A course

Mathematics and
Technology
Mathematics is a living science,
everywhere present in science and
technology

The teacher should have

experienced how science develops
in the real world. He(she) can then
show it.
The scientist:
He(she) asks questions;
He(she) dares to say: I do not know.
He(she) has an open and critical mind.

Helping discovering the

power of the mathematical
method:
Modelling
Problem solving
Mathematical sophistication
Use of computer
Some messages through the course :
Mathematics are useful and constantly
developing around us.
The questions Why? and What is it
useful for? should be encouraged and
deserve an answer.
The beauty of mathematical constructions
Mathematics are much more than numbers

Most subjects treated are too advanced so that

preservice teachers following the course can hope bring
them directly to classroom. The purpose of the course is
rather to teach them how to prepare such kind of
material.
 Mathematics and
technology
New course since winter 2001.
Most students are preservice
secondary school teachers.
Purpose: Discover mathematics
present in everyday technologies
 Joint creation with my colleague:
Yvan Saint-Aubin
Yvan is a physicist and I am a mathematician.
We knew very little of the material of the course
when it was created. We now have enough
material for at least two courses.
The game is to take some technologies
into pieces, to dismantle them in order to
discover and explain the mathematics
that make them work.
We play the game to prepare the course.
We try to teach the students to do the
same.
Description
Two formats:
Flashs-science (1 hour)
More elaborate subject
- 1 week = 3 hours plus 2 hours of
exercices
- or two weeks on one subject
The lectures are of two different types
elementary parts (subject matter for exams)
conference type lectures on advanced parts
Evaluation
Two exams with open book and
personal notes. Non cumulative
contents
A session project on an application of
mathematics (by teams of two, if
possible; by larger teams (4-6)
otherwise )
A half-hour oral presentation of the
project
 The exercices
We have spent a lot of time writing
interesting exercices that make
the students practice modelling
and review their elementary maths
Finding appropriate exam
questions is not a trivial task
A few examples below
 A book for the course
Mathematics and technology (september 2008)
C. Rousseau and Y. Saint-Aubin, Springer-Verlag

Mathmatiques et technologie (october 2008)

C. Rousseau and Y. Saint-Aubin, Springer-Verlag
 Flashs-science
Antennas and radars are parabolic.
Why? (Geometric definition of conics)
Computer vision: calculating the
position of one object from its
position on two photos (The
parametric equations of lines in 3-
dimensional space)
 Covering a territory with antennas for
a mobile phone network (Euclidean
geometry)
 The corresponding exercice
at the exam
We fill a large planar region with nonoverlapping disks of radius r. We
use two methods: in the first method we place the centers of the disks
on a square network and in the second method we place them on a
regular triangular network of equilateral triangles.

(a) (b)
Which method gives the denser filling? Suggestion:
compute the proportion of each square covered by portions
of disks in case (a) and the proportion of each triangle
covered by portions of disks in case (b).
 Physics: unifying the laws of reflection and
refraction. The laws of nature follow
optimization principles. Applications: short
waves, optical fiber
A short look in the architecture of
computers describing logic circuits
The regular tiling of the sphere with twelve
spherical pentagons
 Vorono diagrams (Euclidean
geometry)
 More elaborate subjects
Positioning in space: GPS, GPS signal,
cartography, localization of thunderstorms
(Geometric locus, differential geometry, theory of
finite fields)
How is a musical CDengraved: why 44100
numbers per second? (Elementary Fourier
analysis)
Public key cryptography (Elementary number
theory: congruences)
Error correcting codes: Hamming codes and
Reed-Solomon codes (Linear algebra, finite fields)
Image compression: iterated function systems
(Affine transformations of the plane)
The JPEG format (.jpg) (Elementary Fourier
analysis)
 Robots (Rotations in 3-dimensional space,
change of reference frame)
Friezes and tilings (Symmetries linear
algebra)
Google and the Pagerank algorithm (Markov
chains linear algebra)
The skeleton and the gamma-knife surgery
(Geometry)
Turing machines and DNA computers (The
hierarchy of functions starting from the basic
ones)
Random number generators (Finite fields)
Calculus of variations (Multi-variable calculus)
Sparing and borrowing money
 Some students
projects
(a list on my webpage)
Rollercoasters
The search of boundaries in a photo
Morphing IMAGES
Text compression
Mathematical morphology in treating
images
 Benford law of significant
digit
 How to complete the hole in
Eschers painting
Print Gallery
 Polyhedra and fullerenes

Carbone 60
Truncated
icosahedron
 Vorono diagrams and Delaunay
triangulation in image analysis
Sphere packings and honeycombs
The best skateboard track
Other cryptographic methods
Reed-Mller error-correcting codes
Knots and the action of enzymes on DNA
Digital fingerprintING
Image compression: from fractals to
practical applications
 Penrose tilings
The seasons, the locus of the sunrise and
sunset at a given date, the length of day at a
given date,
Calculation of astronomic distances, from
the ancient Greeks to now
The eclipses
The shape of sand dunes
Phyllotaxy (how to explain spirals in
sunflowers, etc.)
Population growth under constraints
Mathematical modelling of epidemics
Chaos
A remarkable property of
the parabola

All rays parallel to the axis are

reflected to a single point.
Applications: the shape of
many objects among which
Telescope mirrors
 Solar furnaces
Parabolic antennas
Radars
 The corresponding property
of the ellipse

Any ray issued from one focus is reflected to

the other focus.
 Applications: mirrors,
accoustic phenomena
Elliptic mirrors for instance behind
the lamp of a cinema projector

Accoustic phenomena: for instance Par

 Google and the PageRank
algorithm
A search engine 1 1
0 1 0
that does not 2 3
order entries 1 1
1 0 0
properly is 3 3
useless. 1 1
0 2 0 03

0 0 0 0 1
3
1
0 0 0 0
3
Where are we after two
clicks?
1 1 1 11
0
2 6 6 18
2 4 1
0 1
3 9 9
1 5 1
2 0 18 0
6

0 0 1 0 0
9

0 1 0 0
1
6 9
Where are we after n clicks?

0.29 0.29 0.29 0.29 0.29

0.39 0.39 0.39 0.39 0.39
0.22 0.22 0.22 0.22 0.22
Why
0.02 0.02 0.02 0.02 0.22
0.07 0.07 0.07 0.07 0.07

?

 Order of pages

0.29 0.29 0.29 0.29 0.29

0 . 39 0. 39 0 . 39 0. 39 0 . 39
0.22 0.22 0.22 0.22 0.22

0.02 0.02 0.02 0.02 0.22
0.07 0.07 0.07 0.07 0.07

B, A, C, E, D
 Image compression
The easiest way to store an image
inside the memory of a computer is
to store the color of each pixel.

This requires an enormous quantity

of memory!

Can we do better?
 Lets suppose we have drawn a city:

We store in memory the line

segments, circle arcs, etc, which
approximate our image.
We approximate our image by
known
To store a line segment in memory
it is sufficient to store:
the two endpoints of the line
segment
a program explaining to the
computer how to draw a line
segment with given endpoints.

The geometric objects are our

alphabet.
 How to store more complex
images, for instance
landscapes?
We use the same
principle but we
enlarge our
alphabet:
We approximate
our landscape by
fractals, for
instance the fern.
We store in memory a program to
draw the fern. Such a program on
Mathematica
m=15000
L[n_]:=If[1<n<87,2,n]
H[n_]:=If[86<n<94,3,L[n]]
K[n_]:=If[n>93,4,H[n]]
R=Table[K[Random[Integer,{1,100}]],{m}];
F[1,x_,y_]:=0
G[1,x_,y_]:=0.16*y
F[2,x_,y_]:=x*0.85+y*0.04
G[2,x_,y_]:=-x*0.04+y*0.85+1.6
F[3,x_,y_]:=x*0.2-y*0.26
G[3,x_,y_]:=0.23*x+0.22*y+1.6
F[4,x_,y_]:=-x*0.15+y*0.28
G[4,x_,y_]:=x*0.26+y*0.24+0.44
x[1]:=0
y[1]:=0
Do[{x[n+1],y[n+1]}={F[R[[n]],x[n],y[n]],G[R[[n]],x[n],y[n]]},{n,1,m}]
T=Table[{x[n],y[n]},{n,m}];
ListPlot[T, AspectRatio->1, Axes-> False]
 Principle for drawing the
fern
The fern is
the union:

of a stalk
of three copies
of the initial
fern
We can reconstruct the fern
from 4 affine
transformations:
the transformation T1
which sends the large
fern to the fern minus two
branches,
the transformation T2
which sends the large
fern on the left branch,
the transformation T3
which sends the large
fern on the right branch,
the transformation T4
which sends the large
fern on the stalk.
 In order to reconstruct the fern, it
suffices to store in memory this
information!
Algorithm:
We take a point P on the fern.
We choose at random i in {1,2,3,4} and we plot
P1 = Ti(P).
We choose at random i in {1,2,3,4} and we plot P2
= Ti(P).
Etc...

This method is called Iterated function systems .

It works because the fern is self similar.
 Why does it work?
Lets look at the
Sierpinski carpet:
It is a union of three
Sierpinski carpets.

Let us start with a

square and iterate
a construction
algorithm
This works with any initial set!
Lets try another one:
 In practice
Coding: We replace any small
square by the image of
a similar larger square
under a homothety of
ratio composed with
one of 8
transformations:

Identity plus 3
rotations
4 symetries

We adjust contrast.
 Example

First iterate Sixth iterate

Some exercices
 The GPS (Global positioning system)
fully operational since 1995

Network of
orbiting
satellites
whose
position is
known
 The receptor measures the
travelling time t of a signal
from one satellite to the
receptor.
The distance from the
satellite to the receptor is d
= ct
c: speed of light
The points located at a
distance d from a satellite
are on a sphere of radius d,
with center at the satellite.
 The intersection of two
spheres is a circle:

The
intersection
of three
spheres is
two points.
One of them
is excluded
Hence, if we know the
because it is
travelling time of the signals of
non
threerealistic.
satellites to the receptor
we know the position of the
 This is the theory
In practice the satellites have atomic clocks
perfectly synchronized.

The receptor has a cheap clock.

We have a fourth unknown: the shift between

the clock of the receptor and the clocks of the
satellites.

We then need to measure the travelling time

of a signal from a fourth satellite.
4 measured times
4 unknowns
Theshift
between
clocks
The three
coordinates of
position
With this method we get a precision of 20
 Applications of the GPS
Finding ones way in nature
Drawing a map
Managing a fleet of vehicles
Measuring Mount Everest and
observing its growth
Helping blind people
Find ones way on the road
Landing a plane in the fog
GPS are a reference of time!
Electronic equipments can be synchronized
with the help of GPS.

Hydro-Qubec uses this method to

synchronize its lightnings detectors. Once
thunderstorms are localized, one can
reduce the current through lines passing
through zones of thunderstorms so as to
minimize the risk of breakdown of the
electrical network, in case one transit line
receives a lightning.
 A related exam question
Meteorites regularly enter the atmosphere, rapidly
heat up, disintegrate, and finally explode before
hitting the surface of the Earth. This explosion
generates a shock wave that travels in all directions
at the speed of sound v. The shock wave is detected
by seismographs installed at various locations on the
surface of the Earth.

If four stations (equipped with perfectly synchronized

clocks) note the moment that the shock wave arrives,
explain how to calculate both the position and time of
the explosion.
 Signal of the GPS
Shift-register:
Example: we take (q0, q1, q2, q3 )=(1,1,0,0)
000100110101111
001001101011110
010011010111100
100110101111000
001101011110001
011010111100010
110101111000100
101011110001001
010111100010011
Why?
101111000100110
011110001001101
111100010011010
111000100110101
110001001101011
100010011010111
Random number generators
Consider sequences of 0 and 1:
0 and 1 must each appear with
probabilty .
All sequences of length 2 must each
appear with probability .
All sequences of length n must each
appear with probability 1/2n.
Theorem: in the sequences of period
2n 1 generated by the shift-register:
1 appears 2n-1 times and 0 appears
2n-1 1 times,
Each sequence of length 2
appears 2n-2 times except 00 which
appears 2n-2 1 times
Each sequence of length r appears
2n-r times except 00 which
appears 2n-r 1 times, for r <= n.
 Error correcting codes
Principle: we lengthen the message in a
redundant way. This allows to correct some
errors.

Example: We repeat each bit three times. We

want to send 0.
We send 000.
If we receive 000 we decode 0
100 we decode 0
010 we decode 0
001 we decode 0
We have corrected 0 or 1 error.
However

If we receive 110 we decode 1

101 we decode 1
011 we decode 1
111 we decode 1
And the transmission is erroneous.

An error correcting code is efficient if there

are few errors.

This code is not economical: a word of 4

bits is lengthened to 12 bits and we may
only be able to correct one error.
 We can do much better
Hamming code:
We want to send a 4 bits word: u1, u2, u3, u4
We send a 7 bits word. We add (mod 2):
u5 = u 1+ u 2 + u 3
u6 = u 2+ u 3 + u 4
u7 = u 1+ u 2 + u 4
This code can correct one error.
u1 erroneous: u5 and u7 incompatibles
u2 erroneous : u5, u6 and u7 incompatibles
u3 erroneous : u5 and u6 incompatibles
u4 erroneous : u6 and u7 incompatibles
u5 erroneous : u5 incompatible
u6 erroneous : u6 incompatible
u7 erroneous : u7 incompatible