Mandelbrot set

Initial image of a Mandelbrot set zoom sequence with a continuously colored environment
Progressive infinite iterations of the "Nautilus" section of the Mandelbrot Set rendered using webGL
Mandelbrot animation based on a static number of iterations per pixel
Mandelbrot set detail

The Mandelbrot set is the set of complex numbers c for which the function f_c(z)=z^2+c does not diverge when iterated from z=0, i.e., for which the sequence f_c(0), f_c(f_c(0)), etc., remains bounded in absolute value.

The set is closely related to the idea of Julia sets, which produce similarly complex shapes. Its definition and name are due to Adrien Douady, in tribute to the mathematician Benoit Mandelbrot.[1]

Mandelbrot set images are made by sampling complex numbers and determining for each whether the result tends towards infinity when a particular mathematical operation is iterated on it. Treating the real and imaginary parts of each number as image coordinates, pixels are colored according to how rapidly the sequence diverges, if at all.

More precisely, the Mandelbrot set is the set of values of c in the complex plane for which the orbit of 0 under iteration of the quadratic map

z_{n+1} = z_n^2 + c

remains bounded.[2] That is, a complex number c is part of the Mandelbrot set if, when starting with z0 = 0 and applying the iteration repeatedly, the absolute value of zn remains bounded however large n gets. This can also be represented as[3]

z_{n+1} = z_n^2 + c,
c \in M \iff \lim_{n \to \infty}|z_{n+1}| \leq 2.

For example, letting c = 1 gives the sequence 0, 1, 2, 5, 26,…, which tends to infinity. As this sequence is unbounded, 1 is not an element of the Mandelbrot set. On the other hand, c = −1 gives the sequence 0, −1, 0, −1, 0,…, which is bounded, and so −1 belongs to the Mandelbrot set.

Images of the Mandelbrot set display an elaborate boundary that reveals progressively ever-finer recursive detail at increasing magnifications. The "style" of this repeating detail depends on the region of the set being examined. The set's boundary also incorporates smaller versions of the main shape, so the fractal property of self-similarity applies to the entire set, and not just to its parts.

The Mandelbrot set has become popular outside mathematics both for its aesthetic appeal and as an example of a complex structure arising from the application of simple rules. It is one of the best-known examples of mathematical visualization.

History

The first published picture of the Mandelbrot set, by Robert W. Brooks and Peter Matelski in 1978

The Mandelbrot set has its place in complex dynamics, a field first investigated by the French mathematicians Pierre Fatou and Gaston Julia at the beginning of the 20th century. This fractal was first defined and drawn in 1978 by Robert W. Brooks and Peter Matelski as part of a study of Kleinian groups.[4] On 1 March 1980, at IBM's Thomas J. Watson Research Center in Yorktown, Heights, New York, Benoit Mandelbrot first saw a visualization of the set.[5]

Mandelbrot studied the parameter space of quadratic polynomials in an article that appeared in 1980.[6] The mathematical study of the Mandelbrot set really began with work by the mathematicians Adrien Douady and John H. Hubbard,[1] who established many of its fundamental properties and named the set in honor of Mandelbrot.

The mathematicians Heinz-Otto Peitgen and Peter Richter became well known for promoting the set with photographs, books,[7] and an internationally touring exhibit of the German Goethe-Institut.[8][9]

The cover article of the August 1985 Scientific American introduced the algorithm for computing the Mandelbrot set to a wide audience. The cover featured an image created by Peitgen, et al.[10][11] The Mandelbrot set became prominent in the mid-1980s as a computer graphics demo, when personal computers became powerful enough to plot and display the set in high resolution.[12]

The work of Douady and Hubbard coincided with a huge increase in interest in complex dynamics and abstract mathematics, and the study of the Mandelbrot set has been a centerpiece of this field ever since. An exhaustive list of all the mathematicians who have contributed to the understanding of this set since then is beyond the scope of this article, but such a list would notably include Mikhail Lyubich,[13][14] Curt McMullen, John Milnor, Mitsuhiro Shishikura, and Jean-Christophe Yoccoz.

Formal definition

The Mandelbrot set M is defined by a family of complex quadratic polynomials

P_c:\mathbb C\to\mathbb C

given by

P_c: z\mapsto z^2 + c,

where c is a complex parameter. For each c, one considers the behavior of the sequence

(0, P_c(0), P_c(P_c(0)), P_c(P_c(P_c(0))), \ldots)

obtained by iterating P_c(z) starting at critical point z = 0, which either escapes to infinity or stays within a disk of some finite radius. The Mandelbrot set is defined as the set of all points c such that the above sequence does not escape to infinity.

A mathematician's depiction of the Mandelbrot set M. A point c is colored black if it belongs to the set, and white if not. Re[c] and Im[c] denote the real and imaginary parts of c, respectively.

More formally, if P_c^n(z) denotes the nth iterate of P_c(z) (i.e. P_c(z) composed with itself n times), the Mandelbrot set is the subset of the complex plane given by

M = \left\{c\in \mathbb C : \exists s\in \mathbb R, \forall n\in \mathbb N, |P_c^n(0)| \le s \right\}.

As explained below, it is in fact possible to simplify this definition by taking s=2.

Mathematically, the Mandelbrot set is just a set of complex numbers. A given complex number c either belongs to M or it does not. A picture of the Mandelbrot set can be made by coloring all the points c that belong to M black, and all other points white. The more colorful pictures usually seen are generated by coloring points not in the set according to which term in the sequence |P_c^n(0)| is the first term with an absolute value greater than a certain cutoff value, usually 2. See the section on computer drawings below for more details.

The Mandelbrot set can also be defined as the connectedness locus of the family of polynomials P_c(z). That is, it is the subset of the complex plane consisting of those parameters c for which the Julia set of P_c is connected.

Basic properties

The Mandelbrot set is a compact set, since it is closed and contained in the closed disk of radius 2 around the origin. More specifically, a point c belongs to the Mandelbrot set if and only if

|P_c^n(0)|\leq 2 for all n\geq 0.

In other words, if the absolute value of P_c^n(0) ever becomes larger than 2, the sequence will escape to infinity.

Correspondence between the Mandelbrot set and the bifurcation diagram of the logistic map

The intersection of M with the real axis is precisely the interval [−2, 1/4]. The parameters along this interval can be put in one-to-one correspondence with those of the real logistic family,

z\mapsto \lambda z(1-z),\quad \lambda\in[1,4].

The correspondence is given by

c = \frac\lambda2\left(1-\frac\lambda2\right).

In fact, this gives a correspondence between the entire parameter space of the logistic family and that of the Mandelbrot set.

As of October 2012, the area of the Mandelbrot is estimated to be 1.5065918849 ± 0.0000000028.[15]

Douady and Hubbard have shown that the Mandelbrot set is connected. In fact, they constructed an explicit conformal isomorphism between the complement of the Mandelbrot set and the complement of the closed unit disk. Mandelbrot had originally conjectured that the Mandelbrot set is disconnected. This conjecture was based on computer pictures generated by programs that are unable to detect the thin filaments connecting different parts of M. Upon further experiments, he revised his conjecture, deciding that M should be connected.

External rays of wakes near the period 1 continent in the Mandelbrot set

The dynamical formula for the uniformisation of the complement of the Mandelbrot set, arising from Douady and Hubbard's proof of the connectedness of M, gives rise to external rays of the Mandelbrot set. These rays can be used to study the Mandelbrot set in combinatorial terms and form the backbone of the Yoccoz parapuzzle.[16]

The boundary of the Mandelbrot set is exactly the bifurcation locus of the quadratic family; that is, the set of parameters c for which the dynamics changes abruptly under small changes of c. It can be constructed as the limit set of a sequence of plane algebraic curves, the Mandelbrot curves, of the general type known as polynomial lemniscates. The Mandelbrot curves are defined by setting p0 = z, pn+1 = pn2 + z, and then interpreting the set of points |pn(z)| = 2 in the complex plane as a curve in the real Cartesian plane of degree 2n+1 in x and y.

Other properties

Main cardioid and period bulbs

Periods of hyperbolic components

Upon looking at a picture of the Mandelbrot set, one immediately notices the large cardioid-shaped region in the center. This main cardioid is the region of parameters c for which P_c has an attracting fixed point. It consists of all parameters of the form

 c = \frac\mu2\left(1-\frac\mu2\right)

for some \mu in the open unit disk.

To the left of the main cardioid, attached to it at the point c=-3/4, a circular-shaped bulb is visible. This bulb consists of those parameters c for which P_c has an attracting cycle of period 2. This set of parameters is an actual circle, namely that of radius 1/4 around −1.

There are infinitely many other bulbs tangent to the main cardioid: for every rational number \tfrac{p}{q}, with p and q coprime, there is such a bulb that is tangent at the parameter

 c_{\frac{p}{q}} = \frac{e^{2\pi i\frac pq}}2\left(1-\frac{e^{2\pi i\frac pq}}2\right).
Attracting cycle in 2/5-bulb plotted over Julia set (animation)

This bulb is called the \tfrac{p}{q}-bulb of the Mandelbrot set. It consists of parameters that have an attracting cycle of period q and combinatorial rotation number \tfrac{p}{q}. More precisely, the q periodic Fatou components containing the attracting cycle all touch at a common point (commonly called the \alpha-fixed point). If we label these components U_0,\dots,U_{q-1} in counterclockwise orientation, then P_c maps the component U_j to the component U_{j+p\,(\operatorname{mod} q)}.

Attracting cycles and Julia sets for parameters in the 1/2, 3/7, 2/5, 1/3, 1/4, and 1/5 bulbs
Cycle periods and antennae

The change of behavior occurring at c_{\frac{p}{q}} is known as a bifurcation: the attracting fixed point "collides" with a repelling period q-cycle. As we pass through the bifurcation parameter into the \tfrac{p}{q}-bulb, the attracting fixed point turns into a repelling fixed point (the \alpha-fixed point), and the period q-cycle becomes attracting.

Hyperbolic components

All the bulbs we encountered in the previous section were interior components of the Mandelbrot set in which the maps P_c have an attracting periodic cycle. Such components are called hyperbolic components.

It is conjectured that these are the only interior regions of M. This problem, known as density of hyperbolicity, may be the most important open problem in the field of complex dynamics. Hypothetical non-hyperbolic components of the Mandelbrot set are often referred to as "queer" or ghost components.[17][18] For real quadratic polynomials, this question was answered positively in the 1990s independently by Lyubich and by Graczyk and Świątek. (Note that hyperbolic components intersecting the real axis correspond exactly to periodic windows in the Feigenbaum diagram. So this result states that such windows exist near every parameter in the diagram.)

Not every hyperbolic component can be reached by a sequence of direct bifurcations from the main cardioid of the Mandelbrot set. However, such a component can be reached by a sequence of direct bifurcations from the main cardioid of a little Mandelbrot copy (see below).

Each of the hyperbolic components has a centre, which is a point c such that the inner Fatou domain for P_c(z) has a super-attracting cycle – that is, that the attraction is infinite (see the image here). This means that the cycle contains the critical point 0, so that 0 is iterated back to itself after some iterations. We therefore have that P_cn(0) = 0 for some n. If we call this polynomial Q^{n}(c) (letting it depend on c instead of z), we have that Q^{n+1}(c) = Q^{n}(c)^{2} + c and that the degree of Q^{n}(c) is 2^{n-1}. We can therefore construct the centres of the hyperbolic components by successively solving the equations Q^{n}(c) = 0, n = 1, 2, 3, .... Note that for each step, we get just as many new centres as we have found so far.

Local connectivity

Topological model of Mandelbrot set without mini Mandelbrot sets and Misiurewicz points (Cactus model)
Thurston model of Mandelbrot set (abstract Mandelbrot set)

It is conjectured that the Mandelbrot set is locally connected. This famous conjecture is known as MLC (for Mandelbrot Locally Connected). By the work of Adrien Douady and John H. Hubbard, this conjecture would result in a simple abstract "pinched disk" model of the Mandelbrot set. In particular, it would imply the important hyperbolicity conjecture mentioned above.

The work of Jean-Christophe Yoccoz established local connectivity of the Mandelbrot set at all finitely renormalizable parameters; that is, roughly speaking those contained only in finitely many small Mandelbrot copies.[19] Since then, local connectivity has been proved at many other points of M, but the full conjecture is still open.

Self-similarity

Self similarity in the Mandelbrot set shown by zooming in on a round feature while panning in the negative-x direction. The display center pans from (−1, 0) to (−1.31, 0) while the view magnifies from 0.5 × 0.5 to 0.12 × 0.12 to approximate the Feigenbaum ratio \delta.
Self-similarity around Misiurewicz point −0.1011 + 0.9563i.

The Mandelbrot set is self-similar under magnification in the neighborhoods of the Misiurewicz points. It is also conjectured to be self-similar around generalized Feigenbaum points (e.g., −1.401155 or −0.1528 + 1.0397i), in the sense of converging to a limit set.[20][21]

Quasi-self-similarity in the Mandelbrot set

The Mandelbrot set in general is not strictly self-similar but it is quasi-self-similar, as small slightly different versions of itself can be found at arbitrarily small scales.

The little copies of the Mandelbrot set are all slightly different, mostly because of the thin threads connecting them to the main body of the set.

Further results

The Hausdorff dimension of the boundary of the Mandelbrot set equals 2 as determined by a result of Mitsuhiro Shishikura.[22] It is not known whether the boundary of the Mandelbrot set has positive planar Lebesgue measure.

In the Blum-Shub-Smale model of real computation, the Mandelbrot set is not computable, but its complement is computably enumerable. However, many simple objects (e.g., the graph of exponentiation) are also not computable in the BSS model. At present, it is unknown whether the Mandelbrot set is computable in models of real computation based on computable analysis, which correspond more closely to the intuitive notion of "plotting the set by a computer". Hertling has shown that the Mandelbrot set is computable in this model if the hyperbolicity conjecture is true.

The occurrence of π in the Mandelbrot set was discovered by David Boll in 1991.[23] He found that when looking at the pinch points of the Mandelbrot set, the number of iterations needed for the point (−3/4, ε) before escaping, multiplied by ε, was equal to π. Based on this initial finding, Aaron Klebanoff developed a further test near another pinch point (1/4 + ε, 0) in the Mandelbrot set and found that the number of iterations times the square root of ε was equal to π.

A zoom into the Mandelbrot set illustrating a Julia "island" and the corresponding Julia set of the form f_c(z) = z^2 + c\,, in which c is the centre of the Mandelbrot set zoom-in

Relationship with Julia sets

As a consequence of the definition of the Mandelbrot set, there is a close correspondence between the geometry of the Mandelbrot set at a given point and the structure of the corresponding Julia set. For instance, a point is in the Mandelbrot set exactly when the corresponding Julia set is connected.

This principle is exploited in virtually all deep results on the Mandelbrot set. For example, Shishikura proves that, for a dense set of parameters in the boundary of the Mandelbrot set, the Julia set has Hausdorff dimension two, and then transfers this information to the parameter plane.[22] Similarly, Yoccoz first proved the local connectivity of Julia sets, before establishing it for the Mandelbrot set at the corresponding parameters.[19] Adrien Douady phrases this principle as:

Plough in the dynamical plane, and harvest in parameter space.

Geometry

Components on main cardioid for periods 8–14 with antennae 7–13

For every rational number \tfrac{p}{q}, where p and q are relatively prime, a hyperbolic component of period q bifurcates from the main cardioid. The part of the Mandelbrot set connected to the main cardioid at this bifurcation point is called the p/q-limb. Computer experiments suggest that the diameter of the limb tends to zero like \tfrac{1}{q^2}. The best current estimate known is the Yoccoz-inequality, which states that the size tends to zero like \tfrac{1}{q}.

A period-q limb will have q  1 "antennae" at the top of its limb. We can thus determine the period of a given bulb by counting these antennas.

In an attempt to demonstrate that the thickness of the p/q-limb is zero, David Boll carried out a computer experiment in 1991, where he computed the number of iterations required for the series to converge for z = -\tfrac{3}{4} + i\epsilon (-\tfrac{3}{4} being the location thereof). As the series doesn't converge for the exact value of z = -\tfrac{3}{4}, the number of iterations required increases with a small ε. It turns out that multiplying the value of ε with the number of iterations required yields an approximation of π that becomes better for smaller ε. For example, for ε = 0.0000001 the number of iterations is 31415928 and the product is 3.1415928.[24]

Image gallery of a zoom sequence

The Mandelbrot set shows more intricate detail the closer one looks or magnifies the image, usually called "zooming in". The following example of an image sequence zooming to a selected c value gives an impression of the infinite richness of different geometrical structures and explains some of their typical rules.

The magnification of the last image relative to the first one is about 1010 to 1. Relating to an ordinary monitor, it represents a section of a Mandelbrot set with a diameter of 4 million kilometres. Its border would show an astronomical number of different fractal structures.

The seahorse "body" is composed by 25 "spokes" consisting of two groups of 12 "spokes" each and one "spoke" connecting to the main cardioid. These two groups can be attributed by some kind of metamorphosis to the two "fingers" of the "upper hand" of the Mandelbrot set; therefore, the number of "spokes" increases from one "seahorse" to the next by 2; the "hub" is a so-called Misiurewicz point. Between the "upper part of the body" and the "tail" a distorted small copy of the Mandelbrot set called satellite may be recognized.

The islands above seem to consist of infinitely many parts like Cantor sets, as is actually the case for the corresponding Julia set Jc. However, they are connected by tiny structures, so that the whole represents a simply connected set. The tiny structures meet each other at a satellite in the center that is too small to be recognized at this magnification. The value of c for the corresponding Jc is not that of the image center but, relative to the main body of the Mandelbrot set, has the same position as the center of this image relative to the satellite shown in the 6th zoom step.

Generalizations

Animation of the Mandelbrot set for d from 0 to 5
Animation of the set for d from 0.05 to 2

Multibrot sets are bounded sets found in the complex plane for members of the general monic univariate polynomial family of recursions

 z \mapsto z^d + c.\

For integer d, these sets are connectedness loci for the Julia sets built from the same formula. The full cubic connectedness map has also been studied; here one considers the two-parameter recursion  z \mapsto z^3 + 3kz + c , whose two critical points are the complex square roots of the parameter k. A point is in the map if either critical point is stable.[25]

For general families of holomorphic functions, the boundary of the Mandelbrot set generalizes to the bifurcation locus, which is a natural object to study even when the connectedness locus is not useful.

Other, non-analytic, mappings

Of particular interest is the tricorn fractal, the connectedness locus of the anti-holomorphic family

 z \mapsto \bar{z}^2 + c.

The tricorn (also sometimes called the Mandelbar set) was encountered by Milnor in his study of parameter slices of real cubic polynomials. It is not locally connected. This property is inherited by the connectedness locus of real cubic polynomials.

Another non-analytic generalization is the Burning Ship fractal, which is obtained by iterating the mapping

 z \mapsto (|\Re \left(z\right)|+i|\Im \left(z\right)|)^2 + c.

The Multibrot set is obtained by varying the value of the exponent d. The article has a video that shows the development from d = 0 to 7 at which point there are 6 i.e. (d − 1) lobes around the perimeter. A similar development with negative exponents results in (1 − d) clefts on the inside of a ring.

Computer drawings

Still image of a movie of increasing magnification on 0.001643721971153 − 0.822467633298876i
Still image of an animation of increasing magnification

There are many programs used to generate the Mandelbrot set and other fractals, some of which are described in fractal-generating software. These programs use a variety of algorithms to determine the color of individual pixels and achieve efficient computation.

Escape time algorithm

The simplest algorithm for generating a representation of the Mandelbrot set is known as the "escape time" algorithm. A repeating calculation is performed for each x, y point in the plot area and based on the behavior of that calculation, a color is chosen for that pixel.

The x and y locations of each point are used as starting values in a repeating, or iterating calculation (described in detail below). The result of each iteration is used as the starting values for the next. The values are checked during each iteration to see whether they have reached a critical "escape" condition, or "bailout". If that condition is reached, the calculation is stopped, the pixel is drawn, and the next x, y point is examined. For some starting values, escape occurs quickly, after only a small number of iterations. For starting values very close to but not in the set, it may take hundreds or thousands of iterations to escape. For values within the Mandelbrot set, escape will never occur. The programmer or user must choose how much iteration, or "depth", they wish to examine. The higher the maximal number of iterations, the more detail and subtlety emerge in the final image, but the longer time it will take to calculate the fractal image.

Escape conditions can be simple or complex. Because no complex number with a real or imaginary part greater than 2 can be part of the set, a common bailout is to escape when either coefficient exceeds 2. A more computationally complex method that detects escapes sooner, is to compute distance from the origin using the Pythagorean theorem, i.e., to determine the absolute value, or modulus, of the complex number. If this value exceeds 2, the point has reached escape. More computationally intensive rendering variations include the Buddhabrot method, which finds escaping points and plots their iterated coordinates.

The color of each point represents how quickly the values reached the escape point. Often black is used to show values that fail to escape before the iteration limit, and gradually brighter colors are used for points that escape. This gives a visual representation of how many cycles were required before reaching the escape condition.

To render such an image, the region of the complex plane we are considering is subdivided into a certain number of pixels. To color any such pixel, let c be the midpoint of that pixel. We now iterate the critical point 0 under P_c, checking at each step whether the orbit point has modulus larger than 2. When this is the case, we know that c does not belong to the Mandelbrot set, and we color our pixel according to the number of iterations used to find out. Otherwise, we keep iterating up to a fixed number of steps, after which we decide that our parameter is "probably" in the Mandelbrot set, or at least very close to it, and color the pixel black.

In pseudocode, this algorithm would look as follows. The algorithm does not use complex numbers and manually simulates complex-number operations using two real numbers, for those who do not have a complex data type. The program may be simplified if the programming language includes complex-data-type operations.

For each pixel (Px, Py) on the screen, do:
{
  x0 = scaled x coordinate of pixel (scaled to lie in the Mandelbrot X scale (-2.5, 1))
  y0 = scaled y coordinate of pixel (scaled to lie in the Mandelbrot Y scale (-1, 1))
  x = 0.0
  y = 0.0
  iteration = 0
  max_iteration = 1000
  while (x*x + y*y < 2*2  AND  iteration < max_iteration) {
    xtemp = x*x - y*y + x0
    y = 2*x*y + y0
    x = xtemp
    iteration = iteration + 1
  }
  color = palette[iteration]
  plot(Px, Py, color)
}

Here, relating the pseudocode to c, z and P_c:

and so, as can be seen in the pseudocode in the computation of x and y:

To get colorful images of the set, the assignment of a color to each value of the number of executed iterations can be made using one of a variety of functions (linear, exponential, etc.). One practical way, without slowing down calculations, is to use the number of executed iterations as an entry to a look-up color palette table initialized at startup. If the color table has, for instance, 500 entries, then the color selection is n mod 500, where n is the number of iterations.

Histogram coloring

A more accurate coloring method involves using a histogram, which keeps track of how many pixels reached each iteration number, from 1 to n. This method will equally distribute colors to the same overall area, and, importantly, is independent of the maximal number of iterations chosen.

First, create an array of size n. For each pixel, which took i iterations, find the ith element and increment it. This creates the histogram during computation of the image. Then, when finished, perform a second "rendering" pass over each pixel, utilizing the completed histogram. If you had a continuous color palette ranging from 0 to 1, you could find the normalized color of each pixel as follows, using the variables from above.

total = 0
for (i = 0; i < max_iterations; i += 1) {
  total += histogram[i]
}

hue = 0.0;
for (i = 0; i < iteration; i += 1) {
  hue += histogram[i] / total // Must be floating-point division.
}

color = palette[hue]

This method may be combined with the smooth coloring method below for more aesthetically pleasing images.

Continuous (smooth) coloring

This image was rendered with the escape time algorithm. Notice the very obvious "bands" of color
This image was rendered with the normalized iteration count algorithm. Notice the bands of color have been replaced by a smooth gradient. Also, the colors take on the same pattern that would be observed if the escape time algorithm were used.

The escape time algorithm is popular for its simplicity. However, it creates bands of color, which, as a type of aliasing, can detract from an image's aesthetic value. This can be improved using an algorithm known as "normalized iteration count",[26][27] which provides a smooth transition of colors between iterations. The algorithm associates a real number \nu with each value of z by using the connection of the iteration number with the potential function. This function is given by

\phi(z) = \lim_{n \to \infty} (\log|z_n|/P^{n}),

where zn is the value after n iterations and P is the power for which z is raised to in the Mandelbrot set equation (zn+1 = znP + c, P is generally 2).

If we choose a large bailout radius N (e.g., 10100), we have that

\log|z_n|/P^{n} = \log(N)/P^{\nu(z)}

for some real number \nu(z), and this is

\nu(z) = n - \log_P (\log|z_n|/\log(N)),

and as n is the first iteration number such that |zn| > N, the number we subtract from n is in the interval [0, 1).

For the coloring we must have a cyclic scale of colors (constructed mathematically, for instance) and containing H colors numbered from 0 to H  1 (H = 500, for instance). We multiply the real number \nu(z) by a fixed real number determining the density of the colors in the picture, take the integral part of this number modulo H, and use it to look up the corresponding color in the color table.

For example, modifying the above pseudocode and also using the concept of linear interpolation would yield

For each pixel (Px, Py) on the screen, do:
{
  x0 = scaled x coordinate of pixel (scaled to lie in the Mandelbrot X scale (-2.5, 1))
  y0 = scaled y coordinate of pixel (scaled to lie in the Mandelbrot Y scale (-1, 1))
  x = 0.0
  y = 0.0
  iteration = 0
  max_iteration = 1000
  // Here N=2^8 is chosen as a reasonable bailout radius.
  while ( x*x + y*y < (1 << 16)  AND  iteration < max_iteration ) {
    xtemp = x*x - y*y + x0
    y = 2*x*y + y0
    x = xtemp
    iteration = iteration + 1
  }
  // Used to avoid floating point issues with points inside the set.
  if ( iteration < max_iteration ) {
    // sqrt of inner term removed using log simplification rules.
    log_zn = log( x*x + y*y ) / 2
    nu = log( log_zn / log(2) ) / log(2)
    // Rearranging the potential function.
    // Dividing log_zn by log(2) instead of log(N = 1<<8)
    // because we want the entire palette to range from the
    // center to radius 2, NOT our bailout radius.
    iteration = iteration + 1 - nu
  }
  color1 = palette[floor(iteration)]
  color2 = palette[floor(iteration) + 1]
  // iteration % 1 = fractional part of iteration.
  color = linear_interpolate(color1, color2, iteration % 1)
  plot(Px, Py, color)
}

Distance estimates

One can compute the distance from point c (in exterior or interior) to nearest point on the boundary of the Mandelbrot set.[28]

Exterior distance estimation

The proof of the connectedness of the Mandelbrot set in fact gives a formula for the uniformizing map of the complement of M (and the derivative of this map). By the Koebe 1/4 theorem, one can then estimate the distance between the midpoint of our pixel and the Mandelbrot set up to a factor of 4.

In other words, provided that the maximal number of iterations is sufficiently high, one obtains a picture of the Mandelbrot set with the following properties:

  1. Every pixel that contains a point of the Mandelbrot set is colored black.
  2. Every pixel that is colored black is close to the Mandelbrot set.
Exterior distance estimate may be used to color whole complement of Mandelbrot set

The distance estimate b of a pixel c (a complex number) from the Mandelbrot set is given by

b=\lim_{n \to \infty} 2 \cdot \frac{|{P_c^n(c)| \cdot \ln|{P_c^n(c)}}|}{|\frac{\partial}{\partial{c}} P_c^n(c)|},

where

The idea behind this formula is simple: When the equipotential lines for the potential function \phi(z) lie close, the number |\phi'(z)| is large, and conversely, therefore the equipotential lines for the function \phi(z)/|\phi'(z)| should lie approximately regularly.

From a mathematician's point of view, this formula only works in limit where n goes to infinity, but very reasonable estimates can be found with just a few additional iterations after the main loop exits.

Once b is found, by the Koebe 1/4-theorem, we know there's no point of the Mandelbrot set with distance from c smaller than b/4.

The distance estimation can be used for drawing of the boundary of the Mandelbrot set, see the article Julia set.

Interior distance estimation

Pixels colored according to the estimated interior distance

It is also possible to estimate the distance of a limitly periodic (i.e., inner) point to the boundary of the Mandelbrot set. The estimate is given by

b=\frac{1-\left|{\frac{\partial}{\partial{z}}P_c^p(z_0)}\right|^2}
  {\left|{\frac{\partial}{\partial{c}}\frac{\partial}{\partial{z}}P_c^p(z_0) +
        \frac{\partial}{\partial{z}}\frac{\partial}{\partial{z}}P_c^p(z_0)
        \frac{\frac{\partial}{\partial{c}}P_c^p(z_0)}
             {1-\frac{\partial}{\partial{z}}P_c^p(z_0)}} \right|},

where

Analogous to the exterior case, once b is found, we know that all points within the distance of b/4 from c are inside the Mandelbrot set.

There are two practical problems with the interior distance estimate: first, we need to find z_0 precisely, and second, we need to find p precisely. The problem with z_0 is that the convergence to z_0 by iterating P_c(z) requires, theoretically, an infinite number of operations. The problem with any given p is that, sometimes, due to rounding errors, a period is falsely identified to be an integer multiple of the real period (e.g., a period of 86 is detected, while the real period is only 43=86/2). In such case, the distance is overestimated, i.e., the reported radius could contain points outside the Mandelbrot set.

3D view: smallest absolute value of the orbit of the interior points of the Mandelbrot set

Optimizations

Cardioid / bulb checking

One way to improve calculations is to find out beforehand whether the given point lies within the cardioid or in the period-2 bulb. Before passing the complex value through the escape time algorithm, first check that:

 p = \sqrt{  \left(x - \frac{1}{4}\right)^2 + y^2} ,
 x < p - 2p^2 + \frac{1}{4} ,
 (x+1)^2 + y^2 < \frac{1}{16} ,

where x represents the real value of the point and y the imaginary value. The first two equations determine that the point is within the cardioid, the last the period-2 bulb.

The cardioid test can equivalently be performed without the square root:

 q = \left(x - \frac{1}{4}\right)^2 + y^2,
 q \left(q + \left(x - \frac{1}{4}\right)\right) < \frac{1}{4}y^2.

3rd- and higher-order buds do not have equivalent tests, because they are not perfectly circular.[29] However, it is possible to find whether the points are within circles inscribed within these higher-order bulbs, preventing many, though not all, of the points in the bulb from being iterated.

Periodicity checking

To prevent having to do huge numbers of iterations for points in the set, one can perform periodicity checking. Check whether a point reached in iterating a pixel has been reached before. If so, the pixel cannot diverge and must be in the set.

Periodicity checking is, of course, a trade-off. The need to remember points costs memory and data management instructions, whereas it saves computational instructions.

However, checking against only one previous iteration can detect many periods with little performance overhead. For example, within the while loop of the pseudocode above, make the following modifications.

  while (x*x + y*y < 2*2  AND  iteration < max_iteration) {
    xtemp = x*x - y*y + x0
    ytemp = 2*x*y + y0
    if (x == xtemp  AND  y == ytemp) {
      iteration = max_iteration
      break
    }
    x = xtemp
    y = ytemp
    iteration = iteration + 1
  }

Border tracing / edge checking

Edge detection using Sobel filter of hyperbolic components of Mandelbrot set

It can be shown that if a solid shape can be drawn on the Mandelbrot set, with all the border colors being the same, then the shape can be filled in with that color. This is a result of the Mandelbrot set being simply connected. Boundary tracing works by following the lemniscates of the various iteration levels (colored bands) all around the set, and then filling the entire band at once. This can be a good speed increase, because it means that large numbers of points can be skipped.[30]

A similar method operating on the same principle uses rectangles instead of arbitrary border shapes. It is usually faster than boundary tracing because it requires fewer calculations to work out the rectangle. It is inefficient, however, because boundaries are not rectangular, and so some areas can be missed. This issue can be minimized by created a recursive algorithm that, if a rectangle border fails, will subdivide it into four smaller rectangles and test those, and either fill each or subdivide again and repeat the process.

However, this only works using discrete colors in the escape time algorithm. It will not work for smooth/continuous coloring.

Perturbation theory and series approximation

Very highly magnified images require more than the standard 64–128 or so bits of precision that most hardware floating-point units provide, requiring renderers to use slow "bignum" or "arbitrary-precision" math libraries to calculate. However, this can be sped up by the exploitation of perturbation theory. Given

 z_{n+1} = z_n^2 + c

as the iteration, and a small epsilon, it is the case that

 (z_n + \epsilon)^2 + c = z_n^2 + 2z_n\epsilon + \epsilon^2 + c,

or

 z_{n+1} + 2z_n\epsilon + \epsilon^2,

so if one defines

 \epsilon_{n+1} = 2z_n\epsilon_n + \epsilon_n^2,

one can calculate a single point (e.g. the center of an image) using high-precision arithmetic (z), giving a reference orbit, and then compute many points around it in terms of various initial offsets epsilon-zero plus the above iteration for epsilon. For most iterations, epsilon does not need more than 16 significant figures, and consequently hardware floating-point may be used to get a mostly accurate image.[31] There will often be some areas where the orbits of points diverge enough from the reference orbit that extra precision is needed on those points, or else additional local high-precision-calculated reference orbits are needed. By measuring the orbit distance between the reference point and the point calculated with low precision, it can be detected that it is not possible to calculate the point correctly, and the calculation can be stopped. These incorrect points can later be re-calculated e.g. from another closer reference point.

Further, it is possible to approximate the starting values for the low-precision points with a truncated Taylor series, which often enables a significant amount of iterations to be skipped.[32] Renderers implementing these techniques are publicly available and offer speedups for highly magnified images by around two orders of magnitude.[33]

References in popular culture

See also

References

  1. 1 2 Adrien Douady and John H. Hubbard, Etude dynamique des polynômes complexes, Prépublications mathémathiques d'Orsay 2/4 (1984 / 1985)
  2. "Mandelbrot Set Explorer: Mathematical Glossary". Retrieved 2007-10-07.
  3. "Escape Radius, Mu-Ency at MROB". Retrieved 2015-10-21.
  4. Robert Brooks and Peter Matelski, The dynamics of 2-generator subgroups of PSL(2,C), in Irwin Kra (1 May 1981). Irwin Kra, ed. Riemann Surfaces and Related Topics: Proceedings of the 1978 Stony Brook Conference. Bernard Maskit. Princeton University Press. pp. 65–. ISBN 0-691-08267-7.
  5. R.P. Taylor & J.C. Sprott (2008). "Biophilic Fractals and the Visual Journey of Organic Screen-savers" (pdf). Nonlinear Dynamics, Psychology, and Life Sciences, Vol. 12, No. 1. Society for Chaos Theory in Psychology & Life Sciences. Retrieved 1 January 2009.
  6. Benoit Mandelbrot, Fractal aspects of the iteration of z\mapsto\lambda z(1-z) for complex \lambda, z, Annals NY Acad. Sci. 357, 249/259
  7. Peitgen, Heinz-Otto; Richter Peter (1986). The Beauty of Fractals. Heidelberg: Springer-Verlag. ISBN 0-387-15851-0.
  8. Frontiers of Chaos, Exhibition of the Goethe-Institut by H.O. Peitgen, P. Richter, H. Jürgens, M. Prüfer, D.Saupe. since 1985 shown in over 40 countries.
  9. Gleick, James (1987). Chaos: Making a New Science. London: Cardinal. p. 229.
  10. Dewdney, A.K. (1985). A computer microscope zooms in for a close look at the most complicated object in mathematics. http://www.nature.com/scientificamerican/journal/v253/n2/pdf/scientificamerican0885-16.pdf: Scientific American. pp. 16–24. (subscription required)
  11. Fractals: The Patterns of Chaos. John Briggs. 1992. p. 80.
  12. Pountain, Dick (September 1986). "Turbocharging Mandelbrot". Byte Magazine. Retrieved 11 November 2015.
  13. Lyubich, Mikhail (May–June 1999). "Six Lectures on Real and Complex Dynamics" (PDF). Retrieved 2007-04-04.
  14. Lyubich, Mikhail (November 1998). "Regular and stochastic dynamics in the real quadratic family" (PDF). Proceedings of the National Academy of Sciences of the United States of America 95 (24): 14025–14027. doi:10.1073/pnas.95.24.14025. PMC 24319. PMID 9826646. Retrieved 2007-04-04.
  15. Numerical estimation of the area of the Mandelbrot set (2012)
  16. The Mandelbrot set, theme and variations. Tan, Lei. Cambridge University Press, 2000. ISBN 978-0-521-77476-5. Section 2.1, "Yoccoz para-puzzles", p. 121
  17. Exploring the Mandelbrot set. The Orsay Notes by Adrien Douady and John H. Hubbard. page 12
  18. Wolf Jung, March 2002, Homeomorphisms on Edges of the Mandelbrot Set by Wolf Jung
  19. 1 2 Hubbard, J. H. (1993), "Local connectivity of Julia sets and bifurcation loci: three theorems of J.-C. Yoccoz" (PDF), Topological methods in modern mathematics (Stony Brook, NY, 1991), Houston, TX: Publish or Perish, pp. 467–511, MR 1215974. Hubbard cites as his source a 1989 unpublished manuscript of Yoccoz.
  20. Lei.pdf Tan Lei, "Similarity between the Mandelbrot set and Julia Sets", Communications in Mathematical Physics 134 (1990), pp. 587-617.
  21. J. Milnor, "Self-Similarity and Hairiness in the Mandelbrot Set", in Computers in Geometry and Topology, M. Tangora (editor), Dekker, New York, pp. 211-257. (Tangora M. C. (6 January 1989). Computers in Geometry and Topology. Taylor & Francis.)
  22. 1 2 Shishikura, Mitsuhiro (1998), "The Hausdorff dimension of the boundary of the Mandelbrot set and Julia sets", Annals of Mathematics, Second Series 147 (2): 225–267, arXiv:math.DS/9201282, doi:10.2307/121009, MR 1626737.
  23. Klebanoff, Aaron (2001), "π in the Madelbrot set" (PDF), Fractals 9 (4): 393–402, doi:10.1142/S0218348X01000828, MR 1878044
  24. Gary William Flake, The Computational Beauty of Nature, 1998. p. 125. ISBN 978-0-262-56127-3.
  25. Rudy Rucker's discussion of the CCM: CS.sjsu.edu
  26. García, Francisco; Ángel Fernández; Javier Barrallo; Luis Martín. "Coloring Dynamical Systems in the Complex Plane" (PDF). Retrieved 2008-01-21.
  27. Linas Vepstas. "Renormalizing the Mandelbrot Escape".
  28. Albert Lobo Cusidó. "Interior and exterior distance bounds for the Mandelbrot".
  29. "Mandelbrot Bud Maths".
  30. "Boundary Tracing Method". Archived from the original on 2015-02-20.
  31. "Superfractalthing - Arbitrary Precision Mandelbrot Set Rendering in Java".
  32. "SUPERFRACTALTHING MATHS" (PDF).
  33. "Kalles Fraktaler 2".
  34. "Mandelbrot Set". JoCopeda. Retrieved 15 January 2015.
  35. Piers Anthony (1992). Fractal Mode. HarperCollins. ISBN 978-0-246-13902-3.
  36. Arthur C. Clarke (29 September 2011). The Ghost From The Grand Banks. Orion. ISBN 978-0-575-12179-9.
  37. Video: https://www.youtube.com/watch?v=s0UjELAUMjE

Further reading

External links

Wikibooks has a book on the topic of: Fractals
Wikimedia Commons has media related to Mandelbrot set.
This article is issued from Wikipedia - version of the Friday, April 22, 2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.