16.12.2012 Views

Computer Algebra Recipes

Computer Algebra Recipes

Computer Algebra Recipes

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

8.5. RECONSTRUCTING AN ATTRACTOR 351<br />

There is no discernible pattern in the ¯gure, unlike the situation in the previous<br />

recipe. By plotting other pairs, or even triplets, of numbers, one can conclude<br />

(not surprisingly) that there is no underlying chaotic attractor here.<br />

PROBLEMS:<br />

Problem 8-28: Rolling a die<br />

Consulting Maple's Help, produce a random set of 500 numbers from the positive<br />

integers one to six inclusive. This might simulate the rolling of an honest<br />

die. Make a plot of the \time series" and xn+1 versus xn to show the randomness.<br />

Problem 8-29: Literature search<br />

Apply the techniques of attractor reconstruction to some time series data (e.g.,<br />

the Dow Jones index) extracted from newspapers, magazines, or whatever, and<br />

see whether a pattern emerges.<br />

8.5.3 Butter°y Reconstruction<br />

A people's literature is the great textbook for real knowledge of them.<br />

Thewritingsofthedayshowthequalityofthepeopleasnohistorical<br />

reconstruction can.<br />

Edith Hamilton, American classical scholar (1867{1963)<br />

Our ¯nal recipe will illustrate how Lorenz's butter°y attractor can be reconstructed<br />

from time series data extracted from the governing system of three<br />

coupled nonlinear ODEs. The Lorenz system (which was discussed in Section<br />

2.2.1) is entered, along with the initial condition x(0) = 2, y(0) = 5, z(0) = 5.<br />

> restart: with(plots):<br />

> sys:=diff(x(t),t)=sigma*(y(t)-x(t)),diff(y(t),t)=-x(t)*z(t)<br />

+r*x(t)-y(t),diff(z(t),t)=x(t)*y(t)-b*z(t);<br />

sys := d<br />

x (t) =¾ (y(t) ¡ x (t));<br />

dt<br />

d<br />

y(t) =¡x(t) z(t)+r x (t) ¡ y(t);<br />

dt<br />

d<br />

z(t) =x(t) y(t) ¡ b z(t)<br />

dt<br />

> ic:=(x(0)=2,y(0)=5,z(0)=5):<br />

The parameter values are taken to be r =28,b =8=3, and ¾ =10.<br />

> r:=28: b:=8/3: sigma:=10:<br />

The total number of entries in the time series will be taken to be N = 2000.<br />

The total time is T = 50, so the step size ¢ = T=N is 1=40.<br />

> N:=2000: T:=50: Delta:=T/N;

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!