MATLAB Assignment 4 (25 points) due Monday, August 8, 2011
We are going to investigate the Lorenz Attractor using MATLAB.
The Lorenz attractor is an example of a dynamical system. A
dynamical system describes the position of a point in space
(3-dimensional space in this example) over time. So it describes how
a particle might move or how fluid might flow. The Lorenz attractor
is derived from a simplified system to describe convection rolls in
the atmosphere.
The Lorenz attractor is governed by a system of 3 first order
differential equations, with three constants in them. Given the
values of the constants, and a starting value, the evolution of the
Lorenz attractor is deterministic. You can calculate where the
particle will go at every time.
The Lorenz attractor is called an attractor because for
given values of the three constants, the trajectories given by the
equations after a long period of evolution are localized in a bounded
region of space. I.e. the trajectories are attracted to one region of
space.
For certain values of the constants, the behavior of the Lorenz
attractor is also chaotic. This means that if you start two
particles from nearly identical initial conditions, after a long
enough period of time, their trajectories will no longer be close
together.
The chaotic behavior of the Lorenz attractor makes it a strange
attractor and a fractal. A fractal has self-similarity in
a technical mathematical sense on all scales. The types of
trajectories you see in the strange attractor are repeated over all
time scales.
Do a Google Search on the Lorenz Attractor. There is a good article
on it on Wikipedia. You may not understand all of the article, but you can
understand most of it.
- You will need to hand in a cover sheet with at least a one
paragraph description of the Lorenz attractor and what you are doing
in this assignment. I gave you some information above, and there is a
lot more information available on the web. You need to include the
equations to describe it an the values of the parameters you will use
to get the strange attractor.
- In your cover sheet, write a second paragraph describing
the butterfly effect, and tell whether or not you see it in
your subsequent work on Matlab with the strange attractor.
Hint: Google and Wikipedia are your friends, but don't
feel the need to restrict yourselves to these. There are many other
sources of information out there.
- Identify the system of first order differential equations that
governs the attractor. These need to be included on your cover sheet.
You are going to program these equations into MATLAB.
- Identify the values of the three constants which give the
strange attractor. This information also goes in your cover sheet.
- Using ode45 in MATLAB with initial conditions at or close
to (1, -1, 1), find two trajectories that start from nearly the same
point, and, if you can print in color or in a black/white/greyscale
way, use plot3 to show that you get two very different
trajectories from the two very close but not the same starting values.
- Help for plot3 is available in MATLAB (type help plot3),
and there are some instructions here:
MATLAB Instructions/Basics/Help
- Using ode45 and ode15s in MATLAB with initial
conditions (1, -1, 1), over a time scale from 0 to 100 show that you
get two trajectories that start from same point but that
give very different trajectories. If you can print in color
or in a black/white/greyscale way, use plot3 to show that you
get two very different trajectories from the two ODE solvers with
the same starting values. (I believe ode45 gives the better
plot here.) You will want to use the command
t = linspace(0, 100, 10000);
to get your time values; you can plug t directly into ode45 or ode15s
- You will also hand in a copy of the function you use
with ode45 and ode15s to model the dynamical system.
- While you've got the plot on your computer, rotate the plot you
get from ode45
around. Can you see how the Lorenz attractor actually does look like
a butterfly?