Now let's move R up to the value of 3.1
Now we're going to see something slightly different.
Ok. So I'm going to 'go' several times
and what you're going to see is
instead of settling down at a fixed point,
you're going to see an oscillation
between two different values
when x sub t is .558 x sub (t+1) is .7646
Now if I click go again, those two values switch.
And so on and so on forever.
This is called a periodic attractor with the period two.
It's called period 2 because it repeats itself every two time steps.
Now I need to define another term for you
which is the term "state."
The state of the system, in this case, is defined as
a value for x sub t and x sub t+1. That is a point on this graph
That's a state of the system.
So what you can see is, if you watch the point on the graph,
the system is oscillating between two different states.
So this is called a periodic attractor of period 2.
Now, suppose I move R up to 3.2
Set up. Go. And I keep clicking on Go.
What I'm going to see
is again, this same kind of oscillating dynamics
but the exact points that the dot is oscillating between
are different than for the previous value of R.
So this system, with R set to 3.2,
has the same kind of dynamics, the same
kind of periodic dynamics with period two
but the exact values of those two states
that it oscillates between are different than for 3.1
Now let's move R to 3.50
and Go. And we'll see, actually, a different kind of behavior
when the system finally settles down
What we're going to see
is an oscillation, but this time
it turns out that the oscillation is between four different states of the system, not two.
So let me see if I can see that by looking at the blue dot
One. Two. Three. Four.
One. Two. Three. Four.
It's repeating itself.
So this is a periodic attractor with period 4.
So the period has doubled.
And, if you kept doing this -- moving R up by small amounts --
you would find, at some point, the period would double again
that we'd get a period eight attractor
and then a period sixteen attractor
and then a period thirty two attractor
and so on and so on until finally
we reach a state that
we no longer have a periodic attractor any more.
So we can see this when we move R to the extreme
case of 4. And I'll set up and go.
And what you'll see is that the system
starts looking quite random.
So if your growth rate is 4, it turns out that
when you start iterating the system,
it becomes very hard to predict
what the state of the system is going to be
at a later time step without actually running the system
through the whole set of iterations.
So it doesn't settle down into any obvious ordered pattern.
Instead, as it turns out, this is an example of chaos.
Now remember earlier I said that chaos means
sensitive dependence on initial conditions.
So we can see how that works by opening the model called
sensitivedependence.nlogo
This is a model that is similar to our previous model
except it allows us to see the effects of sensitive dependence
on initial conditions by letting us plot two initial conditions and their dynamics.
So here I have R = 4, x0=.2 as before, but now I have another x0
an x zero prime.
So you can think of this as a new initial condition
that's going to be run simultaneously
with this other initial condition although
they're not affecting one another.
Each one is just obeying the mathematics of the logistics map.
And I've set them almost equal. The only difference is that
x0' has a one out in the fifth decimal place.
So when I do 'set up'
you actually see that there's a red dot,
which represents this x0'
and a blue dot, which represents x0.
But the blue dot is being hidden
because it is exactly under the red dot
because they are almost equal.
And the value here, .64, for time step one
is almost equal to the value for time step one over here.
So let's do Go and you can see the two points travel together.
You still can't see the blue point
In this chaotic way... but after a number of time steps
here only sixteen, they start to separate.
And you can see, when I plot this out a little bit more,
Look at this plot here.
The blue plot and the red plot start to be very different
and end up looking very uncorrelated.
So what this is showing is that, even if we start out with very similar initial conditions, almost equal
after some number of time steps,
the behavior of these two different systems
will look very, very different.
Now suppose we didn't know that there was a .1 out here in the fifth decimal place.
That's very possible.
And we were scientists trying to make a prediction
Well, should we predict that the system is going to end up here?
or end up here?
after forty four time steps?
Well it's impossible to know and that's echoing that quotation from Poincare
which said, "prediction becomes impossible."
The significance of this example was stated very eloquently
by the biologist Robert May in 1976 in his paper on the logistic map
when he says, "the fact that the simple and deterministic
equation (that is the logistic map) can possess dynamical trajectories
which look like some random noise, has disturbing practical implications.
It means, for example, that apparently erratic fluctuations
in the census data for an animal population
need not necessarily betoken either the vararies of an
unpredictable environment or sampling errors,
they may simply derive from a rigidly deterministic
population growth relationship." (That is the logistic map).
"Alternatively, it may be observed that in the chaotic regime"
(in our case, this was with R = 4) "arbitrarily close initial conditions can lead to trajectories which,
after a sufficiently long time, diverge widely.
This means that, even if we have a simple model in which
all the parameters are determined exactly,
long-term prediction is nevertheless impossible."
And that's a nice echo of Poincare's "Prediction becomes impossible" quotation.
This is yet another representation of the dynamics of the logistic map
This is a bifurcation diagram and here's what it shows.
On the horizontal axis, it shows R
and we looked at the logistic map's behavior for several of these values,
on the vertical axis, it shows x -- that is the attractor that is reached by the system given a particular value of R.
For example, when we set R to 2.8 and iterated the logistic map until it reached the attractor,
we found that the value that it reached was a little bit more than six.
Sorry point six.
We also see that at certain values of R,
the dynamic changes abruptly. For example,
near R = 3, the dynamic changes from being a fixed point to
a period two attractor. And then as we raise R, we still get
a period two attractor, but just with different values in the
attractor.
Similarly, near R = 3.4, you see a shift from period two to period four.
That's what these four values represent.
And then to period eight.
You see this branching structure until finally,
somewhere near R = 3.55, the dynamics stop being periodic and become chaotic.
This value of R (and I'll tell you what it is a little bit later)
is called "the onset of chaos."
So we've seen in this subunit
how the logistic maps works
and how it exhibits what's called "a period doubling route to chaos."
In the next subunit, we'll see something surprising
about the specifics of this period doubling route.
And see that, even though chaotic systems like the logistic map
are not predictable in detail, there are certain universal properties that they all have.