[English]
What is the connection between a faucet that
tastes the whole mandelbrotte, a
rabbit population, convection
thermal in a fluid and activation
neurons in your brain? It is
this simple equation.
Let's say you want to model a
rabbit population with X rabbit this
year, how many will you have per year
next? Well the most model
simple as I can imagine consists
simply to multiply by a certain
number the growth rate R which
could be say 2 and that
would mean that the population
would double every year. The problem
is just the number of rabbits
would continually increase
exponential.
So I can add the term 1 - x for
represent the constraints of
the environment. And here I imagine that
population
theoretical maximum. So it goes from 0 to
1. And as she approaches this
maximum, this term tends towards zero and this
limits the population.
So here is the logistics method.
Xn + 1 is the population of the year
next and this year's Xnulation.
And if you plot the population of
next year depending on that
of this year, you see that it is
just an inverted parabola. It is
the simplest equation you
can make which has a loop of
negative feedback. More the population
gets bigger here, the more it will be
small the following year. So let's try
an example.
Let's say we are dealing with a
group of rabbits particularly
tactive. So R is equal to 2.6.
Let's choose an initial population of 40
% of the maximum, i.e. 0.4
multiplied by 1 - 0.4 we obtain 0.624.
So the population increased first.
>> But what really interests us,
this is the long-term behavior of
this population. So we can put
this population in the equation. And
to go faster, we can actually
type 2.6 times answer x 1 MB answer.
We obtain 0.61. So the population has a
slightly diminished. Press again
0.619 0.613
0.617
0.615
0.616
0.615.
Continuing to press enter, the
population hardly changes. She
has stabilized as in nature where
populations remain constant only in
gasoline and death balance out.
I now want to make a graph of
this iteration. You see here that she
reaches an equilibrium value of 0.615.
What would happen if I changed the
initial population? I'll just
move this cursor here. And what you
see it's only the first years
change
but the equilibrium population remains
same.
So we can essentially ignore the
initial population.
So what really interests me is
how does this equilibrium population
varies as a function of R the rate of
growth.
Well as you can see, if I
decreases the growth rate, the
equilibrium population decreases. It must have
meaning. And in fact if R goes below
by 1, well then the population drops
and eventually disappears.
So what I want to do is plot
another graph where on the axis of
abscissa, I have R, the rate of
growth. On the disordered axis, I
plots the equilibrium population, that
obtained after many
generations.
For low air, populations
always turn off. The balance is
therefore zero.
When R reaches 1, the population
stabilizes and the higher R is, the more the
equilibrium population is high.
So far, so good. But
Now here's the weird part.
Once R exceeds 3, the graph
divides into 2.
Why?
>> What's happening? Well it doesn't matter
how many times you iterate the equation,
it never stabilizes on a
single constant value. Instead,
it oscillates back and forth between
two values.
>> One year, the population was more
high, the following year, it is more
bass then the cycle repeats.
The cyclical nature of populations is
also observed in nature. THE
number of rabbits varies from year to year
the other, sometimes more, sometimes less. HAS
As R continues to increase, the
fork moves apart then each
divides again.
Now, instead of oscillating between
two values, the populations cross
a cycle of 4 years before repeating itself.
Since the duration of the cycle or period has
doubled, we call these bifurcations
by doubling of period. By increasing
R, we observe more bifurcation by
period doubling. They arrive from
more and more quickly leading to
cycles of 8 16 32 64. Then such that
when R reaches 3.57 at chaos. There
population never stabilizes. She
fluctuates as if at random.
This equation was one of the first
methods for generating numbers
random on computer. That
made it possible to achieve the unpredictable
of a deterministic machine.
No pattern or repetition here. If you
know the initial conditions
exact, you could calculate
precisely the values. So we
treats as pseudo numbers
random. One could believe
the equation will remain chaotic, but
the order returns with the increase of
A.
There are these behavior windows
stable periodic in the midst of chaos.
For example, if R is 3.423, we observe
a stable cycle of 3 years and as
R continues to increase, it divides into
6 12 24 and so on before
return to chaos. In fact, this
equation has periods of all
length 37, 51, 1052, whatever you
want if you just have the right one
value of R.
This bifurcation diagram looks like
a fractal.
The characteristics and the great
scale repeats at larger scales
small.
And indeed, if you zoom in, you see
that it is in fact a fractal.
The best known fractal is the set
of Mandelbrot.
The twist is that the diagram
of bifurcation is part of the set
of Mandelbrot.
Does it work well? Little
recall on the Mandelbrot set, it
is based on this iterated equation. SO
the way it works is that
you choose a number C, regardless
what number in the complex plane, then
you start with Z = 0 and then
you iterate this equation over and over
again. If it diverges towards infinity,
then the number C is not part of
the whole but if this number remains finite
after an unlimited number of iterations,
then it is part of the set of
Mandelbro.
Let's try for example C = 1. So we have 0
squared + 1 which gives 1. Then 1
squared + 1 = 2 squared + 1 = 5
square + 1 = 26. So we see enough
quickly than with c = 1, this equation
will diverge. So the number 1 does not
not part of the Mandelbrot set.
And if we try c = -1, well we have 0²
- 1 = -1 - 1² - 1 = 0. And so we
returns to 0² - 1 = -1. We therefore see that
this function will continue to oscillate
between -1 and 0. So it will remain finite.
And so c = -1 is part of the set
by Mandel Brot. Normally, when we
sees images of the entire
Mandelbrot, it just shows the
boundary between the numbers which make
this iterated equation remains finite and
those who make it diverge, but it does not
doesn't really show how these numbers
remain finite. What we did here,
is that we actually iterated this
equation thousands of times then we have
plotted on the Z axis the value that takes
effectively the iteration. So if we
look aside, what you will see in
done, this is the bifurcation diagram
which is part of this set of
Mandelbrot. So what's going on
really here? Well what does that give us?
shows is that all the numbers in
the main cardioid ends up
stabilize on a single value
constant. But the numbers in this
main bulb, they end up good
oscillate between two values. And in
this bulb, they oscillate between four
values. They have a period of 4 then 8
then 16 32 and so on. Then we
reaches the chaotic part. The part
chaotic of the bifurcation diagram
found here on what we call
the needle of the Mandelbro set,
where the Mandelbrot set becomes
very fine. And you can see this
medal here which looks like a version
smallest of the Mandelbro set
integer. Well what does this show us?
is that all the numbers in the
main cardioid ends up
stabilize on a single value
constant. But the numbers in this
main bulb they end up
oscillate between two values and in this
bulb oscillate between four values.
They have a period of 4 then 8 then 16
32 and one so on. Then we reach
the chaotic part. The chaotic part
of the bifurcation diagram is found
here on what we call the needle of
the Mandelbrot set. where
the Mandelbrot set becomes very
end and you can see this medal
here which looks like a more
small of the Mandelbrot set
integers. Well this corresponds to the
stability window in the diagram
bifurcation with a period of 3.
Now the bifurcation diagram
only exists on the real line because
that we only put real numbers in
our equation. But all these bulbs
outside the main cardioid and
well they also have cycles
periodic for example 3 4 or 5. And
so we see these ghostly images
repeated if we look on the axis z.
done, they also oscillate between these
values.
personally, I find it
} extraordinarily beautiful, but if you
are more pragmatic, you
may ask this
equation really models
animal populations? And the answer
is yes. Especially in
controlled environments that
scientists have set up in
laboratories. What I still find
more incredible, this is the way
This simple equation applies to a
wide range of scientific fields
totally independent
others.
[Music]
The first big confirmation
} experimental is awarded to a
specialist in fluid dynamics
named lipcha. He made a box
rectangular containing mercury and
used a weak gradient of
temperature to cause the
convection. Just fluid cylinder
turning in the opposite direction inside
of his box. That's all that the box
could contain and of course it does not
could not look inside for
See what the fluid did. So he
measured the temperature using a
probe placed at the top. He observed a
regular and periodic peak of
temperature. It's like when
The logistical equation converges towards a
only value.
but by increasing the gradient of
temperature, an oscillation in half
of the initial frequency appeared on
These rolling cylinders. The peaks of
temperature were lower.
they alternated between two heights
different.
he had reached period 2 and in
continuing to increase the temperature,
He observed a doubling of a period at
new. Now he had four
different temperatures before the
cycle is not repeated. Then 8. It was a
fairly spectacular confirmation of the
Theory in an experience
beautifully designed.
but it was only the beginning.
scientists studied the reaction
of our eyes and eyes of the salamander
to flashing lights and they have
discovered a phenomenon of doubling of
period. Once the light has reached
a certain rhythm of flashing, our
Eyes no longer react no longer act
than one on two flashes. It is
incredible in these articles to see
appear the fork diagram,
even if it is a little blurry because it
comes from real world data.
scientists gave a
medicine to rabbits causing a
cardiac fibrillation. I imagine he
thought that there was too much rabbit
outside. Finally, if you don't know this
that is fibrillation, is when your
heart beats extremely
irregular and hardly pumps
blood. And if you don't intervene, you
meurs. They discovered that by going
towards fibrillation, they found the
Route du double period leading to
chaos.
The rabbit first had a beat
periodic then a cycle of two
close beats. Then a cycle
of four different beats before
start again and finally a behavior
aperiodic.
The remarkable aspect of this study is
surveillance in real time of the heart and
The use of chaos theory
to determine when to administer
electric shocks to restore the
periodicity what they have succeeded. SO,
they used chaos to control
a heart and find a more way
} intelligent to administer shocks
} electric to make it beat
normally again. It's really
incredible. And then there is the question
of the tap that tastes. Most of
of course consider the taps
that taste like very
regular and periodical. But a lot
of research showed that once
The flow increases a little, we get a
doubly period. So now,
The drops fall two by two.
of a simple tap that tastes, we can
generate chaotic behavior in
modifying the flow, which leads to
Ask what a tap is really.
well, there is water under pressure
constant and a size opening
constant. And yet, what you
get, it's a drop
chaotic.
so it's a really chaotic system
simple. You can experience this
at home. Go open a right tap
a little and see if you can
get a periodic drop
at home.
The bifurcation diagram appears at
so many different places than that
is starting to seem strange.
Now I want to tell you something
something that will make it even more
strange. There was this physicist
Mitchell Figenbum who studied the moment
where the bifurcations occur. He has
divided the width of each section of
bifurcation by the next one and he
discovered that this relationship converges to
this number 4,669.
what we now call the constant
by Feigenb.
Bifurcations are occurring more and more
faster but in a ratio that
tends towards this fixed value and no one
doesn't really know where this comes from.
constant. It does not seem to be linked to
no other known physical constant,
so that it constitutes in itself
a fundamental constant of nature.
What's even crazier is that he
is not even necessary for the equation
takes the particular form that I
I showed earlier. any equation that
has a single bump. If you
iterate it the same way we do
did it, so you could
use xn + an equal sine of x by
example. If you iterate it again, again
and again, you will also see
bifurcations. Not only that, but
the ratio of when these bifurcations
occur will have the same factor
of scale. 4,669.
Any iterated single-hump function
will give you this constant
fundamental. So why? Well
we speak of universality because it
seems to be something
fundamental and very universal in this
process, in this type of equation and
in this constant value. In 1976,
the biologist Robertm published a
article in nature about this
equation precisely.
This caused a revolution among
those who have studied this subject. This
article has also been cited
thousands of times and in this article, it
calls for teaching
this simple equation to students because
it offers a new intuition on the
way in which simple things,
simple equations can generate
very complex behaviors.
And I still think that today we
doesn't really teach that way.
I mean, we teach equations
simple and simple results because
that these are the easy things to do
and these are the ones that seem logical.
We are not going to sow chaos among the
students, but maybe we should
maybe we should sow some at
less a little. And that's why I
am so excited about the
chaos and this equation because
frankly, how did I reach
37 years without ever having heard of
of Fagenbom's constant?
Since reading James' book
Gleek, Koo, I wanted to make videos
on this subject and now I'm getting started
finally and I hope to do justice to this
topic because I found it
incredibly fascinating and I hope
you too.