honr 300/cmsc 491

16
HONR 300/CMSC 491 Computation, Complexity, and Emergence Mandelbrot & Julia Sets Prof. Marie desJardins February 22, 2012 Based on slides prepared by Nathaniel Wise

Upload: nico

Post on 08-Jan-2016

37 views

Category:

Documents


1 download

DESCRIPTION

Computation, Complexity, and Emergence Mandelbrot & Julia Sets Prof. Marie desJardins February 22, 2012 Based on slides prepared by Nathaniel Wise. HONR 300/CMSC 491. Chapter 8: The Mandelbrot Set & Julia Sets. There once was a young man from Trinity Who took . - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: HONR 300/CMSC 491

HONR 300/CMSC 491

Computation, Complexity, and EmergenceMandelbrot & Julia Sets

Prof. Marie desJardinsFebruary 22, 2012Based on slides prepared by Nathaniel Wise

Page 2: HONR 300/CMSC 491

Chapter 8:The Mandelbrot Set & Julia Sets

There once was a young man from TrinityWho took .But the number of digitsGave him the fidgets;He dropped Math and took up Divinity.

Page 3: HONR 300/CMSC 491

A New Kind of Fractal

The fractals we've looked at are generally self-identical, in that you can look at them at different scales and they look exactly the same.

The Mandelbrot and Julia sets are only self-similar: they have a kind of pattern that's instantly recognizable, at every scale, but no two scales are exactly the same.

We'll see this kind of “different sameness” again when we start to look at chaos theory and chaotic systems.

Page 4: HONR 300/CMSC 491

The Mandelbrot Set Benoit Mandelbrot (1924-2010) is known as the “father of fractal geometry.”

He invented the term “fractal,” and used the new field of computation and digital computers to explore complex mathematical objects that had previously only been studied in the abstract.

The Mandelbrot set is defined using an iterative function:xt+1 = xt + c,

where xt = 0.

The magnitude of a complex number a + bi, is the Euclidean distance of that point from the origin of the complex plane, i.e., √a2 + b2

For a given value c, it turns out that the magnitude of xt+1

will do one of two things:

It will always be smaller than 2 (no matter how large t gets), or

It will eventually diverge (i.e., xt will go to ∞ as t goes to ∞).

The Mandelbrot set is defined as the set of values c for which xt+1 remains smaller than 2.

Page 5: HONR 300/CMSC 491

Computing the Mandelbrot Set

The Mandelbrot set contains those values of c for which the magnitude xt

remains smaller than 2 for all t.

But we have no easy way to know whether the Mandelbrot series diverges for a given value of c!

If we compute the Mandelbrot series for some value c and the magnitude of x

t ever becomes greater than 2, that value c is definitely not in the

Mandelbrot set. (It is a property of the series that if xt is greater than 2,

then subsequent values will always increase.) But a Mandelbrot series may remain below 2 for arbitrarily long before

diverging, and the only way to tell if it will diverge is to compute the sequence for long enough.

Page 6: HONR 300/CMSC 491

The black area corresponds to points in the Mandelbrot set. The colored area represents points not in the Mandelbrot set, where the brightness of the color is proportional to the number of iterations before divergence (i.e., the smallest value of t for which xt ≥ 2).

Page 7: HONR 300/CMSC 491

Julia Sets Long before Mandelbrot, Gaston Julia (1893-1978) had studied a

similar function. (In fact, Mandelbrot started out by studying the Julia set...)

Here, c is fixed complex number (so we talk about “the Julia set for c = some value”) and x

1 is the point being examined (i.e., the point that

is plotted in a display of the Julia set as belonging to that Julia set (or not)).

Julia examined what happens to the series for a given c and x1 as i

increases. As with points in the Mandelbrot set, each such series either diverges, or it does not.

Without the aid of computers, Julia could only sketch relatively crude drawings of these shapes. Today, we can compute the Julia set for any value, to an arbitrary degree of resolution.

Page 8: HONR 300/CMSC 491

Julia Sets

c = -0.375 + 0.61875i c = 0.21875 - 0.575i

The central black areas are points that converge and are a part of the set. The different colors represent how many iterations before that point diverges.

Page 9: HONR 300/CMSC 491

Julia Sets

c = -1.16875 - 0.2875i c = 0.325 + 0.06875i

The central black areas are points that converge and are a part of the set. The different colors represent how many iterations before that point diverges.

Page 10: HONR 300/CMSC 491

Julia Sets

c = -0.04375 + 0.9875i c = -0.3875 - 0.69375i

The central black areas are points that converge and are a part of the set. The different colors represent how many iterations before that point diverges.

Page 11: HONR 300/CMSC 491

The Mandelbrot Set

Some Julia sets consist of infinitely many disconnected regions; others are a single contiguous region (although they may be connected only by arbitrarily fine “filaments”).

The Mandelbrot set serves as a “map” of all the Julia sets.

If a point is inside the Mandelbrot set (colored black), then the corresponding Julia set is contiguous.

The closer a point is to any border area of the Mandelbrot set, the more complex that Julia set will be.

Julia sets often seem to share similar visual characteristics to the corresponding point in the Mandelbrot set.

The NetLogo model posted on the course page lets you explore the Mandelbrot set and corresponding Julia sets:

http://www.csee.umbc.edu/~mariedj/complexity/2012/Mandelbrot.nlogo

Page 12: HONR 300/CMSC 491
Page 13: HONR 300/CMSC 491
Page 14: HONR 300/CMSC 491
Page 15: HONR 300/CMSC 491

The Mandelbrot Set

The Mandelbrot set is perhaps the most complex object in mathematics.

One could spend a lifetime exploring it and never see all of it.

It contains infinitely many imperfect copies of the set within it, none of them matching any other copy.

YouTube user ckorda spent 5 months with about 15 PCs all rendering a video of a Mandelbrot zoom to a depth of 2316 (about 1095):

http://www.youtube.com/watch?v=_QskAoLIzuI

One can zoom as far as your computing power and patience holds up: the NetLogo model can do up to a one-billion zoom, depending on the region.

Page 16: HONR 300/CMSC 491