Professional Documents
Culture Documents
Introduction
This is a math class! Why are we writing?
There is a good chance that you have never written a paper in a math class before. So you
might be wondering why writing is required in your math class now.
The Greek word mathemas, from which we derive the word mathematics, embodies the
notions of knowledge, cognition, understanding, and perception. In the end, mathematics is
about ideas. In math classes at the university level, the ideas and concepts encountered are
more complex and sophisticated. The mathematics learned in college will include concepts
which cannot be expressed using just equations and formulas. Putting mathemas on paper
will require writing sentences and paragraphs in addition to the equations and formulas.
Mathematicians actually spend a great deal of time writing. If a mathematician wants
to contribute to the greater body of mathematical knowledge, she must be able
communicate her ideas in a way which is comprehensible to others. Thus, being able to
write clearly is as important a mathematical skill as being able to solve equations.
Mastering the ability to write clear mathematical explanations is important for
non-mathematicians as well. As you continue taking math courses in college, you will come
to know more mathematics than most other people. When you use your mathematical
knowledge in the future, you may be required to explain your thinking process to another
person (like your boss, a co-worker, or an elected official), and it will be quite likely that
this other person will know less math than you do. Learning how to communicate
mathematical ideas clearly can help you advance in your career.
You will find that writing good mathematical explanations will improve your knowledge
and understanding of the mathematical ideas you encounter. Putting an idea on paper
requires careful thought and attention. Hence, mathematics which is written clearly and
carefully is more likely to be correct. The process of writing will help you learn and retain
the concepts which you will be exploring in your math class.
symbols can correspond to different parts of speech. For instance, below is a perfectly good
complete sentence.
1 + 1 = 2.
The symbol = acts like a verb. Below are a couple more examples of complete sentences.
3xy < 2.
5z R.
9 s 6= t.
Can you identify the verbs? On the other hand, an expression like
5x2 z 10y
is not a complete sentence. There is no verb. Such an expression should be treated as a
noun. Can you identify the nouns in the previous examples?
Formulas and equations need to be contained in complete sentences with proper
punctuation. Here is an example:
The total revenue, R, made from selling widgets is given
by the equation
R = pq,
where p is the price at which each widget is sold and q is
the number of widgets sold. Based on past experience,
we know that when widgets are priced at $15 each, 2000
widgets will be sold. We also know that for every dollar
increase in price, 150 fewer widgets are sold. Hence, if
the price is increased by x dollars, then the revenue is
R = (15 + x)(2000 150x)
= 150x2 250x + 30, 000.
Notice how punctuation follows each of equations. A computation which ends a sentence
needs to end with a period. Computations which do not end sentences are followed by
commas.
A good way to improve your mathematical writing is by reading your writing, including
all of the equations, out loud. Your ears can often pick out sentence fragments and
grammatical errors better than your eyes. If you find yourself saying a series of fragmented
sentences and equations, you should do some rewriting.
There are a couple of other important things to observe in the above example. Notice
how we is used. The use of first person is common in mathematics, especially the plural
we, so dont be afraid to use the word we in the papers you write in your math class.
Another thing to notice is that important or long formulas are written on separate
lines. You can make your mathematical writing easier to read if you place each important
formula on a line of its own. Its hard to pick out the important formulas below:
If d is Bobs distance above the ground in feet, then d =
100 16t2 , where t is the number of seconds after Bobs
Flugelputz-Levitator is activated. Solving for t in the
equation 100 16t2 = 0, we find that t = 2.5. Bob hits
the ground after 2.5 seconds.
This is clearer:
If d is Bobs distance above the ground in feet, then
d = 100 16t2 ,
where t is the number of seconds after Bobs
Flugelputz-Levitator is activated. Solving for t in the
equation
100 16t2 = 0,
we find that t = 2.5. Bob hits the ground after 2.5 seconds.
!!
(3 1) = 0 = 3 = 1 = x = 0.
x
Do not use the equal sign when you really mean the next step is or implies. The above
example is really saying that 1 = 0 = 1! Using arrows instead of equal signs is a slight
improvement, but still not desirable:
(3x 1)2 = 0 3x = 1 x = 0.
For a difficult computation where the reader might not readily follow each step, you can
include words to describe the steps you take.
We want to solve for x in the equation
32x 2(3x ) = 1.
We can rewrite this equation in terms of 3x :
(3x )2 2(3x ) + 1 = 0.
After factoring, this becomes
(3x 1)2 = 1
and it follows that 3x = 1, or x = 0.
However, make sure that your paper has a single flow. Dont explain a calculation using
the two-column method.
32x 2(3x ) = 1 Solve this equation.
(3x )2 2(3x ) + 1 = 0
(3 1) = 0
x
Factor.
x=0
Solve for x.
3 =1
Miscellaneous comments.
Here are a couple of other pointers to help you get started with your mathematical writing.
Dont start a sentence with a formula. While it may be grammatically correct, it
looks strange.
t = 5 when w = 2000, so we can conclude that the new
factory will be completely overrun with cockroaches in 5
years.
f is globberfluxible at x = 3.
Dont turn in pages of unreadable scribbles to your professor. In college, papers are
typed. They are also usually double-spaced with large margins. Mathematics papers
adhere to the same standards as papers written for other classes.
While it is a good idea to type your paper, you may have to leave out the formulas
and insert them by hand later. It is perfectly acceptable to write formulas by hand in
a math paper. Just make sure that your mathematical notation is legible. If you do
decide to type the equations, please be aware that variables in equations and formulas
are usually italicized (to set them apart from the text). Many word processing
programs contain equation editors. In newer versions of Microsoft Word, the equation
editor is available under the Insert menu. Select Object..., and then Equation.1 If
you are going to be writing a lot of technical documents, it might be worthwhile to
learn TEX or LATEX. These are professional mathematical typesetting languages. This
document was written with LATEX. You may also find satisfactory results typing
papers in Maple or some other mathematically oriented software program.
Use mathematical notation correctly. As you learn to write more complicated
formulas, it is all too easy to leave out symbols from formulas. Learn how to use
symbols properly!
Use language precisely and correctly. Make sure that the words you use really mean
what you think they mean. Mathematics requires very precise use of language.
Another thing to avoid is overuse of the word it. Mathematical papers with a lot of
pronouns like it and that tend to be hard to read. It is often hard for the reader
to see what it is referring to. If you, the author, are also having difficulty seeing
what it is referring to, then you may be having some difficulty with the
mathematical ideas; you may need to think more about the ideas you are writing
about.
Try to write as simply and directly as possible. No one likes to read ponderous
pretentious prose.
1
In Microsoft Word, it is also possible to place a button on the tool bar which activates the equation
editor. Select Configure... beneath the Tools menu. In the window that pops up, select the Commands
tab. Under the Insert category you will find the Equation Editor command. Drag the equation editor
icon to the tool bar.
Some papers then state the answer to the problem right after the introduction. Other
papers place the answer at the end. This is a matter of taste. Sometimes, the end result
is the most important thing in the paper. You may need to place the end result at the
beginning to entice the reader. On the other hand, sometimes the method of arriving at
the end result is more important. In such a case, putting the result at the end may be
more sensible.
In any case, it is best to state the result in terms of the original problem using
real-world terms.
The solution is t = 6.
Make sure that the arguments you write are carefully organized. It may help you to
write an outline before you begin writing a mathematics paper. Writing an outline will also
help you think about the concepts more clearly and thus will help you learn the material.
As you write about more advanced mathematical problems, organization will become even
more important.
need to show everything.) The reader of a college mathematics paper will probably not be
interested in reading how to multiply 5 and 74. Leave out what is unimportant. On the
other hand, dont leave out anything which is critical to the key ideas you are trying to
explain. Learning what is important and what is unimportant will help you understand
mathematics better.
You should not assume that the reader is familiar with the problem you are solving.
While you do not need to restate the problem in its entirety, be sure to give an overview of
all important details in the problem. You also should not assume that the reader is in the
same mind set as you. In your writing, state any assumptions which you have made. For
instance, in physics problems, it is often assumed that everything is frictionless. But just
because this assumption is made nearly all the time doesnt mean that your reader will
automatically make this assumption; your reader may not be familiar with physics. Just
because you assume something is true doesnt mean that your reader will. So write it down!
What is n? If n = 8.5 is the above statement true? A better way of stating this is:
For any whole number n, either n or n + 1 is even.
Let f (x) = x2 + 1.
10
In the last example, x is a place holder. It doesnt require a proper introduction. However,
it would be better to write:
Let f (x) = x2 + 1 for all real numbers x.
If describing all the variables gets tedious, try not assigning any variables at all. The
following example clearly needs improvement.
Including units would make this clearer, but the description is still vague.
Let D(t) be the distance in miles at t hours.
Also, be careful that each symbol you use represents only one thing. This can actually
be more subtle than it sounds. The following example seems to be rather clear.
11
I think that the above example would be considered unobjectionable by most readers. It
looks very clear and understandable. The variable P is always standing for the wombat
population. However, notice that in the first paragraph, P is the wombat population in
general. In the next paragraph, P = 0.6272, the wombat population in 1992. And in the
last paragraph, P = 2. The meaning of P appears to be changing every time that it is
used. In the first paragraph, P represents the population at any time. In the other
instances, P represents the population at one particular time. The problem can be fixed
omitting some variables and adding others.
12
While in the above example, we can afford a little bit of sloppiness with the variables, in
more complex problems, this could be a source of potential trouble. When a symbol is used
to represent two different things (even, or perhaps especially, if those things are similar),
the reader (and the writer!) can become confused. A symbol used in two different ways is
not only confusing, but often results in incorrect mathematics!
Just as variables need to be introduced carefully, also be sure not to pull formulas out
of thin air. Tell the reader how you get each formula or what each formula means. Its not
very pleasant to get hit with formulas without any warning.
What should the reader look for in the graph? Why does the graph support the argument?
Be more specific.
13
A good graph should convey relevant and specific information to the reader. The
following graph is vague.
Graphs and diagrams need to be neatly drawn and clearly labeled. Indicate the scale on
the axes. You should point out significant graphical features.
Cooties infections versus time
No. of infections I (in thousands)
1000
900
800
700
600
500
400
300
200
100
10 20 30 40 50 60 70 80 90 100
If you draw a graph by hand, use a straight edge. You may want to generate your graphs
using a computer. Be careful though. Programs like Excel or Microsoft Office generally are
not good at generating mathematical graphs. You will more likely have success using a
math program like Maple.
Any diagrams you draw should also be carefully labeled. Be sure to label everything
that you refer to in your argument.
14
Epilogue
Writing mathematics is not the easiest thing to do. Writing mathematics is a skill which
takes practice and experience to learn. There are many resources here at Purdue Calumet
which are available to you to help you with your mathematical writing. Among these are
the Math Lab and the Writing Lab.
If you have not written mathematics much before, it may feel frustrating at first. But
learning to write mathematics can only be done by actually doing it. It may be hard at
first, but it will get easier with time and you will get better at it. Do not get discouraged!
Being able to write mathematics well is a good skill to learn, and one which you will keep
for a lifetime.
15
17
expressions,
functions.
y = x2 .
or
or
(x + 1)(y 2)
or
A function is a process that takes inputs and makes outputs. The symbol f denotes a
function, but f (3) does not; f (3) is a result or value of a function. So, do not refer to the
function f (3). You mean the value f (3) or the output f (3).
Exception. You may refer to the function f (x) instead of the value f (x). This usage is
technically incorrect, but accepted. The point is, x is a dummy variable here. You are merely
saying that f is a function because it applies to inputs, rather than referring to the specific
output when the specific letter x is input. Similarly, if z is your dummy variable, you may refer
to the function g(z) instead of the value g(z). However, in both cases it would be better
to say the function f or the function g. Or, leave out function and just say f (x).
Examples. Consider
A function is an equation. For instance, f (x) = x2 3.
No! A function is a process. A function is defined by giving an equation that provides a formula
in terms of a dummy variable, but the equation is not itself the function.
Example 2. What is wrong with:
Let f (x) = x2 . In
f(x) = f (2x1) f (3x+2)
(1)
for the first function f you replace the dummy variable by 2x 1 in the equation of
f , getting (2x 1)2 .
Answer: Two things. There is only one function f in (1). It is evaluated twice. Second, there
is no equation of f ; rather, there is a definition of f , the statement that starts with Let.
Correct rewrite: Let f (x) = x2 . In
f(x) = f (2x1) f (3x+2)
for the first evaluation of f you substitute 2x 1 for the dummy variable x in the
definition of f , getting (2x 1)2 .
But why use awkward word expressions like first evaluation of f when you can use brief
symbols:
S Maurer
S Maurer
S Maurer
Final Remark. Please refer to function notation, not functional notation. The adjective
form of (mathematical) function is function, not functional. This is because functional
already has another meaning in ordinary English. Functional notation means any notation
that works well, whether or not it has anything to do with functions, or even with math.
For more information, see the handouts from my manuscript Short Guide to Writing Mathematics.
Lab #1
Do experiment 3.6 " The Computer May Lie" on page 25 of your textbook, preferably
using the Mathematica notebook with the same title. (Alternatively, you may Use the
Function Iterator available at the Dynamical Systems and Technology Project web site).
Choose "generic" seeds, such as 0.836, not whole numbers. Write an essay describing
your findings, answering the questions posed in the Results section, and questions 1 and
2 on page 26. Please do not list or print out the entire orbit for each of your chosen
seeds. Rather, simply give the seed and a short description of the outcome (tends to a
fixed point, goes to infinity, etc.) In each case, the computer is not telling the whole story.
Do you see what else is happening? Comment on this in your essay after explaining what
the computer shows.
Please be sure to review the Lab Report Expectations.
The Dynamical Systems and Technology Project web site is the work of Professor Bob
Devaney at Boston University, and the Mathematica notebooks were written by
Sebastian Marotta.
18.091 Lab 1
The Computer May Lie
Alejandro Ochoa
May 19, 2005
Introduction
Methods
Some of the seeds were chosen by me, while others were randomly generated
by Mathematica using a function. Note that Mathematica is software capable
of doing rational arithmetics, which implies the format in which we type the
number matters. For example, typing 0.1 for a seed will make Mathematica
do all the computations with 1/10 as a oating point number, with limited
precision, while typing 1/10 will force Mathematica to make exact calculations
with that rational number.
For each function, on the table we listed in the rst column the seeds used, and
in the second column the observed behaviour of the orbits of the corresponding
seed. Seeds with ve or six signicant digits were generated by Mathematica.
Double quotation marks indicate the behaviour is the same as described above.
Plots of sample orbits are also included for the rst and third functions to better
illustrate the behavior of such orbits.
Behavior
Fixed point.
Apparently chaotic, bounded in [2, 2].
2. F [x] = x2 2.1,
Seed
1
1
0
0.25
1.985
0.185049
0.283494
0.643112
0.987201
0.252688
Behavior
Diverges to
Behavior
Eventually xed, converges to 0
Periodic
Eventually periodic
Discussion
For the function F [x] = x2 2 the experiment with the computer suggests that
all orbits are chaotic (see Figure 1), except we happened to select the point
x0 = 1 which is xed under iteration by F. However, we can algebraically
verify that the equation F [x] = x has another solution, namely x = 2, so there
are two xed points for this function that the computer would not have found
for us, for example, if it only sampled random numbers. We did not identify
any periodic points either, while we know there are innitely many from theory.
For the second function F [x] = x2 2.1 the computer strongly suggests
that all seeds in (2, 2) diverge to innity, but this is not the case because the
1 9.4
x0 =
,
2
and one of them, x0 1.0330 is in the interval we sampled. Hence, among
all of the points that diverge to innity, there is one that remains xed and was
not detected with the computer.
The function D is by far the most interesting case. Consider x0 = 0.1, which
we know has an exact representation as a rational number, namely x0 = 1/10.
The point 1/10 is eventually periodic, entering the 4cycle
1 2 4 3 1
, , , , , ...
5 5 5 5 5
after the rst iteration. However, the computer shows that x0 = 0.1 is eventually
xed, converging to zero. Looking at Figure 2, we can observe that 0.1s orbit
starts with a seemingly regular orbit as we expect it because it enters a 4
cycle. However, something happens around the 50th iteration, the orbit becomes
irregular and then drops to zero where it remains xed.
It is plausible that the dierence in the outcomes of the orbits of 0.1 and 1/10
is due to the fact that the result of a computation on a oatingpoint number
(the logical representation of decimals on a computer) is bound to lose precision
because oatingpoint numbers only store a nite amount of digits, while the
software that deals with rational arithmetics only needs to store integers and
has an unlimited capacity to do so, always giving an exact answer. Although
the loss of precision is very small, after 50 iterations of the function on 0.1 we
can observe the loss of precision built up to give a wrong answer which is not
reasonably close to the right answer.
For the rst two examples, we were led to an incorrect hypothesis because
we only sampled a nite amount of points. There are a dense and uncountable
amount of points which share the behaviour of the points we sampled in those
intervals. However, there are two and one points of interest respectively for each
F on the given interval, which display the very appealing property of remaining
xed under iterations of F, and the probability that the computer would ran
domly choose these points is close to zero. The third example showed us that
the loss of precision generated by storing numbers as oatingpoint numbers can
become signicant when analyzing orbits of seeds of chaotic systems.
Whenever we use computers to model dynamical systems, we must always
keep in mind the two caveats we discovered, namely, that we should not gener
alize the behavior of an interval of seeds by analyzing just a few seeds, and that
we should be careful when considering results from a large amount of computa
tions on oatingpoint numbers, given that these results are not guaranteed to
be neither correct nor approximately correct.
Lab 1
Jeremy Hurwitz
February 14, 2005
Introduction
The goal of this experiment was to discover the behavior of various functions under iteration
under dierent seeds using Mathematica. The rst function studied was the doubling function,
which was dened as F (x) = 2x (mod 2). The rest of the functions were the family of functions
F (x) = x2 c, where c is any positive constant.
In each case, I attempted to classify the orbits as xed, periodic, tending towards a certain
limit or chaotic. To this end, I ran twenty iterations of each function suing Mathematica to see if
a pattern seemed to be appearing. If unsure, I ran more iterations.
For testing the doubling function, I divided possible seeds into three groups. The rst group was
rational numbers, written in the form p/q, p, qZ+ . The second was rational numbers written as a
decimal expansion. The last group was irrational numbers, such as the square root of 2.
2.1
In the case of rational numbers written in fractional form, one of three results always showed up.
Zero returned the xed point orbit F (x) = 0. If q = xn , for some n, the orbit became eventually
xed at 0. All other orbits were periodic or eventually periodic.
2.2
Decimal inputs returned incorrect results due to rounding errors. Because the computer stores
these numbers in the form
a1 a2 a3
an
+ 2 + 3 + ... + n
(1)
2
2
2
2
after k iterations, the k th term becomes equivalent to 0 (mod 1). Therefore, the computer falsely
stated that all decimal inputs became eventually xed at 0. For example, x0 = 2/5 returned a
periodic orbit, whereas x0 = 0.2 became xed.
2.3
Irrational Numbers
Irrational numbers appeared to be chaotic, at least until the rounding errors compounded enough to
send the orbit to zero. When I calculated the orbit without rounding, the orbits remained chaotic.
F (x) = x2 c, c > 0
I started by testing c = 2. Once I had a feel for how the orbit behaved for dierent seeds, I started
varying c, attempting to nd a pattern that owed between all values of c. As with the doubling
function, I looked for orbits that tended towards periodicity or towards a specic orbit.
3.1
The rst obvious characteristic of the entire family of functions was that if the seed was outside
some specic interval, then it would
tend
towards innity. For c=2, the interval of convergence was
1+ 5 1 5
[2,2]. For c=1, the range was [ 2 , 2 ].
Theorem 1. Let F (x) = x2 c, c > 0. Let p =
1+ 4c+1
2
and q =
1 4c+1
.
2
Then if
4c + 1 and q 2 c = q,
F (x0 ) = q 2q + = p
4c + 1 2q + 2 = p + ( 1)( 4c + 1 + )
3.2
In general, there always existed kcycles which could be found by solving F k (x) = x. For values
not lying on one of these cycles, there were three cases.
When 0 < c < 1, the orbit tended to converge to the xed point p = 1 24c+1 , an attracting
xed point. Note that the other xed point is repelling, so any point within the convergent interval
has to go towards this attracting point.
When c = 1, all orbits tend towards the 2cycle (0,1,0,1,...). As long as |x0 | < 1.68, the computer
could not tell the dierence.
The last case was 1 < c < 2. Within this interval, all orbits appeared to become chaotic.
Summary
The most important result of this experiment was the rounding issue that arose when a decimal
was used as the seed of the doubling function. Unless very careful analysis is used as to how the
computer is treating the number and how errors are propagated through the iterations, computer
results must be taken with a large grain of salt.
From a standpoint of generating and studying chaos, the main result was how easily very sim
ple, predictable systems became chaotic. The doubling function, which was completely predictable
for all rational numbers, showed no patterns for irrational numbers. For the family of functions
F (x) = x2 c, orbits were predictable (excluding individual kcycles) for all values of c outside the
range (1,2]. Within that range, unless the orbit was periodic, it was always chaotic.
Input
Correct Orbit
Computers Output
0.2
periodic
goes to zero
1/5
periodic
periodic
1/9
periodic
periodic
0.23
periodic
goes to zero
5/17
periodic
periodic
0
xed point
xed point
1/2
eventually xed
eventually xed
1/10
eventually periodic eventually periodic
1/4
eventually xed
eventually xed
x R Q (10 trials) no pattern
no pattern
Table 3: x2 2
Input
2
0
2
1
1
1/2
2
3
8 other tests
Eventual Behavior
Fixed
Eventually Fixed
Eventually Fixed
Fixed
Eventually Fixed
Chaotic
eventually xed
eventually xed
chaotic
Data
Table 4: x2 3
Input
2
3
0
2
1+ 13
2
Eventual Behavior
Periodic
to innity
to innity
periodic
xed
periodic
c
4
1.5
1
.5
behavior
2 xed points, all other tests to innity
chaotic or towards innity
towards 2cycle or towards innity
towards .366025 or towards innity