6. JULIA SETS


Gaston Julia

Gaston Julia was a French mathematician who lived from 1893 to 1978. He is remembered today for the fractal sets which bear his name, the Julia sets

It is interesting how he came to study these sets. He had been looking at some work of Sir Arthur Cayley called ‘The Newton-Fourier Imaginary Problem’ about finding the roots of the equation f(z) = z3 + c = 0 using a method of iteration.  He wondered if one could predict which of the three roots a given starting value of z would reach as a limit.  The details are not important to us here, and we know that Julia found the solution to the problem difficult – not surprising in an age before computers! Today we know that there are ‘basins of attraction’, and these basins are infinitely complex. These basins for f(z) = z3 – 1 = 0, coloured red, blue and green, are pictured at left. This figure is a fractal, and is also an example of a Julia set.


Julia sets

If you have been working through these chapters in sequence, you will have already investigated the Mandelbrot set. In a sense, the Julia sets are a simpler concept than the Mandelbrot set, but we have put them in reverse order here for a very simple reason. For, as we shall see, the Mandelbrot set plays an essential role in classifying the many different types of Julia set. We repeat here the short introduction given for the Mandelbrot set.

As with the Mandelbrot set, to understand the Julia sets, we need to understand complex numbers – the numbers of the form z = x + iy where x, y are real, and i2 = –1. Number z can be represented in the complex plane by the point (x, y).

Now if c is a fixed complex number in the z-plane, let us consider the transformation of the plane under the mapping z z2 + c. Each point z in the plane maps to a new point z2 + c. Suppose now we obtain a whole sequence of points by starting with a given z and carrying out this transformation repeatedly. Thus:

z z2 + c (z2 + c)2 + c ((z2 + c)2 + c) + c ... (*)

It turns out that this sequence of points does one of two things. Either
(a) it remains in some bounded part of the plane (say within a circle of suitably large radius), or
(b) it ‘escapes to infinity’, getting further and further away from the origin.

Now for any given value of c, we define the Julia set Jc to be the set of points z in the complex plane which never escape. Here are some examples (above right and below):

In each of these figures, the actual Julia set is designated in black, although this is obvious in only one of the figures. Normally when a Julia set is constructed, there is a system of colouring. Points which never escape are coloured black. Other points of the plane are coloured according to the speed with which they escape. This is done by choosing a large circle centred at the origin, and counting the number of iterations it takes for a point to get beyond this circle. The result of this process is particularly evident in the top diagram above, looking at the shades of green in the background. Where black points are not evident in a Julia set, they are revealed by zooming in on the set, looking at it with a greater magnification.


Applet

As mentioned above, the Mandelbrot set gives a wonderful way of classifying the Julia sets. We can explore this using the very neat applet devised by Rowan Seymour and used here with his permission. This applet appears on the web site

http://www.refract.ijuru.com

and the author can be contacted at rowanseymour@gmail.com .

Rowan recently obtained his PhD in Audio-visual Speech and Speaker Recognition from Queen's University Belfast, and is shortly going out to Rwanda to train computer programmers there.

Place the applet window in the upper right corner of your screen so that it is visible while you work through the text.

We notice that a Mandelbrot set appears in the left frame and a Julia set in the right frame. In each window the centre point is marked with a cross. The Mandelbrot or Julia set can be moved relative to this cross by dragging with the mouse, or using the arrow keys. Clicking on a frame activates it.

There are various other controls.

• We can zoom in on a set by using the mouse wheel (not Macintosh), or by using the +/– keys, or by entering a magnification number in the Zoom window.
• The colour palette can be shifted using the Z and X keys, and scaled using the C and V keys.
• The colour palette can be inverted using the I key and reversed using the R key.

Notice that in the left frame the coordinates of the cross appear below the figure. These are the c-values which give rise to the Julia set Jc which appears in the right frame.


For any choice of sets, the applet works to continually refine the fractals by taking greater and greater numbers of iterations (*).

This applet gives us an excellent way of determining the relationship between the Mandelbrot set and the various Julia sets.


Properties of Julia sets

• In the applet, place the cross, c, within the Mandelbrot set, M. Describe the corresponding Julia set, Jc. What property of Jc appears to remain constant while c is in M? What happens when point c lies outside M? What happens if c lies near the border of M? (You might like to try zooming in on M while looking at this.)





















# It can be shown that the Julia set Jc is connected (all in one piece) if and only if c is a point in the Mandelbrot set.

But there is much more detail than this!

• Compare the Julia sets Jc for points c in the main cardioid body of M and points c in the large circular disk to the left of the cardioid. In the latter case, notice how many solid regions meet at a common vertex. Now look at the Julia sets Jc for points c in the circular disk atop the cardioid. For these Julia sets, how many regions meet at a vertex? Now do some exploring on your own.





















# It is clear that in any given Julia set Jc the number of regions meeting at a vertex depends on which portion of the Mandelbrot set M the point c lies in. For example, it point c lies in the circular disk to the left of the main cardioid body of m, there are two regions meeting at each vertex of Jc . The self similar nature of the fractal ensures that the arrangement is the same at all such vertices. You might notice an interesting sequence of numbers (at a vertex) by starting with points in the large disk at the left of the cardioid body of M, and working around the cardioid in a clockwise direction, at each step choosing the next largest disk.


• Finally, place the cross, c, on one the fine filaments of the Mandelbrot set. You will notice that the corresponding Julia set Jc is also of filament form. Now zoom in on the Mandelbrot filament. Notice how closely, in general structure, the two filaments are alike.

This is just a small sample of the relationships between the sets M and Jc . And while the mathematics is very interesting, don't lose sight of just how beautiful these sets are!


Bibliography

Excellent fractal book including Julia sets ...

Fractals for the Classroom, Peitgen, H-O., Jürgens, H., Saupe, D. (Springer-Verlag 1992)

Two online background discussions ...

http://en.wikipedia.org/wiki/Julia_set

http://mathworld.wolfram.com/JuliaSet.html

Another applet ...

http://math.bu.edu/DYSYS/applets/JuliaIteration.html