Lyapunov Exponents for the 2-Dimensional Logistic Equation
Click here for the original high-resolution PNG version. Click here for more images created with lyapunov.cc Introduction I want to show a real-life example of the use of PNGwriter. This program, which is included with the PNGwriter source (in examples/), will output an image characterizing the lyapunov exponents of the logistic equation when the parameter R is not constant but a periodic series of 2 values. What does that mean? Take a simple equation, and iterate it (take an initial value, put it into the equation, get what it spits out, and plug it in again to the equation, etc, etc). This equation has a number in it somewhere called a parameter. Imagine you iterate the equation 1000 times with the parameter at a certain value. You see if the resulting series of points is periodic or chaotic. Moreover, you determine a coefficient that depends upon this parameter, which tells you how chaotic the equation is for that given parameter value. This is called a Lyapunov exponent. Was that too rushed? Ultimately the trajectory (the sequence of the resulting values of iterating the function on an initial value) will be chaotic or periodic based solely on its parameter. Here's a 1-dimensional example: x_n is the nth iterate of an initial value, x_0, which is a number between 0 and 1. x_n+1 is the nth-plus-one iterate. Take x_n+1 = R * x_n * ( 1 - x_n ) where R is this parameter. Take R = 2, for example, and you'll realise that the trajectory is periodic: Initial x_0 = 0.45 2*0.45* ( 1-0.45 ) = 0.49545 2*0.49545* ( 1-0.49545 ) = 0.499959 2*0.499959* ( 1-0.499959 ) = 0.5 It converged to its steady-state solution in just 3 iterations.... subsequent iterations will give 0.5. This is because the equation is very stable for R = 2. This would mean that, if you plotted a line of pixels, each representing a value of R from 0 to 4, you'd see varying degrees of red, black and blue (following the convention that red is chaos, black is neither chaos nor stable and blue is stable. Thus the pixel representing R = 2 would be quite blue. Imagine that instead of varying the parameter continuously from 0 to 4 (for example), you replace this simple parameter with a series... for example ABBB. This means that, for a given value of A (between 1 and 4) and a given value of B (between 1 and 4) you can now iterate the equation first with A as the parameter, then with B then with B then with B and then A again, looping on until you do this 1000 times. You can calculate how chaotic this setup is for these fixed values of A and B. Take A and B as your x-y coordinates, and the lyapunov exponent as the colour value: chaotic is red, super-stable is deep blue, and in between is black. That's where the image comes from!
syntax highlighted by Code2HTML, v. 0.9.1 |
||
© 2002, 2003, 2004, 2005, 2006, 2007 Paul Blackburn |