Welcome to Fractal Forums

Fractal Math, Chaos Theory & Research => Mandelbrot & Julia Set => Topic started by: bkercso on May 28, 2016, 09:32:29 PM




Title: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 28, 2016, 09:32:29 PM
Based on my positive experience with coupled logistic map (http://www.fractalforums.com/new-theories-and-research/bifurcations-of-coupled-system/ (http://(http://www.fractalforums.com/new-theories-and-research/bifurcations-of-coupled-system/)), I constructed a coupled Mandelbrot set. One did it before me (http://link.springer.com/article/10.1007%2Fs11071-010-9845-9#/page-1 (http://link.springer.com/article/10.1007%2Fs11071-010-9845-9#/page-1)), but I cannot download this paper yet... Anyway, this CMS is higly depends not only on the coupling factor, but on the initial values of the secound set (x2,y2). The only way to get another M-set without shift it on the complex plane is take one (or some) iterations on 2nd set in each point before coupling. Obviously, we start the first set from (x=0,y=0).

The coupled set:
Code:
Initial conditions at (x0,y0) point:
x:=0; y:=0;   // 1st set
x2:=x0; y2:=y0;   // 2nd set; it is equal with 1 iteration from (x2=0,y2=0)
coupling:=0..0.5   // choose a value

Iterations:
xtemp:=sqr(x)-sqr(y)+x0;   // 1st set
ytemp:=2*x*y+y0;
x2temp:=sqr(x2)-sqr(y2)+x0;   // 2nd set
y2temp:=2*x2*y2+y0;
x:=xtemp+coupling*(x2temp-xtemp);   // linear, symmetric coupling
y:=ytemp+coupling*(y2temp-ytemp);
x2:=x2temp+coupling*(xtemp-x2temp);
y2:=y2temp+coupling*(ytemp-y2temp);

I'm woking on higher quality images...


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 28, 2016, 09:43:58 PM
Its strange, but this is not a multiple attractor fractal: if I start the two sets (or only the 2nd one) from different initial conditions, I get noise.


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 12:11:15 AM
The antialiasing is tricky a little bit at these fractals (coupled mandelbrot/burning ship/burning bird).
The well known procedure is the follow:
Code:
After iterations, the pixel value is "iter".
Antialiasing:

iter:=iter+1-ln(ln(abs(z)))/ln(ORDER)

where:
z: the complex number after it reached the bailout, z=sqrt(sqr(x)+sqr(y))
(you have to use large bailout for antialiasing, eg. 10 000)
ORDER: order of the fractal
(2 for mandelbrot/burning ship/burning bird, 8 for 8th order mandelbulb etc.)

Now, the ORDER of coupled fractals change with coupling. I didn't find a proper solution for this problem, so I used the brute force method: made some trials and fitted functions.

Coupled Mandelbrot with coupling=0.1
without antialiasing:
(http://s33.postimg.org/pfuqeynun/mandelbrot_0001.png)

with normal antialiasing (ORDER=2):
(http://s33.postimg.org/9f914ls1r/mandelbrot_0002.png)

and with calculated ORDER:
(http://s33.postimg.org/fhcj0q81r/mandelbrot_0003.png)


The collected data:
(http://s33.postimg.org/rs1t9t2kf/Untitled.png)

And the fitted funtions:
Code:
Calculate ORDER of coupled systems:

Mandelbrot:
   if coupling<0.0001  then ORDER:=2
   else if coupling<0.001 then ORDER:=0.87*ln(coupling)+10
   else if coupling<0.003 then ORDER:=-0.89*ln(coupling)-2.1
   else if coupling<0.1 then ORDER:=3
   else if coupling<0.25 then ORDER:=-1.1*ln(coupling)+0.48
   else ORDER:=2;
  
Burning ship:
   if coupling<0.0001  then ORDER:=2
   else if coupling<0.001 then ORDER:=0.87*ln(coupling)+10
   else if coupling<0.003 then ORDER:=-1.5*ln(coupling)-6.6
   else if coupling<0.1 then ORDER:=2.3
   else if coupling<0.25 then ORDER:=-0.28*ln(coupling)+1.7
   else ORDER:=2;

Burning bird:
   if coupling<0.0001  then ORDER:=2
   else if coupling<0.001 then ORDER:=0.87*ln(coupling)+10
   else if coupling<0.002 then ORDER:=3.7
   else if coupling<0.1 then ORDER:=-0.29*ln(coupling)+1.8
   else if coupling<0.25 then ORDER:=-0.56*ln(coupling)+1.2
   else ORDER:=2;
  end;


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 12:41:36 AM
The importance of the idea that start the 2nd mandelbrot set after 1 iteration at every point: if we start the 2nd set with a constant shift, we get something similar (depends on the shift):

(http://s33.postimg.org/5o8dcjq67/mandelbrot_0004_coup_0_1.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 12:55:33 AM
And then the high quality images:

Img #1: Coupled Mandelbrot set, coupling=0.1
(http://s33.postimg.org/egslfdydb/fra_kep0001.png)

Img #2: Zoom at the neck
(http://s33.postimg.org/lap3zumv3/mandelbrot_0001.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 01:23:28 AM
Img #3: Coupled Mandelbrot set, coupling=0.01, zoom at the second neck
This is a stunningly beautiful fractal! ;D
(http://s33.postimg.org/i4gjzfedb/fra_kep0002.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: quaz0r on May 31, 2016, 01:31:24 AM
moar zoom


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 02:39:57 AM
Img #4: processing a video to here... (couplig=1E-4)
(http://s33.postimg.org/unujrpmdb/fra_video_C_00500.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: 0Encrypted0 on May 31, 2016, 02:50:47 AM
In reference to Reply#5, Img #3
This CMS is the closest technique I have seen compared to what Jock Cooper creates.
I really want to try this in Ultra Fractal!


(http://www.fractal-recursions.com/fractals/fractal-91395d.jpg) (http://www.fractal-recursions.com/fractals/fractal-91395d.jpg)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: Chris Thomasson on May 31, 2016, 07:46:38 AM
Its strange, but this is not a multiple attractor fractal: if I start the two sets (or only the 2nd one) from different initial conditions, I get noise.

This is very nice work. Thank you.  :)

FWIW, I have always had fun trying to get the fractal to extend into and morph with the non-escaping bulbs of the MSet. Here is a shot of a bulb that is getting a bit close to being filled in with strange Julia sets:

https://plus.google.com/101799841244447089430/posts/JLZbNFroQSj (https://plus.google.com/101799841244447089430/posts/JLZbNFroQSj)

https://plus.google.com/101799841244447089430/posts/H9FSCf4gJkh (https://plus.google.com/101799841244447089430/posts/H9FSCf4gJkh)



Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 12:41:33 PM
Thanks the valuable input mates! O0

Img #5: coupling=0.1
No, you cannot make a zoom from here which fills the screen with divergent patterns! The very thin divergent (ie. colorful) layer consists of interspiral regions.
(http://s33.postimg.org/glhv9m8e7/fra_kep0003.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 03:08:30 PM
Others also play with this Coupled Mandelbrot, but in a different way: while I use the same constants for the 2 sets and shifted them from each other with 1 iteration at each point, they use different constants. So the authors of the article below couple 2 differnt points of the M-set which are not in special relation, while I couple 2 points are on the same trajectory.
Anyway, these coupled systems are good for 3D fractal generation!
http://download.springer.com/static/pdf/797/art%253A10.1007%252Fs11071-016-2606-7.pdf?originUrl=http%3A%2F%2Flink.springer.com%2Farticle%2F10.1007%2Fs11071-016-2606-7&token2=exp=1464700149~acl=%2Fstatic%2Fpdf%2F797%2Fart%25253A10.1007%25252Fs11071-016-2606-7.pdf%3ForiginUrl%3Dhttp%253A%252F%252Flink.springer.com%252Farticle%252F10.1007%252Fs11071-016-2606-7*~hmac=850399e3d4f3f2278fb733a19f87abe5971981443483c7aa8c4e53c81894a263 (http://download.springer.com/static/pdf/797/art%253A10.1007%252Fs11071-016-2606-7.pdf?originUrl=http%3A%2F%2Flink.springer.com%2Farticle%2F10.1007%2Fs11071-016-2606-7&token2=exp=1464700149~acl=%2Fstatic%2Fpdf%2F797%2Fart%25253A10.1007%25252Fs11071-016-2606-7.pdf%3ForiginUrl%3Dhttp%253A%252F%252Flink.springer.com%252Farticle%252F10.1007%252Fs11071-016-2606-7*~hmac=850399e3d4f3f2278fb733a19f87abe5971981443483c7aa8c4e53c81894a263)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on May 31, 2016, 03:50:28 PM
They use the same constant at both sets (as me), but started the 2nd set with a constant shift from the 1st set, as in Reply #3. See equation (3) and below:
http://download.springer.com/static/pdf/117/art%253A10.1007%252Fs11071-013-0785-z.pdf?originUrl=http%3A%2F%2Flink.springer.com%2Farticle%2F10.1007%2Fs11071-013-0785-z&token2=exp=1464702377~acl=%2Fstatic%2Fpdf%2F117%2Fart%25253A10.1007%25252Fs11071-013-0785-z.pdf%3ForiginUrl%3Dhttp%253A%252F%252Flink.springer.com%252Farticle%252F10.1007%252Fs11071-013-0785-z*~hmac=8ddfedf35b594e2e7adb6b261e77077e9605ff491377e3deea3241cf9ac775d5 (http://download.springer.com/static/pdf/117/art%253A10.1007%252Fs11071-013-0785-z.pdf?originUrl=http%3A%2F%2Flink.springer.com%2Farticle%2F10.1007%2Fs11071-013-0785-z&token2=exp=1464702377~acl=%2Fstatic%2Fpdf%2F117%2Fart%25253A10.1007%25252Fs11071-013-0785-z.pdf%3ForiginUrl%3Dhttp%253A%252F%252Flink.springer.com%252Farticle%252F10.1007%252Fs11071-013-0785-z*~hmac=8ddfedf35b594e2e7adb6b261e77077e9605ff491377e3deea3241cf9ac775d5)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: Max Sinister on May 31, 2016, 07:42:47 PM
@0Encrypted0: This one I liked most.


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 01, 2016, 01:02:40 PM
Img #6: A picture from the video of Img #4 (coupling=1E-4):
(http://s33.postimg.org/xinwnaoa7/fra_video_B_00382.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 01, 2016, 10:51:50 PM
Img #7: coupling=1E-4
(http://s33.postimg.org/6mqubh84v/fra_kep0004.png)

Img #8: Coupled 0.1*Mandelbrot + 0.9*Burning ship (started both from (0,0) @ constant (x0,y0) )
(http://nocache-nocookies.digitalgott.com/gallery/19/4917_16_06_16_12_00_52.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 01, 2016, 11:39:40 PM
Coupled Mandelbrot + Burning ship:
Code:
Initial values:

x=0; y=0;
x2=0; y2=0;


Iterations:

xtemp:=sqr(x)-sqr(y)+x0;   { Mandelbrot }
ytemp:=2*x*y+y0;
x2temp:=sqr(x2)-sqr(y2)+x0;   { Burning ship }
y2temp:=2*abs(x2*y2)+y0;

x:=xtemp+coupling*(x2temp-xtemp);
y:=ytemp+coupling*(y2temp-ytemp);
x2:=x2temp+coupling*(xtemp-x2temp);
y2:=y2temp+coupling*(ytemp-y2temp);
zk:=sqr(x)+sqr(y);

Img #9: Zoom of the above. It combines the emblematic patterns of Mandelbrot and Burning ship. It not was easy to find such a picture; I found it in the positive half of the imag. axis.
center (Re), center (Im), range (Re), range (Im):
-1.7105465724241176E+0000
 4.2582773728477242E-0003
 9.7538511027023023E-0010
 6.5025674018015351E-0010
(http://s33.postimg.org/3lemavvjj/fra_kep0002.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: claude on June 02, 2016, 01:20:28 AM
I had a go at this in Fragmentarium (attached) - the main thing I'm not sure about is how to choose the secondary C value for the coupling, so I just offset it from the primary C value...
(https://mathr.co.uk/misc/2016-06-02_coupled_mandelbrot_set_1.jpg)
(https://mathr.co.uk/misc/2016-06-02_coupled_mandelbrot_set_2.jpg)
(https://mathr.co.uk/misc/2016-06-02_coupled_mandelbrot_set_3.jpg)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 02, 2016, 01:30:17 AM
Wow, cool! :dink:
You did a different thing than me, because I don't adjust the constant, but the initial values of x and y. At 1st set (x,y)=(0,0), at 2nd set (x2,y2)=(c.re,c.im), where c.re, c.im the constants for both sets. This is equal with 1 iteration between the sets.


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 02, 2016, 01:39:42 AM
3 coupled Mandelbrot sets with symmetric "hierarchical" coupling:

Code:
3 coupled Mandelbrot set:

Initial values:

x:=0; y:=0;
x2:=x0; y2:=y0;  { = 1 iteration }
x3:=sqr(x0)-sqr(y0)+x0; y3:=2*x0*y0+y0;   { = 2 iterations }


Iteration:
xtemp:=sqr(x)-sqr(y)+x0;
ytemp:=2*x*y+y0;
x2temp:=sqr(x2)-sqr(y2)+x0;
y2temp:=2*x2*y2+y0;
x3temp:=sqr(x3)-sqr(y3)+x0;
y3temp:=2*x3*y3+y0;

x:=xtemp+coupling*(x2temp-xtemp+coupling*(x3temp-xtemp));
y:=ytemp+coupling*(y2temp-ytemp+coupling*(y3temp-ytemp));
x2:=x2temp+coupling*(x3temp-x2temp+coupling*(xtemp-x2temp));
y2:=y2temp+coupling*(y3temp-y2temp+coupling*(ytemp-y2temp));
x3:=x3temp+coupling*(xtemp-x3temp+coupling*(x2temp-x3temp));
y3:=y3temp+coupling*(ytemp-y3temp+coupling*(y2temp-y3temp));
zk:=sqr(x)+sqr(y);

Img #10: 3 coupled Mandelbrot sets with symmetric "hierarchical" coupling (coupling=0.001): "second generation" hybride patterns
(http://s33.postimg.org/7k87c7h7z/fra_kep0001.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: claude on June 02, 2016, 02:00:08 AM
Wow, cool! :dink:
You did a different thing than me, because I don't adjust the constant, but the initial values of x and y. At 1st set (x,y)=(0,0), at 2nd set (x2,y2)=(c.re,c.im), where c.re, c.im the constants for both sets. This is equal with 1 iteration between the sets.

aha! thanks for the explanation, makes more sense


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 02, 2016, 02:22:57 AM
Cheers!

Img #11: Finally, there already are divergent regions on the edge of these spirals at these 3 coupled sets:
(http://s33.postimg.org/v0f2obt67/fra_kep0003.png)

Img #12: Where the mixed / hybride patterns are
(http://s33.postimg.org/cyv3yuq9b/mandelbrot_0002.png)

Img #13: a zoom of Img #12
(http://s33.postimg.org/wu96esupb/mandelbrot_0003.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: Chillheimer on June 02, 2016, 11:03:00 AM
wow, guys you are doing awesome and probably groundbreaking work here!
keep it up! :)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: Max Sinister on June 02, 2016, 10:53:37 PM
@claude: Your images look as if Fragmentarium wanted to honor its name (OK, there's a bad pun implied by me).


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 03, 2016, 01:38:32 AM
Video #1: Zoom to Img #4 (HD)
 (https://www.youtube.com/watch?v=bly8ie8xqo8[/url)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 03, 2016, 02:52:28 AM
Sinusoidal coupling:

Code:

Sinusoidal coupling:

coupling=coupling_original*(sin(iterationNumber*frequency)+1)/2;


Img #14: 2 coupled Mandelbrot sets with sinusoidal coupling, coupling=0.01, frequency=0.01
(http://s33.postimg.org/8ej9hwshb/fra_kep0001.png)

Img #15: coupling=0.01, frequency=0.001
(http://s33.postimg.org/8zpuv1u0f/mandelbrot_0001.png)

Img #16: Zoom of the above, coupling=0.01, frequency=0.001 (maxiter=8000); empty patterns...
(http://s33.postimg.org/o54suwsof/fra_kep0002.png)

Img #17: coupling=0.01, frequency=0.01
(http://s33.postimg.org/3o2828a3z/fra_kep0003.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: 0Encrypted0 on June 03, 2016, 05:27:19 AM
Last month, I attended the International Fractal Art Symposium, where Christian Kleinhuis gave a talk on hybrid fractals--either alternating or interpolating functions each iteration.  Since I already have an alternating functions formula in the UF formula database, I decided to write an interpolating functions formula.  You can find it in lkm3.ufm and there are some sample parameter sets in lkm3.upr.  The attached image is one example, interpolating between 2 Newtons functions, 2 Mandelbrot, and 2 Julia on each iteration.


Would someone please give a brief explanation of the difference between alternating, interpolating and coupled formulas?


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 04, 2016, 02:29:35 AM
I edited the pictures in Reply #25, because my program contains an error (I called a local and a global variable with the same name).
This sin (or cos) coupling results partially empty patterns.

Img #18: And a new image with cos coupling at a place from Video #1:
Code:
coupling:=1E-4*(1E-3*cos(iterationNumber*1E-3)+1)
(http://s33.postimg.org/oqzskdghb/fra_kep0001_1_E_3cos.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 04, 2016, 02:39:32 AM
Img #19: this is without sin coupling, just simple coupled Mandelbrots (2 sets):
Set: (place of Img #15, without sinusoidal modulations of coupling)
(http://s33.postimg.org/ou1muc97z/mandelbrot_0001.png)

(http://s33.postimg.org/poauuitv3/mandelbrot_0002.png)

Zoom:
(http://s33.postimg.org/fdv49fgrz/mandelbrot_0004.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 05, 2016, 02:03:28 AM
3 different coupling for Img #18:

Img #20: original image, simple coupling
Code:
coupling: 1E-4
(http://s33.postimg.org/pls7v0gcf/fra_kep0001_original.png)

Img #18 (again): coupling with cosinus ripple
Code:
coupling:=coupling_original*(1+Amplitude*cos(iterationNumber*Frequency));

coupling_original: 1E-4
Amplitude=Frequency: 1E-3    // relative amplitude
(http://s33.postimg.org/funjj1yv3/fra_kep0001_1_E_3cos.png)

Img #21: coupling with random ripple (the same random number series was used for each pixel)
Code:
coupling:=coupling_original*(1+Amplitude*(random-0.5)*2);   // @each iteration

coupling_original: 1E-4
Amplitude: 1E-3    // relative amplitude
random: 0..1
The image is not reproducible, depends on the values of the random number series
(http://s33.postimg.org/r4v09iif3/fra_kep0001_random.png)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on June 05, 2016, 11:53:13 PM
I opened a new topic for fractals with random noise. I link it because these produce similar patterns:
http://www.fractalforums.com/mandelbrot-and-julia-set/noisy-ifs-fractals-new-patterns/msg93591/#msg93591 (http://www.fractalforums.com/mandelbrot-and-julia-set/noisy-ifs-fractals-new-patterns/msg93591/#msg93591)


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: 0Encrypted0 on July 31, 2016, 06:22:58 AM
Alternate, Interpolate and DEcombinate are the current methods to hybridize formulas in Mandelbulb 3D.
Is the Coupled Mandelbrot Sets (CMS) method applicable to 3D?
Could it be added to the above methods?


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: bkercso on August 23, 2016, 12:12:26 PM
I think it should be the same as the interpolate method. Please see the code in the 1st page.


Title: Re: Coupled Mandelbrot Sets (CMS)
Post by: claude on August 27, 2016, 08:31:57 PM
Don't know if this paper has been mentioned:

http://arxiv.org/abs/1604.04880

Quote
Real and complex behavior for networks of coupled logistic maps
Anca Radulescu, Ariel Pignatelli
(Submitted on 17 Apr 2016)

Many natural systems are organized as networks, in which the nodes interact in a time-dependent fashion. The object of our study is to relate connectivity to the temporal behavior of a network in which the nodes are (real or complex) logistic maps, coupled according to a connectivity scheme that obeys certain constrains, but also incorporates random aspects. We investigate in particular the relationship between the system architecture and possible dynamics. In the current paper we focus on establishing the framework, terminology and pertinent questions for low-dimensional networks. A subsequent paper will further address the relationship between hardwiring and dynamics in high-dimensional networks.

For networks of both complex and real node-maps, we define extensions of the Julia and Mandelbrot sets traditionally defined in the context of single map iterations. For three different model networks, we use a combination of analytical and numerical tools to illustrate how the system behavior (measured via topological properties of the Julia sets) changes when perturbing the underlying adjacency graph. We differentiate between the effects on dynamics of different perturbations that directly modulate network connectivity: increasing/decreasing edge weights, and altering edge configuration by adding, deleting or moving edges. We discuss the implications of considering a rigorous extension of Fatou-Julia theory known to apply for iterations of single maps, to iterations of ensembles of maps coupled as nodes in a network.

Comments:    20 pages, 16 figures
Subjects:    Dynamical Systems (math.DS)
Cite as:    arXiv:1604.04880 [math.DS]