code breakers to guess part of the message. Knowing that a
letter could not be encoded as itself, helped them narrow down the possibilities for
which of the possible codes was being used that day and the bombes could then
search for that day’s settings.
Suppose the word “weather” is known to be part of the plain text for the following
coded message. The word can be tried in different positions, as follows
The third of these possible positions cannot be the place where the word “weather”
was in the text because w cannot code to W and h cannot code to H.
In the 1970s, Rivest, Shamir and Adleman came up with a way to encode messages
which did not rely on the key being kept secret. The system they invented is
sometimes known as RSA. The system encodes numbers rather than letters but it is
possible to have a system for making letters or words into numbers. The table below
shows the steps in the process by using a simple example.
Enciphering
M C
1 13 (mod 22) = 1
2 23 (mod 22) = 8 (mod 22) = 8
3 33 (mod 22) = 27 (mod 22) = 5
4 43 (mod 22) = 64 (mod 22) = 20
Remember from the section on Caesar ciphers that to find 64 (mod 22) you can find
the remainder when 64 is divided by 22. However, working with powers of numbers
can lead to calculations with very large numbers. Fortunately mathematics can help.
Suppose you want to work out ab (modulo 15). Suppose a = r (modulo 15) and
b = R (modulo 15) .
Deciphering
M C
1 17 (mod 22) = 1
8 87 (mod 22) = (82 mod22) × (82 mod22) × (82 mod22) × (8 mod 22) (mod 22)
= ((64 mod 22)3 × 8) mod 22
= (203 × 8) mod 22
= 64 000 mod 22
=2
5 57 (mod 22) = (52 mod22) × (52 mod22) × (52 mod22) × (5 mod 22) (mod 22)
= ((33 mod 22) × 5) mod 22
= ((27 mod 22) × 5) mod 22
= (5 × 5) mod 22
=3
To keep writing “mod 22” is making this look more complicated than necessary so, for
deciphering 20, a different notation has been used. ≡ means “has the same
remainder when divided by 22”.
The tables above demonstrate that deciphering takes you back to the original number
for the particular examples tested. This will always work. The proof depends on
Euler’s theorem, which states that
M φ(m) ≡ 1mod m
φ(m) is the number of integers smaller than m and having no common factor with with
m. If m = pq where p and q are prime, ϕ(m) = ( p −1)(q −1) and this is what we called
A in the description of the RSA algorithm.
Using ≡ to mean “has the same remainder when divided by m” and remembering the
formula for deciphering is M = C D (modulo m)
C≡ME
C D ≡ (M E )D
≡ M ED = M ED−1 × M
m = pq is known; to break the code needs the individual primes, p and q to be known.
m = 22 does not give a secure code as it is obvious that the primes were 2 and 11.
The primes need to be much larger for the code to be secure.
655 427 is the product of two primes. To find the primes a spreadsheet could be
used. This is the start of a spreadsheet showing one way to find the primes:
Rather than just trying to divide by primes, all integers are tested so 1 does go into
655 427 but this does not help. The prime factors are found further down:
Constructing the spreadsheet and finding the factors took about 1 minute so 655 427
would not give a secure code.
C Timed self
doing spreadsheet
The factorising can be speeded up by using Fermat’s factorising method. This
depends on the following result. D Reflection on
2 2 significance of
p+q p−q
m pq
= = − personal timing
2 2
Multiplying out the brackets on the right hand side gives
p 2 + q 2 + 2 pq p 2 + q 2 − 2 pq
− pq
=
4 4
p+q p−q
If p and q are large prime numbers, they will both be odd so and are
2 2
whole numbers.
We are trying to factorise m = 655 427 . 655 427 = 809.58... so the smallest possible
value of x is 810.
For each possible value of x find the value of x2 − m . When this is a square number,
y2 , the factors are ( x − y )( x + y ) . Doing the calculations on a spreadsheet gives the
following:
The factors are found after testing 156 numbers whereas the previous method took
439 numbers. It is so quick to do for a 6 digit value of m that it makes little difference
but it shows that computing speed and efficient methods of factorising can enable the
RSA code to be broken unless the prime numbers used are very large. The current
record for successfully factoring a number is 200 digits.
“In the eighties it was generally held that prime numbers of fifty odd digits would
suffice. However, developments went much faster than foreseen, and it is a
precarious matter to venture upon quantitative forecasts in this field. When Rivest
challenged the world in 1977 to factor RSA-129, a 129 digit number (from a special
list), he estimated that on the basis of contemporary computational methods and A No conclusion
computer systems this would take about 1016 years of computing time. Seventeen
years later it took only eight months in a world-wide cooperative effort to do the job.”
D Many opportunities for
http://www.cwi.nl/en/RSA reflection: What about
the future? Will faster
Bibliography
computers eventually
Cryptology timeline www.math.cornell.edu/~morris/135/timeline.html
eliminate the ability to
encode messages? What
An introduction to cryptography other implications could
http://www.math.sunysb.edu/~scott/papers/MSTP/crypto/crypto.html there be?
Derangements http://www.mathlab.mtu.edu/~eewestlu/ma3210_lecture19.pdf
http://en.wikipedia.org/wiki/Cryptanalysis_of_the_Enigma
http://www.cwi.nl/en/RSA
M AT H S C O U R S E WO R K
Proving Euler’s Totient Theorem
TA B L E O F C O N T E N T S
Maths Coursework 3
Introduction 3
The Theorem 3
Applications 7
Conclusion 7
M AT H S C O U R S E WO R K
Proving Euler’s Totient Theorem A Aim
Introduction
My interest in Euler as a mathematician was first sparked when, on completing a listener
crossword, the hidden message “Read Euler, he is the master of us all” was revealed, so
when I saw the inclusion of his name on the list of prompt words there was really no
option but to go for him. Euler was a mathematician in the 18th century and is responsible
for the first proofs of many great many number of conjectures and problems. In number
theory alone his accomplishments include proving the two square theorem and Fermat’s
little theorem as well as doing a great deal of work that later led to the first proof of the
four square theorem. His achievement that I am going to focus on though is less well
known, it is a generalisation of Fermat’s little theorem that has come to be known as
A Introduction
Euler’s totient theorem.
and rationale
The Theorem
Euler’s totient theorem1 states that for relatively prime a and n:
B Class knew
about modular
aΦn ≡ 1 (mod n) arithmetic,
so this didn't
need definition.
Where Φn is Euler’s totient function
1
http://en.wikipedia.org/wiki/Euler's_theorem
2
http://mathworld.wolfram.com/TotientFunction.html
Proving Euler’s Totient Theorem
B Small n in
Below is a table of the totients of the numbers up to 20. text. Condoned
N Φ N
2 1
3 2
4 2
5 4
6 2
7 6
8 4
9 6
10 4
11 10
12 4
13 12
14 6
15 8
16 8
17 16
18 6
19 18
20 8
3 http://mathworld.wolfram.com/FermatsLittleTheorem.html
Proving Euler’s Totient Theorem
ap ≡ a (mod p)
or
ap-1 ≡ 1 (mod p)
It is immediately apparent that this fits in with Euler’s totient theorem for primes p, as we
have seen Φp, where p is a prime, is always p-1.
Take two numbers a and p which are relatively prime, and where p itself is prime.
Consider the set of the multiples of a { a, 2a, 3a, 4a, 5a ..... (p-1)a }
We can take out a factor of ap-1 from the left hand side
ap-1 ≡ 1 (mod p)
or
ap ≡ a (mod p)
QED.
Lemma 1: Each number in the first set must be congruent to one and only one number in
the second and each number in the second set must be congruent to one and only one
number in the first. This may not be obvious at first but can be proved through three
logical steps.
(1) Each number in the first set must be congruent to one of the elements in the second as
all possible congruences save 0 are present, none will be congruent to 0 as a and p are
relatively prime.
(2) A number cannot be congruent to two numbers in the second set as a number can only
be congruent to numbers which differ by a multiple of p, as all elements of the second
set are smaller than p a number can only be congruent to one of them.
(3) No two numbers in the first set, call them ba and ca, can be congruent to the same
number in the second. This would indicate that the two numbers were congruent to
each other ba ≡ ca (mod p) which would indicate that b ≡ c (mod p) which is not true
as they are both different and less than p itself.
Therefore, through these three steps Lemma 1 is proven.
This set will have Φn elements (Φn is defined as the number of numbers relatively prime to n)
Consider the set aN, where each element is the product of a and an element of N { a, an1,
an2... anΦn }
Each element in set aN will be congruent to an element in set N (mod n), this is follows by
the same argument as in lemma 1 and so the two sets will be congruent to each other
Therefore { a x an1 x an2 x ... x anΦn } ≡ { 1 x n1 x n2 x ... x nΦn } (mod n)
4
http://planetmath.org/?op=getobj&from=objects&id=335
By taking out a factor of aΦn from the left hand side we get
aΦn { 1 x n1 x n2 x ... x nΦn } ≡ { 1 x n1 x n2 x ... x nΦn } (mod n)
If we then divide through by { 1 x n1 x n2 x ... x nΦn } which is valid as all elements are
relatively prime to n we get
A Complete.
aΦn ≡ 1 (mod n) Achieves aim.
Concise!
QED.
Applications
Unlike some of Euler’s other work in number theory such as his proof of the two square
theorem Euler’s totient theorem has very real uses and applications in the world and like
much of number theory those uses are almost exclusively in the world of cryptography
and cryptanalysis. Both Fermat’s little theorem and Euler’s totient theorem are used in the
encryption and decryption of data, specifically in the RSA encryption system5, whose
protection revolves around large prime numbers raised to large powers being difficult to
factorise.
D Reflection. Linking
mathematical ideas
Conclusion
This theorem may not be Euler’s most elegant piece of mathematics (my personal
favourite is his proof of the two square theorem by infinite descent) or at the time seemed
like his most important piece of work at the time but this, in number theory at least, is
probably his most useful piece of mathematics to the world today.
This proof has given me a chance to link up some of the work I have done in the largely
separate discrete mathematics and sets relations and groups options. These two options
appear to me to be the purest sections of mathematics that I have studied but are for
whatever reason seldom linked in class, this project has allowed me to explore the links
between them and use knowledge from one in relation to the other, broadening my view
of maths.
A Conclusion
C Personally
involved
5
http://www.muppetlabs.com/~breadbox/txt/rsa.html#7
When the game is started, the player is presented a grid of squares. The size of the grid
varies from the difficulty level chosen by the player. The larger the grid the higher the
difficulty level and skills required. The number of mines also varies from difficulty level;
usually in the beginner level there are 10 mines in a 9x9 grid. The game progresses by
the player’s clicks on a square. The square clicked will either a mine or reveal a square
with a digit. If the square revealed turns out to be a mine, the game is over. The game
progresses as the player clicks on squares without mines and reveals digits with the
highest digit being 8. The digit indicates the number of adjacent squares that contain
mines. Using this information, players can deduce the certainties of other squares on
whether they are mine-filled or free and continue on with the game by either flagging a
square or proceed to clearing other squares. For example, if you left-clicked a square,
and an 8 digit appeared, then you would know that the square is surrounded by 8
mines, all 8 of its adjacent squares. You would then have to go elsewhere to test for
non-mine squares. The player is declared the winner when either all mine’s are flagged
or you cleared all the squares that does not contain mines.
The game can be played for two basic purposes; play for speed or play to just win. To
play for speed simply try to clear the minefield as fast as you can without detonating a
mine and to play to win; simply clear the minefield. Each contradicts each other as
focusing for speed decreases your chances of winning and focusing on just winning
decreases your speed. D Some reflection
but very simplistic
Controls
For a player to clear a mine simply left click a square.
Players can then ‘’flag’’ a square by right clicking on a square they believe to inhabit a
mine thus eliminating it.
Right-clicking again on the flag symbol of a square will change the flag into a question-
mark graphic to indicate the possibility of the square may or may not be a mine.
Right clicking once more on a question mark graphic transforms the square back to its
original form.
The player can press down both left and right clicks at the same time on a square to A Lack of
show the square’s adjacent squares. coherency
through limited
explanation
Basic Strategy
The key of minesweeper is logical deduction from the digits of the squares. Simple
logical strategies cases such as if there is a 3 in a corner, all the squares adjacent to is
100% mines. Another example would be if an uncovered square is with a digit 1, and
there is only one covered square touching it, that square must be a mine. In general,
clearing a board with speed requires a form of pattern recognition and the brain to B, E No mathematical
quickly identify as well as make connections in identifying the clues which state that strategy outlined.
there’s a 100% probability in a square being a mine. Missed opportunity
for presentation using
mathematical language
Element of Luck
The first few moves of minesweeper without a doubt require luck. The player can only
give wild guesses because of the fact the player have either no digits or anything to refer
to. The chance of a player clicking on a mine on the first move is 0% as the mines
generate randomly after your first click. On your second click however, on a beginner’s
A No explanation
difficulty would be around 12% and 38% on expert difficulty. Because the placement of
the mines in every game is random, the choice of your starting move can’t be based on of why these
anything but luck. Even if you start in the corner, you may either be lucky enough to get values were used
a 3 or be unlucky and get a 1 which really doesn’t help a whole lot for your first move.
Experts do however recommend beginning the middle area of the grid opposed to the
corners.
It is only after the first few moves of minesweeper is complete when the player can truly
deduct and rely on logic of where the mines are for the rest of the game.
Minesweeper also have the factor of sheer luck in cases where the player has no choice
but to take a guess in clearing a square as the digits of the squares show all the
surrounding squares each having the same probability of possessing a mine. C Opportunity for
an example here
There are times when a grid generated can only be solved with the element of guessing.
Refer to Figure 3.
Figure 3
The player must guess whether or is the mine. The probability of A and B being
B Incorrect
mines are equal, 50/50.
terminology used
Another scenario that requires guessing is when an un-clicked square is completely or
partially surrounded by mines. The player has no numbers touching the un-clicked
square thus the player has again no information about the probability of whether or not
the square will be a mine.
C, D An
Strategies when facing this scenario that will allow the player to avoid simple guesses appropriate
are to play the rest of the game and ignore the square. In the game when the number of logical
un-clicked and un-flagged squares equals the number of mines left unmarked then all implication
the remaining spaces are mines. Some versions of the game automatically flag these
spaces once all the other squares in the game have been clicked as well as reveal them
once all the flags have been placed.
If the number of un-clicked squares adjacent to the number of square equal to the
number on that square, all the un-clicked squares are definitely all mines.
For any square with a digit, if the digit of mines you've found adjacent to that square is
equal to the number of the square, all other squares adjacent to that numbered square
must be safe. If you know the square to the right of a 1 is a mine, then you can deduce
that all the other squares next to that 1 do not contain mines and you can clear them to
obtain more digits to progress in the game. B, E Student
work is descriptive
rather than
mathematical.
Example case Example case
1 2
C Use of
a and b are personal examples
safe to open,
as the 3 is
already a and b must
satisfied with be mines; the
3 adjacent only squares
mines. that can
provide those
demanded by
the 3 are a
and b.
Multiple square strategies are a great for players who want to win in beginner
difficulties quickly, and a must in the higher difficulties. For these complex puzzles, the
player would need to consider more than one square a time
If you have two adjacent numbers, the difference between those numbers is equal to
the difference in the number of mines for the 3 squares adjacent to each that are not
adjacent to the other number. For example: if these numbers differ by 3, all of the
adjacent squares to the higher number not shared by the other are mines, and all the
opposite ones are safe. C Opportunity for
an example
Sometimes it can be known that there are a certain number of mines in a certain
number of squares (without necessarily knowing which the mines are and which are
safe) and you can often utilize this information to find out information about other
squares.
One would recognize the variables in the game are the unopened squares, and the
constraints are the adjacent squares that are opened.
The algorithm consists of trying every combination of mines that satisfies all the
numbers in the adjacent squares, and making a conclusion from there.
In expert difficulty, this would be a time-consuming process for even the most powerful
computer, but minesweeper players who are experienced at the expert level would be
able to quickly see which squares need this procedure, and where one might expect it to
succeed.
Analysis
One of the outstanding problems in logic and computer science is determining whether A Lack of
questions exist whose answer can be quickly checked by an artificial intelligence, coherency
however requires a much longer time to solve without knowing the answer. So far no
one has proved that any of them really does require a long time to solve.
Conclusion
In the school’s courses, students play a few games and attempts to deduct and evaluate
the various positions the squares that definitely possess mines and the squares
definitely are safe. The exercise is proven to give students a sense of given the rules, the
A Implicit
evidence and conjecture, then prove or test. Students who become ‘’masters’’ of this
game learn truly useful reasoning, logic and deductive techniques such as proof by aim here but
contradiction or the role of counter, all which can be applied to high levels of incorrectly placed
mathematics such as proving theorems.
It is indeed interesting how gaming can contribute “practical” things to people's lives
and how this simple game such as minesweeper might just be the thing that can a solve D Superficial
today’s mathematical mystery. It truly is no surprise, with the amount of logical reflection
deductions, time and brainpower required to play this game.
Bibliography
Loy, Jim. "Minesweeper Strategy." Minesweeper Strategy. 1996. Web. 6 Jan. 2010.
<http://www.jimloy.com/games/mine.htm>.
Lipman, Lisa. "Games may help solve math mysteries." MSNBC. Associated Press, 2000.
Web. 6 Jan. 2010.
<http://www.math.hawaii.edu/lab/Game%20may%20help%20solve%20math%20myste
ries.htm>.
Introduction A Rationale
included in
From the stimulus word “Harmony”, I chose to look at the transmission of sound introduction
waves in music. As a keen musician myself, I was curious to understand more about how
electronic instruments emit sound. Sound travels as a transverse wave, so takes the same
basic shape as a mathematical sine wave. Electronic instruments use these waves to
produce sounds of different frequencies, volume and timbre.
Oscillators 1 are used in most electronic instruments to produce sound waves. These
sound waves are sine waves, with different frequencies, changing the pitch of the note,
and different amplitudes, varying the volume of the note. The timbre of the sound is
changed once the wave has left the oscillator. It travels through various filters which
change the type of sound. This allows for a great range of sounds, that can imitate C Personal
different instruments. It is this technology that is used in electronic keyboards. interest
in subject
I decided to focus on the changing of frequencies to produce different notes. By area, but it
modelling chords, and series of notes using sine waves, I can demonstrate how the sound is limited
waves travel in the air, and how the different frequency waves relate to each other in
different chords. I will look specifically at the relationships within each type of chord,
and whether there is any difference between chords that are perceived to be in harmony A Aim not really
and those that are perceived as discordant. clear compared
with what is done
Investigation A Undefined
term and not clear
I started by working out how to model a single note. I decided to use “Middle” A, as for non-musicians
it has an exact frequency of 440 Hz 2, and is the note that Western orchestras tune to. I
wanted to make the sound wave of this note equal to the sine wave, which meant making
220Hz equivalent toππ. This produces the following graph:
Typing error
4 y
−2π −π π 2π
−2
1
http://en.wikipedia.org/wiki/Synthesizer
2
http://www.phy.mtu.edu/~suits/notefreqs.html
A Lack of clarity
1. The Major Chord
for non-musicians
Using an A major chord, I can look at the relationship between the notes in a major
chord. The following graph demonstrates how this chord travels through air:
y
−π −π/2 π/2 π
−1
−2
A No mathematical explanation
of how these are determined
Equation 1: y=sinx
Equation 2: y=sin(554.37/440x)
B Erroneous notation. This means
Equation 3: y=sin(659.26/440x)
Equation 4: y=sin2x 554.37 but should be 554.37x .
440x 440
I altered the wavelengths, or frequencies, of the graphs by changing the coefficient of
x. In order to get the most accurate graph, I used the original frequencies of the other
notes 3. The table below shows more information about the notes in this chord.
The frequency ratios in this chord appear to be neatly spaced, with the third gap in
the chord equalling the sum of the first 2 gaps in the chord. The ratio of the gaps is 1:1:2. C Student
This may indicate the start of a pattern. Similarly, it could be that the third gap is double created own
the first or second gap. I will now look at different chords to see if this conjecture holds way of looking
for patterns in
3
the chord.
http://www.phy.mtu.edu/~suits/notefreqs.html
true. The gaps in the ratios do not directly relate to the difference in semitones, as the
semitones do not have equal frequencies-the frequencies of semitones diverge as the
frequencies increase, and the notes get higher in pitch.
Minor chords are also considered to be in harmony. Again using an A chord, I have
modelled a minor chord and looked at its features. These waves look very similar to the
major chord waves, as only one note has changed; the C# has gone down by a semitone.
Therefore the graphs are very similar. The graph for the minor chord looks like this:
y
−1
−2
This graph looks very similar to the major graph, but the actual values, shown in the
following table, are slightly different.
Again, because the E in the middle of the chord has not moved, the sum of the first 2
gaps in the chord is equal to the third gap. The ratio of the gaps is now 2:3:5. However,
to test this theory properly, we need to look at chords that differ from the original major
chord more dramatically.
An inversion chord is the same as a major chord, but the notes are in a slightly
different order. A first inversion chord starting on A will actually be in the key of F
major. The graph of such a chord looks like this:
2 y
−1
−2
Equation 1: y=sinx
Equation 2: y=sin(523.25/440x) Incorrect value
Equation 3: y=sin(698.46/440x) (698.46) compared
Equation 4: y=sin2x with table
The table of values for this chord shows that my original conjecture does not hold
true for all chords. However, the gaps are still in a neat ratio of 1:2:2.
This chord, like the first inversion, is another different rearrangement of the notes in
a major chord.
−1
−2
Interestingly, in this chord, the frequencies of the notes are evenly spaced, as the ratio
of the gaps is 1:1:1.
I will now look at some discordant notes, and see how they compare with the chords
I have already looked at.
This chord is also known as “The Devil’s Chord”, as it is considered to be the most
unpleasant sounding chord in music. As it is only 2 notes, it cannot be modelled in
exactly the same way as the other chords, but can still be graphed:
2 y
−1
−2
Equation 1: y=sinx
Equation 2: y=sin(622.25/440x)
I will now look at the augmented 6th, another discordant chord. This chord however,
has 3 notes, so can be more easily compared to the other chords I have looked at. An
Augmented 6th is also known as a dominant 7th, which was one of the first discordant
chords to be used in music in the West.
2 y
−1
−2
Equation 1: y=sinx
Equation 2: y=sin(523.25/440x)
Equation 3: y=sin(739.99/440x)
Equation 4: y=sin2x
In this chord, the gaps in the ratios are not easily relatable to each other, suggesting
that discordant sets of notes are not as natural as those that sound harmonic.
The intervals known as a “perfect fourth” and “perfect fifth” are, as the name
suggests, considered to be the most harmonic intervals. Interestingly, the graphs of these
intervals show that the notes in the perfect fourth cross at exactly 3π, and the notes in the
perfect fifth cross at 2π.
y
1
0.5
x
0
3π/2 2π 5π/2
−0.5
Equation 1: y=sinx
Equation 2: y=sin(587.33/440x)
0.5
x
0
9π/4 11π/4 13π/4
−0.5
−1
A, D Lack of explanation
Equation 1: y=sinx
and a lost opportunity for
Equation 2: y=sin(659.26/440x)
reflection on how results
fit with conjecture
Conclusion
This evidence would suggest that notes which are traditionally considered to be “in
harmony” are mathematically “neater” than those that are not in harmony. The
mathematical nature of the notes in traditional harmony suggests that the sound waves are
more pleasing to our brain because of the way the notes fit together. The ratios of
frequencies of chords in harmony are more concise than the discordant notes. The sound
B More poor use
of terminology
waves produced by oscillators are in the same form as notes produced by mechanical
instruments, so the same note ultimately sounds the same, or very similar, to our ear. For
this reason, this theory does not only apply to sound coming to electronic instruments, but
every kind of instrument.
Evaluation
Bibliography
Maths Exploration
Newton-Raphson method
A Introduction
includes aim
and rationale
Rationale- For this project I chose to research and analyse the Newton-Raphson method,
where calculus is used to approximate roots. I chose this topic because it looked extremely
interesting and the idea of using calculus to approximate roots, seemed intriguing.
The aim of this exploration is to find out how to use the Newton-Raphson method, and in
what situations this method is used
A, B Lack of
f ( xn )
xn +=
1 xn − , f '( xn ) ≠ 0 mathematical explanation
f '( xn )
of this terminology
This is called the Newton-Raphson formula
The first thing you do when approximating a root is to make an initial estimate in terms of
a positive integer and find the tangent of the function at that point.
B Poor use of subscripts
Let the initial estimate = x₁
throughout exploration
The intercept of the tangent at x₁ with the x-axis will be closer to the root than the initial
estimate. This intercept can be called x₂. E No
explanation of
Calculating x₂ is done by studying the triangle bounded by the x-axis, the line x=x₁, and how student
the tangent to the line at x₁. This only works when the gradient of the function at x₁ is not knows it will
be closer
equal to zero.
An example of a triangle is shown in the following graph (where the square root of 7 is
being estimated):
B Graph could
be larger and
clearer, zooming in
on the pertinent
domain and range
f ( x2 )
x=
3 x2 −
f '( x2 )
x₃ will be even closer to the root than x₂ was, and this procedure can be repeated an
infinite number of times.
f ( xn )
xn +=
1 xn − A Not an
f '( xn )
explanation
Where n= the number of iterations
The more iterations that are carried out the more accurate the answer is. In practice the
iteration is usually only repeated to x₁₀ or less.
For example, if we use this method to approximate the value of root 7 , if the equation
y x 2 − 7 is used.
=
C Good demonstration
The formula for this function is: of applying unfamiliar
mathematics
f ( xn )
xn +=
1 xn −
f '( xn )
xn 2 − 7 E Differentiation
xn +=
1 xn −
2 xn within syllabus
Using the Newton-Raphson formula for this function the following table showing the
value of x for different numbers of iterations:
x₁ 3.0000000000
x₂ 2.6666666667
x₃ 2.6458333333
x₄ 2.6457513123 A Lack of detailed
x₅ 2.6457513111 explanation
x₆ 2.6457513111
x₇ 2.6457513111
x₈ 2.6457513111
x₉ 2.6457513111
x₁₀ 2.6457513111
The table shows that this method will only need a few iterations and you already have the
value of x to 3 decimal places and after the fifth iteration you have the values of x up to 10
decimal places.
A Graph
does not aid
coherence and
adds nothing to
previous graph.
f ( xn )
xn +=
1 xn −
f '( xn )
xn 2 − 3
xn +=
1 xn −
2 xn
Using the Newton-Raphson formula for this function the following table showing the
value of x for different numbers of iterations:
x₁ 2.0000000000
x₂ 1.7500000000
x₃ 1.7321428571
x₄ 1.7320508100
x₅ 1.7320508076
x₆ 1.7320508076
x₇ 1.7320508076
x₈ 1.7320508076
x₉ 1.7320508076
x₁₀ 1.7320508076
The table again shows that this method will only need a few iterations and you already
have the value of x to 3 decimal places and after the fifth iteration you have the values of x
up to 10 decimal places.
Using a spreadsheet, I repeated the same method for all roots x where x ∈ Z + in the
domain 1 ≤ x ≤ 7 , and a table of iterations was produced: B Good use
of notation
1 2 3 4 5 6 7
x₁ 1.0000000000 1.5000000000 2.0000000000 2.0000000000 2.5000000000 2.5000000000 3.0000000000
x₂ 1.0000000000 1.4166666667 1.7500000000 2.0000000000 2.2500000000 2.4500000000 2.6666666667
x₃ 1.0000000000 1.4142156863 1.7321428571 2.0000000000 2.2361111111 2.4494897959 2.6458333333
x₄ 1.0000000000 1.4142135624 1.7320508100 2.0000000000 2.2360679779 2.4494897428 2.6457513123
x₅ 1.0000000000 1.4142135624 1.7320508076 2.0000000000 2.2360679775 2.4494897428 2.6457513111
x₆ 1.0000000000 1.4142135624 1.7320508076 2.0000000000 2.2360679775 2.4494897428 2.6457513111
x₇ 1.0000000000 1.4142135624 1.7320508076 2.0000000000 2.2360679775 2.4494897428 2.6457513111
x₈ 1.0000000000 1.4142135624 1.7320508076 2.0000000000 2.2360679775 2.4494897428 2.6457513111
x₉ 1.0000000000 1.4142135624 1.7320508076 2.0000000000 2.2360679775 2.4494897428 2.6457513111
x₁₀ 1.0000000000 1.4142135624 1.7320508076 2.0000000000 2.2360679775 2.4494897428 2.6457513111
The table shows that after the first few iterations you get a very precise value of x
Although the main function of this method is to approximate roots it can also be used
for non-linear equations. C, E Student could have mentioned
other roots, such as cube roots.
For example the Newton-Raphson method can be used to find when tan(x) = x or cos(x)= x
The Newton Raphson formula can easily be adjusted for these equations:
f ( xn )
xn +=
1 xn − C Student created
f '( xn ) own example, but did
not follow through to
When f ( x)= x − tan x complete application
with graphical analysis
to see how the process
worked for this case.
Becomes:
xn − tan( xn )
xn +=
1 xn −
1 − sec 2 ( xn )
D Opportunity for
And when f ( x)= x − cos x reflection on any
special implications to
non-linear functions
Becomes:
xn − cos ( xn )
xn +=
1 xn −
1 + sin( xn )
Reflection:
This was a very interesting exploration, as at the start I had no idea what the formula meant
and the logic behind it. The idea of using the x-intercept of the gradient to get closer and
closer to the correct value of x was amazing. Up to now, I always had a calculator to find
B Confused
the root or zero of a number. If I didn’t I would just leave it as the 3 or 5 . Although,use of "root"
this method may not be practical to use on a maths test, it has quite a few advantages. This and "zero"
method gives us the chance to approximate roots up to thousands of decimal places. Even
after the first five iterations the value of x is given to 10 decimal places. When finding the
root of x, it is quite easy to work out x₂ in your head, which is very close to the precise
value. Furthermore, this formula is implemented in technology such as autograph, where
Newton-Raphson Iteration is very easy to use and find, as well as by using spreadsheets
where it is simple to implement the formula for any number of iterations. Nevertheless,
it’s easy to find the zero or root by finding the x-intercept, which is much easier than using
the Newton-Raphson method on the calculator or any graphing program. This method
would be useful though in any field involving zeros or roots where a high-level of
precision is needed.
D Superficial
Bibliography: reflection with
no depth
Internet Sources:
"The Newton-Raphson Method ." S.O.S. Math. N.p., n.d. Web. 20 Jun 2010.
<http://www.sosmath.com/calculus/diff/der07/der07.html>.
Image:
Florence Nightingale headed a group of 38 nurses who went to clean up the hospitals for
the British soldiers in the Crimea in 1854. She found that most of the deaths were due to
diseases which could be prevented by basic hygiene, such as typhus and cholera. Her
improvements were simple but they had an enormous effect:
“She and her nurses washed and bathed the soldiers, laundered their linens, gave
them clean beds to lie in, and fed them”3.
When she returned to Britain she made a detailed report to the Government setting out
what conditions were like and what needed to be done to reduce deaths in the hospitals.
Nothing was done, so she tried again, making another statistical report and included in it
three new statistical diagrams to make data collated by William Farr more accessible to
people who could not get their minds around tables of figures. These were her polar area
diagrams or rose diagrams, sometimes also known as ‘coxcombs’. The first showed how
many men had died over the two years 1854-5, the second showed what proportions of
men had died from wounds in battle, from disease and from other causes, the third
showed how the number of deaths had decreased once “sanitary improvements”4 had
been introduced.
I decided I would try to recreate the second of these diagrams which is the most
complicated and the most shocking. It is called “Diagram of the causes of mortality in the
army in the east”. A copy of it is below:
Figure 1
The basic ideas are very simple. The blue area represents deaths due to disease, the red
area represents death due to wounds in battle and the black area represents death due to
other causes.
I tried to find a copy of the data which this diagram represented, but I had no luck, so I
C Student
decided to make sure I understood exactly how the diagram was made and to make my
own version of some data which I did have to hand. Once I tried to understand the
demonstrates
diagram in detail I found there were some problems. personal interest
in topic.
The First Problem
D Good critical
reflection on
problems related
to objective
this area or this area
In other words, were the colours separate, or overlapping? The articles I read didn’t make
it clear. O’Connor says that “The area of each coloured wedge, measured from the centre
A Good research
as a common point, is in proportion to the statistic it represents”5, which makes it seem
demonstrated
that all colours are wedged shaped, or sectors, so the colours overlap. However, Lienhard
commented that in the November 1854 section “battle deaths take up a very small portion
of each slice”6, which makes it sound as though the slice has three separate portions, and
Brasseur says that “she also divided the areas within each of the wedges to show which
portion of the mortality data for that month could be allotted to each cause of death”4. I
decided to construct polar area diagrams for a set of data with the colours separate and
with the colours overlapping to see if putting theory into practice would make it clearer to
C Created
me. The data I used was taken from the IB grade distribution statistics for the past 15
personal example
years at my own institution. I used the numbers taking Higher Mathematics, Standard
Mathematics and Mathematical Studies to be represented by my three colours. I took the
old Mathematical Methods course to be the same as Standard Mathematics.
2π
To fit 15 sectors into the circle I needed each arc to subtend an angle of radians at
15 E Area of sector
1 2π 2 π 2 using radians is
the centre. The area of each sector would then be A = × r =r where r is the
2 15 15 in syllabus and
radius of the sector. Since the area needs to be proportional to the statistic, I needed to is relevant to
15A exploration.
find the radius, so I used r = and just used a scale which would allow me to draw
π
a reasonable sized diagram. To create a polar area diagram with overlapping sectors I just
used this formula on each of the numbers of students taking the various options.
I then used a geometric program (GeoGebra) to draw the sectors all with a common
2π
centre, each with an angle of radians, and with the radii as given in the table. I drew
15
the Higher sectors first with the Studies on top of these, and the Standard on top of these.
This was the result:
Figure 2
Polar area diagram to show students
taking Mathematics at one school
(colours overlapping)
Blue represents the number of students taking
Higher Maths.
Brown represents the number of students
taking Mathematical Studies.
Green represents the number of students
taking Standard Maths.
The colours are not solid, so where colours overlap there is a different colour. The blue
overlapping the brown makes a pink here, and the green overlapping the blue makes a
darker green. In 2003 and in 2004 there were an equal number of students taking Higher
and Standard so three separate colours cannot be seen on the diagram.
Next I worked out the radii needed if the colours were not to overlap. For this I used
15 A1 15 ( A1 + A2 )
cumulative areas to work out the radii. R1 = , R2 = and
π π B Good use
15 ( A1 + A2 + A3) of notation
R3 = .
π
Numbers taking Mathematics year on year
Numbers (A) Radii
Higher (A1) Studies (A2) Standard (A3) R1 R2 R3
1995 1 24 0 2.2 10.9 10.9
1996 4 15 0 4.4 9.5 9.5
1997 8 10 0 6.2 9.3 9.3
1998 6 31 0 5.4 13.3 13.3
1999 9 17 0 6.6 11.1 11.1
2000 10 20 0 6.9 12.0 12.0
2001 4 31 1 4.4 12.9 13.1
2002 5 21 2 4.9 11.1 11.6
2003 4 15 4 4.4 9.5 10.5
2004 5 29 5 4.9 12.7 13.6
2005 1 28 0 2.2 11.8 11.8
2006 3 16 2 3.8 9.5 10.0
2007 8 13 0 6.2 10.0 10.0
2008 11 29 14 7.2 13.8 16.1
2009 10 23 15 6.9 12.6 15.1
This gave a diagram with Higher numbers at the centre and Standard numbers at the
edge, like this:
Figure 3
Polar area diagram to show students
taking Mathematics at one school
(colours separate)
Blue represents the number of students taking
Higher Maths.
Brown represents the number of students
taking Mathematical Studies.
Green represents the number of students
taking Standard Maths.
This diagram is incomplete in that it has not got the dates on it, but I was interested in the
basic shape it would make rather than seeing it as a finished article to represent the data.
I decided to do the same thing but with Studies in the middle and Higher at the edge to
see how different it would look.
Figure 4
Polar area diagram to show students
taking Mathematics at one school
(colours separate)
Blue represents the number of students taking
Higher Maths.
Brown represents the number of students
taking Mathematical Studies.
Green represents the number of students
taking Standard Maths.
C Good
demonstration of
personal interest
This feels very different. The blue section actually looks less significant, to my eye, being
put at the edges. This made me think of something else I had read in Brasseur’s article, D Critical
“Nightingale arranged these colored areas so that the main cause of death (and the largest reflection on
sections)—deaths by disease—would be at the end of the wedges and would be more personal work
easily noticed.”4 I am sure that Brasseur thought that the colours were separate, and not
overlapped. However, comparing my diagrams to Nightingale’s original in Figure 1, I
became sure that she did mean them to be overlapped. I noticed that in the lefthand rose
in figure 1 (representing the second year) there is a wedge with blue at the edge followed
by a wedge with blue at the edge: "Blue" is a
typing error.
Figure 5
A zoom in of part of figure 1
This can happen in a diagram like my figure 2 of overlapping colours, but would be
impossible if the colours are separate as in figures 3 and 4. From this I deduced that the
colours on the diagram must be overlapping.
D Very good
The Second Problem example of
critical reflection
My diagrams were unlike Nightingale’s ones in that the total area of the sectors in figure
2 represented the total number of students taking the IB at this school over the 15 years.
Nightingale’s statistics were rates of mortality. Basically they can be thought of as
percentages of soldiers who died, but, as before, when I read through the articles again, I
was unsure what they were percentages of. Gill and Gill have table (Table 2) in their
article with headings “No. of soldiers admitted to the hospital” and “No. (%) of soldiers
who died”3. This might suggest that Nightingale was working with percentages of
soldiers who were admitted into hospital. Lewi is more definite and refers to the actual
statistic of one wedge of the third of Nightingale’s polar area diagrams as follows: “The
mortality during the first period was 192 per 1,000 hospitalized soldiers (on a yearly
basis)”9. However, Brasseur refers to the statistic in a wedge of Nightingale’s first
diagram as being “the ratio of mortality for every 1,000 soldiers per annum in the field”4,
in other words a percentage of the army actually on duty.
I decided to create a polar area diagram to act as an analogy to the possible situations as
follows:
Nightingale’s data My data
Number of soldiers in the army in a month Number of students taking the IB in a year
Number of soldiers taken to hospital Number of students taking Maths Studies
Number of soldiers dying of wounds Number of students gaining grade 7
Number of soldiers dying of disease Number of students gaining grade 6
C Another good
Number of soldiers dying for other reasons Number of students gaining grade 5
personal example
My analogy of drawing a diagram showing the numbers of soldiers dying as a percentage
of those admitted to hospital would then be the number of students gaining a grade above
4 as a percentage of those taking Mathematical Studies. I decided to do this one by hand,
partly to prove I could, and partly to see if it would throw any extra light on the
construction of the diagrams.
I gathered the data, found the percentages and used the percentages as A in the usual
15A
formula r = to find the radii needed to construct the diagram. The data is here:
π
Numbers gaining top three grades in Mathematical Studies
As percentage of those taking
Studies Radius required for each
Grade 7 Grade 6 Grade 5 Taking Total
Studies in year % grade 7 % grade 6 % grade 5 R7 R6 R5
1995 7 10 4 24 25 29.16667 41.66667 16.66667 11.80 14.10 8.92
1996 2 9 3 15 19 13.33333 60.00000 20.00000 7.98 16.93 9.77
1997 1 4 2 10 18 10.00000 40.00000 20.00000 6.91 13.82 9.77
1998 5 12 11 31 37 16.12903 38.70968 35.48387 8.78 13.60 13.02
1999 2 6 7 17 26 11.76471 35.29412 41.17647 7.49 12.98 14.02
2000 3 4 7 20 30 15.00000 20.00000 35.00000 8.46 9.77 12.93
2001 3 8 8 31 36 9.67742 25.80645 25.80645 6.80 11.10 11.10
2002 1 8 4 21 28 4.76190 38.09524 19.04762 4.77 13.49 9.54
2003 0 1 8 15 23 0.00000 6.66667 53.33333 0.00 5.64 15.96
2004 3 9 7 29 34 10.34483 31.03448 24.13793 7.03 12.17 10.74
2005 1 11 9 28 29 3.57143 39.28571 32.14286 4.13 13.70 12.39
2006 2 4 5 16 21 12.50000 25.00000 31.25000 7.73 10.93 12.22
2007 1 8 3 13 22 7.69231 61.53846 23.07692 6.06 17.14 10.50
2008 0 3 17 29 54 0.00000 10.34483 58.62069 0.00 7.03 16.73
2009 0 5 5 23 48 0.00000 21.73913 21.73913 0.00 10.19 10.19
Figure 6
Polar area diagram to show
percentages of students taking
Mathematical Studies who gained
grades above 4
Red represents the number of students
gaining grade 7.
Blue represents the number of students
gaining grade 6.
Green represents the number of students
gaining grade 5.
The purple areas represent coinciding B Student
numbers of students gaining grade 5 and 6. needs to watch
consistency of
location of colours
used here.
One thing which I learnt from this exercise is that you have to be very careful about your
scale and think through every move before you start if you don’t want to fall off the edge
of the paper! It is a far more tense experience drawing a diagram by hand because you
know that one slip will make the whole diagram flawed. A computer slip can be corrected
before you print out the result. My admiration for Florence Nightingale’s draftsmanship C This section
was heightened by doing this. shows a lot of
The other thing which drawing by hand brought out was that, if you draw the arcs in in personal interest.
the appropriate colours, the colouring of the sectors sorts itself out. You colour from the
arc inwards until you come to another arc or the centre. The only problem came when
two arcs of different colours came in exactly the same place. I got around this problem by
colouring these areas in a totally different colour and saying so at the side.
At this point in my research someone suggested some more possible websites to me, and
following these up I found a copy of Nightingale’s second diagram which was clear
enough for me to read her notes, and a copy of the original data she used. The first of
these was in a letter by Henry Woodbury suggesting that Nightingale got her calculations
wrong and the radii represented the statistics rather than the area.7 The letter had a
comment posted by Ian Short which led me to an article by him8 giving the data for the
second diagram and explaining how it was created.
The article by Short8 was a joy to read, although I could only work out the mathematical
equations, which were written out in a way which is strange to me ( for example
“$$ \text{Area of sector B} = \frac{\pi r_B^2}{3}=3$$”8 ) because I already knew what
they were (The example had a sector B in a diagram which I could see had
1 2π 2 π 2
=areaB = rB rB ). The two things I found exciting from this article were the
2 3 3
table of data which Nightingale used to create the second diagram, and an explanation of
what rates of mortality she used. She described these as follows; “The ratios of deaths
and admissions to Force per 1000 per annum are calculated from the monthly ratios
given in Dr. Smith’s Table B”4 and I had not been able to understand the meaning of this
from the other articles. (Brasseur adds that “Dr. Smith was the late director-general of the
army.”4). Using Short’s article I was able to work out what it meant.
I will use an example of data taken from the table in Short’s article, which is in turn taken
from “A contribution to the sanitary history of the British army during the late war with
Russia” by Florence Nightingale of 18598. In February 1855 the average size of the army
was 30919. Of these 2120 died of ‘zymotic diseases’, 42 died of ‘wounds & injuries’ and
361 died of ‘all other causes’. This gives a total of 2120 + 42 + 361 = 2523 deaths. 2523
2523
out of 30919 means that ×1000 = 81.6003 men died per 1000 men in the army in
30919
that month. If the size of the army had stayed at 30919, with no more men being shipped
in or out, and the death rate had continued at 81.6 deaths per 1000 men per month over 12
months, the number of deaths per annum would have been 81.6003 ×12 = 979.2 per 1000
men in the army. In other words 979.2 deaths per 1000 per annum.
This understanding of the units used allowed me to finally understand why O’Connor
says of the death rate in January 1855, “if this rate had continued, and troops had not been
replaced frequently, then disease alone would have killed the entire British army in the C, D Interesting
Crimea.”5 The number of deaths due to disease in January 1855 was 2761 and the
background
2761
average size of the army was 32393. This gives a rate of ×1000 ×12 = 1022.8 information and
32393 reflection
deaths from disease per 1000 per annum. Another way of looking at it is that if 2761 had
dies each month from disease, 2761×12 = 33132 would have died in 12 months, but there
were only 32393 in the army!
As an aside, I noticed that O’Connor quoted the mortality rate for January 1855 as “1,023
per 10,000 being from zymotic diseases”5. Another example that we should not trust
everything we see in print.
Having sorted this out I was ready to attempt my recreation of figure 1. I decided to do
the right hand rose only, covering April 1854 to March 1855. The following table shows
the data taken from Short’s article in blue and my calculations in black:
Average Wounds
size of Zymotic & Z/S*1000*12 Radius W/S*1000*12 Radius O/S*1000*12 Radius
army diseases injuries Other (Az) for (Aw) for (Ao) for
Month (S) (Z) (W) (O) (1 d.p.) Zymtotic (1 d.p.) Wounds (1 d.p.) Other
Apr-54 8571 1 0 5 1.4 2.3 0.0 0.0 7.0 5.2
May-54 23333 12 0 9 6.2 4.9 0.0 0.0 4.6 4.2
Jun-54 28333 11 0 6 4.7 4.2 0.0 0.0 2.5 3.1
Jul-54 28722 359 0 23 150.0 23.9 0.0 0.0 9.6 6.1
Aug-54 30246 828 1 30 328.5 35.4 0.4 1.2 11.9 6.7
Sep-54 30290 788 81 70 312.2 34.5 32.1 11.1 27.7 10.3
Oct-54 30643 503 132 128 197.0 27.4 51.7 14.1 50.1 13.8
Nov-54 29736 844 287 106 340.6 36.1 115.8 21.0 42.8 12.8
Dec-54 32779 1725 114 131 631.5 49.1 41.7 12.6 48.0 13.5
Jan-55 32393 2761 83 324 1022.8 62.5 30.7 10.8 120.0 21.4
Feb-55 30919 2120 42 361 822.8 56.1 16.3 7.9 140.1 23.1
Mar-55 30107 1205 32 172 480.3 42.8 12.8 7.0 68.6 16.2
Az is the death rate per 1000 per annum from disease, Aw is the death rate per 1000 per
annum from wounds and Ao is the death rate per 1000 per annum from other causes. For
2π π
this diagram there are 12 divisions so each sector has an angle of = and an area of
12 6
1π 2 π 2 12A
r = r . So for each radius r = .
26 12 π
I will show my final polar area diagram side by side with Nightingale’s original version:
A Demonstration
Figure 7. of the fulfillment
Nightingale’s original of the aim
“Diagram of the causes of
mortality in the army in the
east” and my recreation.
I have to admit that I felt rather proud once I had done this! However, looking at the
C Personal
September 1854 wedge I realised that the two diagrams didn’t correspond. In
Nightingale’s original diagram I can see that there are more deaths from other causes than ownership
from wounds. In my version there are fewer deaths from other causes than from wounds.
All other versions of the original in other articles I looked at ( Gill and Gill3, Brasseur4,
O’Connor5, Woodbury7, Riddle10, Small11, Lienhard6) are as the original, but the table in
Short definitely shows fewer deaths from other causes than from wounds8.
Conclusion
I started out to try to lean how to recreate the polar area diagram which Florence
Nightingale made to communicate to other people just how bad the situation was in army
C Good link to
hospitals. This diagram shouts a need for reform. Look at it. The blue represents deaths
which could be avoided with a bit of organisation and care. The red represents deaths due human factor
to the actual battles. Florence Nightingale had copies of her report containing her
diagrams published at her own expense and sent them to doctors, army officers, members
of parliament and the Queen. Following her persistent lobbying a commission was set up
to improve military barracks and hospitals, sanitary codes were established and
procedures were put in place for more organised collection of medical statistics4. It is a
very shocking picture with a huge snowball of social change behind it. It has been an
A Very good
exciting adventure to drill down to a real understanding of its construction.
conclusion
However, the biggest lesson I have learnt from this research is that you can’t trust what
you read. As I have argued in the main text, I am moderately sure that Brasseur thought
the colours of the second diagram did not overlap4, I think O’Connor got his death rates
wrong for January 18555, and I think Short may have transcribed the data incorrectly for
September 18548. According to Brasseur, Florence Nightingale cross checked her data D Reflection
and was systematic about addressing objections to her analysis4. Everyone can make on what has
mistakes, and errors can propagate if we just quote what someone else says without been learned
looking for corroboration.
I have been left with a desire to find out more about this tenacious woman who wouldn’t
let society mould her into a genteel wife. Also, if I ever get the chance, I would like to get
a look at one of the 2000 copies of “Notes on Matters Affecting the Health, Effiency and
Hospital Administration of the British Army. Founded Chiefly on the Experience of the
Late War” which Florence Nightingale had published in 1858, to see the actual table of
data and check the numbers for September 1854.
References/Bibliography
1.Duthie, Eric ed. The Children’s Book of Famous Lives.Odhams Press Ltd, London 1957
6. Lienhard, John H., Nightingale’s Graph, The Engines of Our Ingenuity. 2002
<http://www.uh.edu/engines/epi1712.htm> viewed 26th July 2009
9. Lewi, Paul J. Florence Nightingale and Polar Area Diagrams, Speaking of Graphics.
2006 < www.datascope.be/sog/SOG-Chapter5.pdf> viewed 26th July 2009
Mathematical Exploration
I have decided to centre this mathematical exploration around the topic of weather. C A good
To begin, I looked at the stimuli given to me by my teacher. While looking at different ideas, personal
weather was a topic that stood out to me, because never before, had I thought about the perspective
mathematical significance in weather. Would it be possible to predict tomorrow’s weather using
Mathematics? I did some research about the significance in meteorology, but did not find anything
very interesting in particular. Nevertheless, I wanted to stick to the topic of weather and decided to
relate it to a more personal matter. Having lived in a country like Germany for nearly all my life,
one is face to face with rain almost every day of the year. It comes and goes, no matter what. I thus
decided to write this exploration about rain, that is falling objects. Is it possible to model rainfall?
What would I need to consider? I first decided on the factors, which I wanted to account for in my
model and then looked a little closer at rain itself: There are different kinds of rain, and one would
have quite different experiences walking through a light drizzle than being outside on the street A Clarity
during a thunderstorm. needed on
I thus decided to explore different sizes of raindrops, first very small drops with diameter how these
d ≤ 0.008cm and then larger drops with diameter d ≥ 0.125cm . In addition, falling objects are not values were
only dependent upon gravity and mass itself, but on air resistance. Raindrops thus experience an chosen
upward drag, which must be accounted for when modelling rainfall. The velocity upon impact, thus
terminal velocity of the raindrop, and the time it takes the raindrop to reach the ground were
determined.
Finally, the model was extended to a parachute jump, which in a way is a very, very large raindrop
falling down from the sky. However, there is one essential difference: the parachute itself, as a
skydiver will always experience two periods of falling: the free-fall and the slow descent, after the
parachute is activated. This must be accounted for when developing a model.
Introduction:
Rain is a phenomenon we come across on a regularly basis in our day-to-day life. But what do we
actually know about it? We know it is liquid precipitation and that rain can be of different intensity,
such as light rain (a drizzle) or heavy rain (a storm). This exploration aims to develop a model for
rainfall.
In order to correctly model rainfall, I decided to first determine the factors related to rainfall, which
are thus of importance to the model: As mentioned above, the intensity of rain is never the same. It
is a measure of the amount of rain that falls over time, thus measured in the height of the water
layer covering the ground in a period of time, millimetres per hour. But not only the intensity of rain
matters. In order to determine the other factors, I found it easiest to just think of a falling object,
such as a pen, which is dropped at a certain height. What factors will have an impact upon the pen?
Gravitational force would be the first thing that springs to mind, but also acceleration and velocity
are important factors to consider, as is the distance of the pen from the ground and the time it takes
the pen to reach the ground.
To develop the first model taking into account the factors mentioned above, I did some research
about the forces upon falling objects and found Newton’s Second law of motion, which states that
force equals mass times acceleration. 1 This can be written as:
dv d 2 s
a= =
dt dt 2
We thus obtain a differential equation for a raindrop with constant acceleration, for which, in order
A Further
to solve it, we first need to set some initial conditions: The raindrop is starting at rest from 1000
metres in the air, thus the origin. By doing some research, it was obtained that the constant clarity needed
acceleration of falling objects, on which the only significant force in gravitational force, is
g = 9.81ms −2 . 2 It can also be stated that, at origin, v(0) = 0, s(0) = 0,t = 0 .
dv
=g
dt
dv = gdt
∫ dv = ∫ g dt
v = gt + k
To find the constant of integration, the initial conditions explained above must be used, thus
v(0) = 0,t = 0
0 = g(0) + k
k=0
Thus:
v = gt
ds
=v
We can now use another position equation, dt , to continue. Again, basic integration is used to
solve the differential equation.
1
“Newton’s Three Laws of Motion“, Astronomy 162 Stars, Galaxies, and Cosmology,
<http://csep10.phys.utk.edu/astr161/lect/history/newton3laws.html>, 1 May 2010
2
“Acceleration of Gravity and Newton’s Second Law of Motion“, The Engineering ToolBox,
<http://www.engineeringtoolbox.com/accelaration-gravity-d_340.html>, 3 May 2010
v = gt
ds
v=
dt
ds
= gt
dt
ds = gtdt
∫ ds = ∫ gt dt
g 2
s= t +k
2
To determine the constant of integration, the initial conditions have to be used again, thus
s(0) = 0,t(0) = 0 . Thus,
B Explanation
needed
g 2 why t(0) is
0= (0) + k
2 included here
k=0
We now get:
g
s = t2
2
−2
As stated above, s = 1000m and g = 9.81ms . These values can be substituted in the above
equation to solve for time, t:
9.81 2
1000 = t
2
2000 = 9.81t 2
2000
t2 =
9.81
2000 E Accuracy
t=
9.81 is too great
t = 14.278 for three
significant
To determine the velocity of the raindrop upon arriving on the ground we can use the first equation, figures.
v = gt .
However, when observing the constant acceleration model, it can be said that it will work for any
size of raindrop, since neither acceleration, nor velocity is dependent on mass.
dv
When taking a closer look at the equation for acceleration, = g , we find that it is not possible to
dt
determine the terminal velocity of the raindrop, as g = 9.81ms −2 and acceleration can thus never be
zero. This means, that the raindrop will not stop accelerating, which in reality, is not the case. 3
Fdrag ∝ adrag ∝ v
Thus,
adrag = kv
−1
Experiment shows that k = 12.2s .
−2
As stated above acceleration, when not accounting for air resistance, is constant, g = 9.81ms .
In order to find a new differential equation for this model, we now need to account for the constant
acceleration and the upward drag. Acceleration is thus given by:
dv
a=
dt
dv
= g − kv
dt
Using this differential equation we can easily find the terminal velocity of the raindrop. Terminal
velocity of a falling object is a constant velocity reached when there is no resultant acceleration,
acceleration is thus zero. We can thus obtain the terminal velocity from the model:
0 = g − kv
g = kv
g
v=
k
3
Emerson, Brandon, “Raindrop Modeling“, <bpemer.people.wm.edu/Raindrops.doc>, 28 April
2010
4
“The Forces at Work – Gravity“, < http://ffden-
2.phys.uaf.edu/211.fall2000.web.projects/vlad%20paverman/forces.htm>, 22 May 2010
dv
∫ g − kv = ∫ dt
1 E Good
− ln (g − kv ) = t + a mathematics
k
demonstrated
ln(g − kv) = −kt + a
g − kv = e− kt + a B Incorrect
g − kv = e− kt × ea notation for
ea is a constant and can thus be written as b to simplify the equation. constant
We now get:
g − kv = b × e− kt
To obtain the constant of integration, b, we consider the initial conditions, which are the same as in
the first model, v(0) = 0,t = 0 .
Thus,
B Poor use of
g − k(0) = b × e− k (0) mathematical language
g=b
Plugging in the constant of integration and solving for the velocity, v we get:
B Poor
g − kv = ge − kt use of
terminology
g − ge− kt = kv ("prove")
g − ge− kt
v= E Good
k
g example of
v = (1 − e− kt ) connecting
k
different areas
We can now graph the equation above to prove that the terminal velocity, which was determined of mathematics
earlier, is correct. The graph should, at one point, have a gradient of zero. Thus, once the graph
reaches a horizontal plateau, we can determine the terminal velocity. E Incorrect
analysis
Point at which the gradient of
the curve is firstly equal to zero,
x=1, y=0.804
v 9.81
y= (1 − e−12.2t )
12.2
B Poor
display of
graph and
associated
data
As seen above, the graph is horizontal, and the point at which the gradient is equal to zero is
E
reached very fast, within a minute. It can thus be assumed that the droplets travel with terminal
velocity for almost the entire fall. This seems rather odd at first, why would a falling body Incorrect
experience zero acceleration for such a long time? This is due to the retarding force of air analysis
resistance, which exists because air molecules collide into a falling body and create an upward force
opposite gravity. Eventually, this upward drag balances the weight of the falling object, which will
thus continue to fall at constant, terminal velocity. 5
ds
=v
As before, we now use a second position equation, dt . Thus:
g
v= (1 − e− kt )
k
ds g
= (1 − e− kt )
dt k
g
ds = (1 − e− kt )dt
k
5
Elert, Glenn, “Speed of a Skydiver (Terminal Velocity)“, The Physics
Factbookhttp://hypertextbook.com/facts/JianHuang.shtml, 10 May 2010
g
∫ ds = ∫ k (1 − e
− kt
)dt
g 1
⇒s= t + e− kt + c
k k
Solving for the constant of integration, c, we use the initial conditions, s(0) = 0,t(0) = 0 . We then
get:
g 1
0= (0 + ) + c
k k B Poor
g use of
⇒c=− 2
k mathematical
language
Plugging in the constant of integration:
g 1 g
s= (t + kt
)− 2
k ke k
g 1 1
s = (t + − )
k kekt k
We can now substitute in the known values, s = 1000, g = 9.81, k = 12.2 , to obtain:
9.81 1 1
1000 = (t + − )
12.2 12.2e12.2t 12.2
Instead of solving the equation algebraically, it is better to just graph the equation
9.81 1 1
y= (t + − )
12.2 12.2e 12.2t
12.2 and find the intersection with the equation y = 1000 . This can be
done on the graphing display calculator. However, to avoid complicated fractions it is best to first
simplify:
9.81 9.81
12200 = 9.81t + −
12.2e12.2t
12.2
We thus graph the two equations and find the point of intersection of the two graphs.
y Point of
intersection of
the two
graphs:
(1243.7109,
y = 12000 Typing mistake
12200)
condoned
9.81 9.81
y = 9.81t + 12.2t
−
12.2e 12.2
x
We thus get obtain that, at y=12200, x=1243.711. This number now needs to be divided by sixty, to
determine the time in minutes it will take the raindrop to reach the ground.
1243.711
= 20.729
60
It thus takes the raindrop approximately 21 minutes to reach the ground. Having found the time, we
can now determine the raindrop’s velocity upon reaching the ground. We found that
g
v= (1 − e− kt )
k
Thus:
9.81
v= (1 − e(−12.2 × 20.729) )
12.2
v = 0.804m / s
We will now again extend the model of rainfall. It is unreasonable to assume that all droplets are
smaller than 0.008cm. We will thus consider larger raindrops with a diameter d ≥ 0.125cm . To
correctly model the rainfall of the larger drops, we will immediately account for the upward drag, B Error:
which in this case, experiment shows, is proportional to velocity squared and the constant of mentioned as
proportionality is c = 0.0097 . "c" here but
We can thus alter the model for smaller raindrops according to the new conditions. Thus: referred to as
"k" from here on
dv
= g − kv 2
dt
dv
∫ g − kv 2
= ∫ dt
The integral on the left seems rather difficult, as we need to use partial fractions to integrate. Using
partial fractions, we know that:
1 1 1 C New mathematics
2
= − to this student and
g − kv 2 g × (v k + g ) 2 g × (v k − g )
not covered in class.
We can now use this to integrate, and, as g is a constant, we can rewrite it as: Good demonstration of
personal involvement
dv dv
∫2 g × (v k + g )
−∫
2 g × (v k − g ) ∫
= dt
1 dv 1 dv
×∫ ×∫
v k− g ∫
− = dt
2 g k v+ g 2 g
v k+ g B Notation
ln
1 v k− g error—
× =t +C constant
2 g k
appears late
v k+ g
ln = (t + C) × 2 g k
v k− g
ln
kv + g
kv − g
( )
= 2 kg t
Simplifying:
e
(2 kg )t = kv + g
E Good
kv − g
mathematical
e
(2 kg )t
( kv − g = kv + g ) manipulation of
the equations and
(2 ) − 1 =
kg t (2 ) + 1
kg t
kv e ge expressions here
(2 ) + 1
kg t
g e
v=
k e(2 ) − 1
kg t
We have now found the formula for the velocity. This is a differential equation for the distance s.
Simplifying further we get:
g e(2 kg )t + 1
v=
k e(2 kg )t − 1
ds g e(2 kg )t − 1 + 2
=
dt k e(2 kg )t − 1
ds g 2
= 1 + (2 kg )t
dt k e − 1
g 2
∫ ds = k ∫ 1 + e(2 kg )t − 1 dt
In order to integrate this fraction we use the substitution:
u=e
(2 kg )t
du = 2 kge
(2 kg )t dt B Variable t
is here, but
du obscured by the
⇒ dt =
2 kgu brackets—possible
g g problem with the
du
s= t + 2 ∫ equation editor
k k 2 kgu(u − 1)
g g du
2 ∫
s= t+
k k g u(u − 1)
1 1 1
We can now use partial fractions to solve this last integral. We know that = − .
u(u − 1) (u − 1) u
Therefore:
du 1 1 E Sophisticated
∫ u(u − 1) = ∫ (u − 1) − u dt level of mathematics.
, which can be written as: Integration involves
1 1 dt dt partial fractions,
∫ (u − 1) − u dt = ∫ (u − 1) − ∫ u substitution, and multiple
This can now be integrated. stages. Could be tidied
dt dt further, so not elegant,
∫ (u − 1) − ∫ u = ln u − 1 − ln u but impressive
As we used a substitution for u above, we now need to substitute u back in.
u = e( )
2 kg t
We thus obtain:
ln e( ) − 1 − ln e( )
2 kg t 2 kg t
Substituting the part back into the equation above we get a solution for s:
s=
g
k
g
t + k 2 g ln e
( )
(2 kg )t − 1 − ln e(2 kg )t
We can now substitute in the known values, s = 1000, g = 9.81, k = 0.0097 , to obtain:
9.81
1000 =
0.0097
t+
9.81
(0.0097)2 × 9.81
ln e(
2
( 0.0097 × 9.81 t ) − 1 − ln e(2 )
0.0097 × 9.81 t
)
In order to solve for t, we can graph the two equations y = 1000 and
y=
9.81
0.0097
t+
9.81
(0.0097)2 × 9.81
ln e( (
2 0.0097 × 9.81 )t
− 1 − ln e(
2 0.0097 × 9.81 )t
)
and find the point of
intersection of the two graphs.
y
Point of
y = 1000 B x and y in
intersection:
x=1000, wrong order.
y=31.445 Correctly given
below, so condoned
9.81
y=
9.81
(
t + (0.0097)2 × 9.81 ln e
0.0097
(2 ) − 1 − ln e(2
0.0097 × 9.81 t )
0.0097 × 9.81 t
)
x
We thus obtain that at y=1000, x=31.445. The value we found for x is the time it takes the raindrop
to reach the ground. This value can now be substituted into the equation for velocity:
g e(2 kg )t + 1
v=
k e(2 kg )t − 1
We thus obtain:
9.81 e(2 9.81× 0.0097 )× 31.445 + 1
v=
0.0097 e(2 9.81× 0.0097 )× 31.445 − 1
v = 31.802m / s
dv
= g − kv 2
dt , as explained above. To find the terminal velocity we calculate v when acceleration is
equal to zero. Thus:
g = kv 2
g
v2 =
k
g
v=
k
9.81
v=
0.0097
v = 31.802
It is thus reasonable to assume that the large raindrops also travel at terminal velocity for most of
the time.
To apply the concept of raindrop modelling to human beings, one could now consider a parachute
jump. As I do not know much about skydiving, I first did some research around the basics of it.
A typical parachute jump consists of the individuals jumping out of an aircraft at approximately
4,400 metres altitude. The period of free-fall takes about a minute, then the parachute is opened to
slow the landing down, which takes between five and seven minutes.
To model a parachute jump, we can say that the parachutist, just like any other falling object,
−2
experiences constant acceleration g = 9.81ms , wince all objects fall with the same rate of
Typing error
acceleration regardless of their mass.
We also need to account for air resistance and can thus use the same model as before, when
modelling the fall of small raindrops.
dv
a=
dt
dv
= g − kv
dt
However, as the size of humans is not comparable to the size of small raindrops, we need to find a
different value for k, as the air resistance is not the same in this case. Also, the significance of air
resistance will ultimately depend on the position of the skydiver in the air. For example, a skydiver
in the spread eagle position will encounter more air resistance than a parachutist, who falls feet or
D Meaningful
head first, which related back to the cross sectional area of the falling object as explained earlier.
As explained earlier, the amount of air resistance is subject to change throughout the free fall, as it reflection
is also dependent upon speed. The air resistance is small at the start of the jump and the parachutist
accelerates downward. As the speed of the free faller increases, the amount of air resistance goes up
as well, thus slowing down acceleration. The air resistance force will eventually be equal and
opposite to the downward force due to gravity and at this point the skydiver’s speed remains
constant. Terminal velocity has thus been reached. 6
In order to develop a model for the parachute problem, one would thus have to divide the free fall
into three parts, each with a different amount of air resistance and then also develop a model for the
slow descent after the parachute is deployed. The mathematical process however is very simple and
the same as in the second model for small raindrops, k is just exchanged for the respective values of
the air resistance, that is the constant of proportionality in reference to velocity.
Having extended the mathematical exploration to the problem of a parachute jump, one can now
critically reflect upon the work done. We can start by looking at the first model, which merely looks
at raindrops as falling objects. It is an unrealistic model as such, as it does not account for air A Good conclusion
6
„KS§ 9K Air resistance and friction“, slideshare,
<http://www.slideshare.net/scienceinteractive/ks3-9k-air-resistance-and-friction>, 23 May 2010
resistance. The results could thus be used as a rough idea on how to develop the other models, but
other than that, were hypothetical.
The second model, which looks at smaller raindrops, takes air resistance into account and thus gets
a lot closer to reality. However, there are certain uncertainties to this model. To begin, we put all
raindrops with a diameter smaller that 0.008cm into one group, the results are thus not specific to
the respective size of the raindrop. In addition, we made the assumption that the amount of air
resistance is always that same. In reality, the amount of air resistance is subject to a number of
factors. As stated earlier, it depends upon the cross sectional area and speed of the object and thus
changes throughout the fall. However, this is only a minor inaccuracy, as we found out that
raindrops reach terminal velocity after a very short period of time and thus travel at terminal
velocity, and thus the same speed, for almost the entire fall. The air resistance, however, can also be
influenced by environmental factors, such as a strong wind.
These uncertainties are the same for the third model, looking at larger raindrops.
Bibliography:
Elert, Glenn. "Speed of a Skydiver (Terminal Velocity)." The Physics Factbook. Web. 10 May
2010. <http://hypertextbook.com/facts/JianHuang.shtml>.
Meade, Douglas B. "Ode Models for the Parachute Problem." Web. 16 May 2010.
<http://www.math.sc.edu/~meade/papers/sr-parachute.pdf>.
Phoebus, Ronald, and Cole Reilly. "Differential Equations and the Parachute Problem." 10 May
2004. Web. 15 May 2010.
<http://online.redwoods.cc.ca.us/instruct/darnold/DEProj/sp04/coleron/paper1.pdf>.
Smith, David A., and Lawrence C. Moore. "Raindrops." Math DL The MAA Mathematical Science
Digital Library. The Mathematical Association of America. Web. 05 May 2010.
<http://mathdl.maa.org/mathDL/?pa=content&sa=viewDocument&nodeId=315&pf=1>.
"Student Project: Skydiving." Pure Mathematics 30. Alberta Education, Feb. 2009. Web. 18 May
2010. <http://education.alberta.ca/media/930712/03%20pure30%20feb-
09%20skydiving%20student%20project_pdf%20final.pdf>.
Spirals in Nature
A Aim and
rationale
When researching about mathematics in nature I found that certain spirals are found in shell
shapes. The Nautilus is a marine mollusk with a spiral shell with partitions to create buoyancy. included in
The adult mollusk can grow up to 25-30 centimetres across and the shell can withstand depths of introduction
up to 650 metres underwater. The chambers of the shell are separated but are interconnected via
a tube running through them. The tube pours gas or liquid through the tube to move the creature
around to sink or float respectively). The Nautilus shell’s curves are logarithmic and equiangular
with slightly different proportions to other spirals such as the Golden ratio.
I have a fossil of an ammonite which has a spiral in it. I want to find out if this has an
Archimedian spiral (a spiral described in polar coordinates by r = aθ , where a is a constant), or a
logarithmic spiral (a spiral described in polar coordinates by r = kecθ where c = cot φ ), or
something else altogether.
Here I have scanned the fossil and will attempt to model the spiral shown. In order to model the
spiral, I have marked several points along it, according to the picture below and have computed
C Developed
the x- and y-coordinates.
own example
There were 31 points. These are the coordinates and a plot of the coordinates to make sure I have
the right shape.
x y
648 -639
434 -741
221 -786
7 -773 C Good
-212 -742 example of
-421 -586 personal
-639 -216 engagement
-635 105
-537 323
-421 461
-207 581
2 599
221 537
434 336 C Applying unfamiliar
506 114
mathematics of polar
429 -158
225 -332 curves and polar points
2 -349
-212 -212
-283 7
-198 203
7 305
221 198
221 -46.8 B Very nice
7 -126 linking between
-105 11
the scan, the data
11 167
73 96 and the graph
96 33
7 -47
-25 74
Now I want to see if this spiral can be described as r = aθ , so I need to find the radius and angle
for each point. I will find the angle in radians. The radius is easy:=r x 2 + y 2 . The angle is a
bit more complicated because I want the point ( −25, 74 ) to have an angle of less than π, but I
want the point ( −105,11) to have an angle of nearly 3π because you can see from the plotting of
y
the points that you have to go round more than a whole revolution to get there. arctan gives
x E Good
an angle in the 1st or 4th quadrant so I worked out that if I know which quadrant a point is in and I understanding
keep the points all in order I can work out the angle by thinking about how many times I need to of the use
y of the
add
= π, so θ arctan + nπ where I define n in th following spreadsheet:
x periodic
x y quadrant n θ r logθ logr nature of
648 -639 4 8 24.35433596 910.0687 1.386576 2.959074 tangents
434 -741 4 8 24.09177962 858.7415 1.381869 2.933862 above SL
221 -786 4 8 23.83603866 816.4784 1.377234 2.911945 student
7 -773 4 8 23.57100028 773.0317 1.372378 2.888197 expectations
-212 -742 3 7 23.28364524 771.6916 1.367051 2.887444
-421 -586 3 7 22.93895658 721.5518 1.360574 2.858268
-639 -216 3 7 22.31711851 674.5198 1.348638 2.828995
-635 105 2 7 21.82727704 643.6226 1.339 2.808631
-537 323 2 7 21.44963438 626.6562 1.33142 2.797029
-421 461 2 7 21.16042999 624.3092 1.325524 2.7954
-207 581 2 7 20.76261275 616.7739 1.317282 2.790126
2 599 1 6 20.41701336 599.0033 1.309992 2.777429
221 537 1 6 20.02993254 580.6979 1.301679 2.76395
434 336 1 6 19.50836196 548.8643 1.290221 2.739465
506 114 1 6 19.0711525 518.6829 1.280377 2.714902
429 -158 4 6 18.49667356 457.1706 1.267094 2.660078
225 -332 4 6 17.87436924 401.0598 1.252231 2.603209
2 -349 4 6 17.28449019 349.0057 1.237657 2.542833
-212 -212 3 5 16.49336143 299.8133 1.217309 2.476851
-283 7 2 5 15.68323333 283.0866 1.195436 2.451919
-198 203 2 5 14.91009692 283.5719 1.17348 2.452663
7 305 1 4 14.11422015 305.0803 1.149657 2.484414
221 198 1 4 13.29693121 296.7238 1.123751 2.472352
221 -46.8 4 4 12.35768885 225.901 1.091937 2.353918
7 -126 4 4 11.05107279 126.1943 1.043404 2.10104
-105 11 2 3 9.320396808 105.5746 0.969434 2.02356
11 167 1 2 7.788208383 167.3619 0.891438 2.223657
73 96 1 2 7.203847123 120.6027 0.857564 2.081357
96 33 1 2 6.614281384 101.5135 0.820483 2.006524
7 -47 4 2 4.860238346 47.51842 0.686658 1.676862
-25 74 2 1 1.896595442 78.1089 0.277975 1.892701
If the spiral follows an Archemedian spiral, r = aθ , so plotting r against θ should give a straight
line of gradient a intersecting the vertical axis at the origin. I have plotted rand θ and fitted a line
of best fit by computer.
I will try using the computer to draw r= 0.041132 θ 3 + 0.287964 θ 2 + 1.73643θ + 64.6669 on top
of the data.
E Good
understanding
of relationship
between the
graphs and
polar graphs
Now I will try fitting a logarithmic spiral. In the Encyclopaedia Britannica it said that the curve
was of the form r = kecθ where c = cot φ . I think c is a constant, so
ln r = ln kecθ
= ln k + ln ecθ
= ln k + cθ ln e
= ln k + cθ
So if I plot log r against θ I should get a straight line with gradient c and y-intercept ln k .
θ r lnr
24.35434 910.0687 6.81352
24.09178 858.7415 6.755468
23.83604 816.4784 6.705 lnr
23.571 773.0317 6.65032
23.28365 771.6916 6.648585
22.93896 721.5518 6.581404
22.31712 674.5198 6.514001
21.82728 643.6226 6.467112
21.44963 626.6562 6.440398
21.16043 624.3092 6.436646
20.76261 616.7739 6.424502
20.41701 599.0033 6.395267
20.02993 580.6979 6.364231
19.50836 548.8643 6.307851
19.07115 518.6829 6.251293
18.49667 457.1706 6.125057
17.87437 401.0598 5.994111
θ
17.28449 349.0057 5.855088
16.49336 299.8133 5.70316
15.68323 283.0866 5.645753
14.9101 283.5719 5.647466
14.11422 305.0803 5.720575
13.29693 296.7238 5.692802
12.35769 225.901 5.420097
11.05107 126.1943 4.837823
9.320397 105.5746 4.659418
7.788208 167.3619 5.120158
7.203847 120.6027 4.792501
6.614281 101.5135 4.620192
4.860238 47.51842 3.861117
1.896595 78.1089 4.358104
So
= r 45.09 × e0.123083θ
D Opportunities for
considering accuracy
and whether
using less accurate
parameters would
impact reasonableness
D Meaningful
reflection on
which is the
best approach
I think this looks very good. The cubic example looked very good too, however, so now I am
going to try to compare them. Every point on the spiral has a value of theta and a value of r and
each model has an approximate value of r for that theta. I have decided to find the absolute error
in each approximation of r and I will sum these errors for each of my models to see which has
the least sum and hence is the closest to the real points on the spiral.
A Huge,
detailed
tables stop the
exploration from
being concise.
The Archemedian spiral had an absolute error sum of over 6000. The logarithmic spiral had an
absolute error sum of just over 914 and the cubic spiral which I just made up because the points
when I plotted r against θ looked as if they could follow a polynomial curve, has the smallest
error sum of about 840. I expected the logarithmic spiral to fit best because I had read that
A Good
Nautilus shells follow these curves, but that doesn’t seem to be the case for my fossil.
conclusion
It would be interesting to find other examples of ammonites from the many photographs
available on the web, and to model their curves by cubic spirals and logarithmic spirals and to D Limited
see which seem to fit more closely to find out whether the cubic shape is a good model for these reflection
spirals in general or whether it was a coincidence in this one case. on how the
exploration could
be extended
Bibliography
http://www.bsu.edu/web/math/exchange/01-01/allen.pdf
Murphy, Amanarth. "Maths of Nature and Nature of Maths Chapter 1." Scribd, 11
Jan. 2009. Web. 6 Jan. 2010. <http://www.scribd.com/doc/21990600/Maths-of-
Nature-and-Nature-of-Maths-Chapter-1>.
Images
Chambered Nautilus. Digital image. Inquiry By Design, Proportion. 7 Nov. 2006.
Web. 6 Jan. 2010.
<http://blog.lib.umn.edu/myee/architecture/Nautilus%20Shell%202.gif>.
Finding a Pattern
The task is to find out how many moves it takes 64 disks to
transfer from one pole to another pole of the Tower of Hanoi. Solving the
puzzle using a smaller number of disks will be easier to analyze and
understand. Let’s look at how the Tower of Hanoi is solved using 1, 2,
and 3 discs.
B These diagrams
Figure B. Three Moves
are helpful, but
better clarity could
Move 1: move disk 2 to post B be achieved by
Move 2: move disk 1 to post C identifying which
Move 3: move disk 2 to post C is disk 1 and 2.
moves to transfer two discs to pole B (step 3 of Figure C). The number of
moves required shall be the variable M.
The next step is to transfer the remaining disc from A to C (step 4
of Figure C).
The last step is to transfer the discs from pole B to C. It will require
the same number of moves required in the first step also known as the
variable M. In Figure C it takes three moves to transfer n-1 discs from
pole B to pole C (step 7 of Figure C).
Reflecting on the steps you can see that to solve the puzzle you
move M amount of moves twice (from A to B, and from B to C). You also
make a single move from A to C. Mathematically the recursive pattern
should look like this:
B Not an
# of discs Total Moves Equation 2M + 1
equation
1 1 2(0) + 1 = 1
2 3 2(1) + 1 = 3
3 7 2(3) + 1 = 7
Table A.
In Table A under the Equation column, you are able to clearly see
why there is a recursive pattern present. The total moves from the 1 disc
puzzle become the M of the second puzzle as shown in the bold and
black numbers. The total moves for the 2 disc puzzle become the M of
the third puzzle as shown in the bold and grey numbers. Using the
recursive pattern, you are able to find the number of moves it takes for a
4, 5 disc puzzle and for an n disc puzzle as shown in Table B on page 4.
C More diagrams
would help here.
1 1 2(0) + 1 = 1
2 3 2(1) + 1 = 3
3 7 2(3) + 1 = 7
4 15 2(7) + 1 = 15
5 31 2(15) + 1 = 31
n T 2(M) + 1 = T
Table B.
E Poor display
of knowledge—not
Referring back to the original task, the number of moves it takes to a recognized
solve a 64 disc puzzle is still unknown. It is possible to use the recursive geometric sequence
in differences
pattern and equation 2M + 1 to find the answer but in order to do this
one needs to know what M is. This reveals a weakness in recursive
patterns because in order to know what M is for 64 discs, one must
know M for 63 discs, M for 62 discs, and so on. This process would be
very time consuming and therefore not an efficient way of solving the
D Some
problem. reflection
here
Another problem with using the recursive pattern is if you wanted
to graph the data to compare the # of discs with the Total Moves. If one
wanted to use the recursive equation 2M + 1 one would discover that it is
not possible. This is because of the nature of a recursive equation. The M
variable is unknown unless the previous terms are known. Therefore it is
not possible to graph data using a recursive equation.
Perhaps a non-recursive way of looking at the puzzle would be
more efficient and useful. Using the data obtained from the recursive
pattern we can find an explicit pattern in the puzzle. An explicit pattern is
E Explicit versus
basically a non-recursive pattern meaning it doesn’t need information
recursive recognizes
from previous terms to obtain an answer. two important
mathematical ideas
related to patterns.
Creating a Formula
First, take the values from the first two columns of Table B. These values
are displayed below.
# of discs Total Moves
1 1
2 3
C Trying to find
3 7
a formula by using
4 15
graphical means
5 31
indicates independent
thinking.
n T
Table C
B Well-labelled
graph; however,
student should try
to avoid wasted
open space by
adjusting the domain.
Graph A.
Graph B compares the data points with the quadratic function y=x2.
B Student doesn't
acknowledge that
discrete data are
being modelled
with a continuous
function.
Graph B.
In graph B, the quadratic function y=x2 has a similar trend to the
data points except the quadratic function forms a parabola. This is
unsuitable to the data points because they do not form a trend in the E Confusing why
negatives are
shape of a parabola. Even when negative numbers are used (though this considered here
is unrealistic as it’s impossible to have negative # of moves), the trend is
Graph C compares the data points with the exponential function y=2x.
Graph C.
Graph D shows the function y=2x - 1 compared with the data points.
Graph D.
Reflection A Rationale
identified at end
I chose this topic because I remember playing this simple but
challenging game at Science World. While playing the game I did not
naturally think of it as a mathematical concept but naturally I would
identify patterns in the game as I solve the puzzle and the difficulty
increases. I chose this topic because the history of it fascinates me. Also I
knew that the game had several mathematical concepts that were I
recently learned. I wanted to see if my mathematical knowledge of
transformations, functions, and patterns could be put to use and solve
the legend behind The Tower of Hanoi.
When exploring the mathematics of this game, I learned to see
things in life in a mathematical point-of-view as well as seeing
mathematics in a realistic point-of-view. The process of solving the
legend showed the relationships that mathematical concepts have with
each other as well as with life. I learned that everything is connected
mathematically and even the simplest things can be dissected and
analyzed in a mathematical way.
Conclusion D Rather
simple reflections
Words Cited
Bogomolny, A. "Tower of Hanoi from Interactive Mathematics Miscellany and
Puzzles." Interactive Mathematics Miscellany and Puzzles, Index.
knot.org/recurrence/hanoi.shtml>.
Dr. Math. "Math Forum: Ask Dr. Math FAQ: Tower of Hanoi." The Math Forum @
<http://mathforum.org/dr.math/faq/faq.tower.hanoi.html>.
"LHS: Tower of Hanoi Facts." Lawrence Hall of Science. Lawrence Hall of
<http://lawrencehallofscience.org/java/tower/towerhistory.html>.
Weisstein, Eric W. "Tower of Hanoi -- from Wolfram MathWorld." Wolfram
MathWorld: The Web's Most Extensive Mathematics Resource. Wolfram
<http://mathworld.wolfram.com/TowerofHanoi.html>.
A The introduction
includes a clear and
explicit description of
the aim, as well as a
rationale. The plan of
action is not explicitly
explained.
B Constraints relative to
this specific study are
appropriately defined.
C The Mathematical
exploration relates to
personal interest.
C A specific model is
chosen.
B The variable P
used in the formula is
explained at this point.
The definition (fluid
density) is made at the
next paragraph.
B The (differential) ∆
operator is not defined
in the equations provided.
It is implicitly explained
at the next line.
A, B Interesting
explanations on the
method and process
that will be used.
E Very good
understanding of the
complexity of the
modelling.
A, B, C Although the
explanations are not
correctly expressed,
the symmetry of the
quadratic function is
used in order to avoid
repetition of calculations.
This helps making the
work concise.
C Personal example.
A Appropriate
explanations.
B, E Technology is used
in order to determine
this function.
E Interesting idea.
E Appropriate use of
mathematics, beyond the
scope of the Math HL
syllabus.
E Correct use of
calculus.
A Extended research
has been conducted.
D Critical reflection on
the research made and
its applications to the
current problem.
E Good understanding
of the subject is
demonstrated.
B Good use of
technology, in order to
generate (correct) data.
E Good reasoning,
supported by appropriate
explanations
B, E Correct and
appropriate data are
generated.
A, B, E Final results
for the drag force are
produced. The lines of
reasoning are consistent
and easy to follow
through.
B Appropriate display of
the results.
B, E The data
manipulated are
displayed and the linear
relationship is visible.
The unit of the y-axis
is not properly expressed
(confusing notation). The
line of best fit is also
displayed.
B ...function...
B Interesting
display of the data
which offers a visual
representation of the
data presented in the
table.
B It is important
to mention that this
function is an estimation
as more calculations are
necessary to prove it.
D Critical analysis of
the initial assumptions/
simplifications made for
the investigation.
B, D Interesting
suggestion for possible and
realistic improvements.
No explanation of
notation.
The mathematics
requires more clarity and
explanation.
Source?
Probability of what?
Aim and
rationale
Citation?
Evidence of reflection
Labels?
Evidence of understanding
A1
See bibliography 2
wikipedia
Good use
of geogebra
B2
B2
A2
E1
B3
B4
B4
E2
B7 OK - student
tabulates results
A Confusion between
explanationanddiagram.
Clear well
presented
photo.
Consideration of error is a
good example of meaningful
reflection.
Personal
engagement.
Aim.
Defines variables.
variables
Personal
engagement.
What is N?
Good observation
A little
too wordy.
Good
observation.
Reflection.
Personal
engagement
Personal
example.
Personal
example.
Limited data
A+ Introduction and
background information
In bibliography
Personal engagement
D+
B - Representation of fractions.
D + Good analysis.
More analyses
B - Why n-1?
Why?
Good point.
Student reflects
on exploration.
A Rationale.
D+
A Defines types
of domes.
D+
C+
B+
D + Meaningful
Does it?
Could be clearer.
D+
B - Define x
2.44?
D+
D+
D+
A+
C+
D+
B+
B+
B+
C + New to student.
E+
D+
D+
C+
E+
D+
D+
D+
— 35.479...
D+
B - Missing dx
D+
D+
Good conclusion
B + x- and y- defined.
What is B?
D -, E - Why do you
think a linear regression
makes sense?
D-
C+
D+
A + Aim
C+
B+
Units in table.
E+
Parameters
E+
B - Two “a’s”
D+
D+
C+
B+
D+
B+
D+
D+
A + Returns to aim.
D+
A + Good introduction.
C + Topic is personal.
A + Clear aim.
C+
D+
D+
B + Clear graph.
y-axis explained.
B + Variables defined.
Condone * for
multiplication
x condoned here
mathematics correct