Endemyon
Forums Freshman
Posts: 14
|
|
« on: December 12, 2013, 06:48:11 PM » |
|
Hello Everybody ! I tried a formula wich obtains interesting results: http://www.fractalforums.com/index.php?action=gallery;sa=view;id=15275It's obtained from iteration of Zn+1=(Zn^3+c)*(Zn^2+c) I assume it's possible to do every formula's of type: Zn+1=(Zn^p+c)*(Zn^q+c) with p,q>1 integers because the two parts of the equation are mandelbrot's formulas. Maybe it's even possible to multiply it with other terms of type (Zn^k+c) It's possible to obtained corresponding julia sets just like the normal way, but they're not totally symetric. It seems like a composition of the two sets ! Does anyone is familiar with this kind of thing and tell me more about it ? Friendly, a fractal fellow Edit: A corresponding Julia's set: http://www.fractalforums.com/index.php?action=gallery;sa=view;id=15278And it appears to me that i might have misplaced this topic, could a great asministrator deplace it to the appropriate area ?
|
|
« Last Edit: December 14, 2013, 05:39:13 AM by Endemyon, Reason: New image, Wrong Topic »
|
Logged
|
|
|
|
cKleinhuis
|
|
« Reply #1 on: March 03, 2014, 08:33:45 AM » |
|
hey there, nice meeting you as well, i move it to the math section, what you did there is a kind of hybrid fractal, and yes, basically this is one form of hybridisation, combining existing formulas to new ones, you connected the two z^2 and z^3 mandelbrot equation with the multiplication operation
|
|
|
Logged
|
---
divide and conquer - iterate and rule - chaos is No random!
|
|
|
Rychveldir
Forums Newbie
Posts: 8
|
|
« Reply #2 on: March 06, 2014, 11:58:15 AM » |
|
In your Formula Is c a constant? And did you use at all? Or was it something like or I also like exploring generalized versions of the mandelbrot set using polynomials of the form They can generate really interesting effects.
|
|
|
Logged
|
|
|
|
cKleinhuis
|
|
« Reply #3 on: March 06, 2014, 02:33:42 PM » |
|
i think the "c" is just the z0 value in this case but anyways, all of the combinations should provide interesting results
|
|
|
Logged
|
---
divide and conquer - iterate and rule - chaos is No random!
|
|
|
element90
|
|
« Reply #4 on: March 07, 2014, 12:11:16 PM » |
|
The difficulty with the formula
z = (z^3 + c)(z^2 + c)
is that the critical points can't be easily determined except for the point at zero.
The formula expands to
z = z^5 +cz^3 + cz^2 + c^2
so the critical points are the solutions of
f'(z) = 5z^4 + 3cz^2 + 2cz = 0
so
z(5z^3 + 3cz^2 + c) = 0
which implies that there is a critical point at 0, the remaining 3 critical values have to calculated for each value of c (the location in the complex plane).
|
|
|
Logged
|
|
|
|
|
element90
|
|
« Reply #6 on: March 07, 2014, 01:51:03 PM » |
|
For higher level polynomials from a programming point of view it is even easier, the coefficients can be used to produce the companion matrix and the eigenvalues are the roots, provided you have the routines to do it.. It just has to be done for each location. Implementing this in a script with a program such as Ultra Fractal won't be easy.
I looked at
z = (z^2 + c)(z^2 + c)
so
f'(z) = 4z^3 + 2cz = 0
z(4z^2 + 2c) = 0
which has three critical points 0, -sqrt(-c) and sqrt(-c) which can be plugged into a UF or Gnofract4d script which it turns out produces three identical pictures which isn't surprising as the pictures are properly formed M4 Mandelbrots.
|
|
|
Logged
|
|
|
|
element90
|
|
« Reply #7 on: March 07, 2014, 02:41:38 PM » |
|
The formula z = (z^2 + c)(z^4 + c) has somewhat easier critical points to determine, for the values see the code below in the Gnofract4d script. The critical points are solutions of f'(z) = 6z^5 + 4cz^3 + 2cz = 0 z(6z^4 + 4cz^2 + 2c) = 0 which gives zero and the square roots of the roots of a quadratic. M2.M4 { ; M2 Mandelbrot multiplied by an M4 Mandelbrot init: ; z = 0 z0 = #pixel ; z = sqrt((-z0 + sqrt(z0*z0 - 24*z0))/12) ; z = sqrt((-z0 - sqrt(z0*z0 - 24*z0))/12) z = -sqrt((-z0 + sqrt(z0*z0 - 24*z0))/12) ; z = -sqrt((-z0 - sqrt(z0*z0 - 24*z0))/12) loop: z2 = z*z z = (z2 + #pixel)*(z2*z2 + #pixel) bailout: @bailfunc(z) < @bailout default: float param bailout default = 4.0 endparam float func bailfunc default = cmag endfunc } Critical point 0 https://copy.com/hqSr3VYbJBuOCritical point z = sqrt((-z0 + sqrt(z0*z0 - 24*z0))/12) and sqrt((-z0 + sqrt(z0*z0 - 24*z0))/12) https://copy.com/kqLZT9DwdGFJCritical point z = sqrt((-z0 - sqrt(z0*z0 - 24*z0))/12) and -sqrt((-z0 - sqrt(z0*z0 - 24*z0))/12) https://copy.com/UMOwFuqWj8edCombined https://copy.com/BGUDDqTrGMt7What usually happens is that areas in one critical point picture that lack Mandelbrot buds are provided with them from an other critical point picture. Oddly the areas at the top and the bottom have areas where this isn't so, assuming I haven't made a mistake.
|
|
« Last Edit: May 26, 2014, 02:55:45 PM by element90, Reason: Replaced file links. »
|
Logged
|
|
|
|
jdebord
|
|
« Reply #8 on: March 10, 2014, 10:11:17 AM » |
|
There are analytical formulas for polynomials up to degree 4. See for instance: http://dlmf.nist.gov/1.11#iiiSince we are looking for the roots of the derivative, this applies for polynomials up to degree 5.
|
|
|
Logged
|
|
|
|
Endemyon
Forums Freshman
Posts: 14
|
|
« Reply #9 on: March 31, 2014, 11:21:38 PM » |
|
Hey guys, thanks for your answers ! I can't resist to show you this julia set, obtained by Zn+1=(Zn^2+c)*(Zn^3+c)*(Zn^4+c) with C=0,7937+i0,0904 http://www.fractalforums.com/index.php?action=gallery;sa=view;id=15802I don't understand why you need to find the roots of critical points (if it's not to have a completely connex Julia set ) ? It's possible to have an approximation for higher degrees with the newton's method, but I think you don't learn something new
|
|
|
Logged
|
|
|
|
jdebord
|
|
« Reply #10 on: April 01, 2014, 10:12:46 AM » |
|
The critical point is required only for the Mandelbrot set, not for the Julia set.
The Julia set iteration always start with the pixel coordinates.
|
|
|
Logged
|
|
|
|
Endemyon
Forums Freshman
Posts: 14
|
|
« Reply #11 on: April 01, 2014, 06:47:10 PM » |
|
The critical point is required only for the Mandelbrot set, not for the Julia set. The Julia set iteration always start with the pixel coordinates. If you do a Mandelbrot set for the equation Zn+1=Zn^2+C (the classic). C is the point which parcours the pixel coordinates and Z0=0. So you don't need a critical point So that doesn't not explain me why it's needed, is it from the software you use? Now, why I was talking about critical point for a julia set: If you do a julia set for the equation Zn+1=Zn^2+C. C is the chosen constant and Z0 is the point which is the pixel coordinates. If you try do a julia set where C is the critical point of mandelbrot set, you obtain a julia set where all converging parts are connected. ( for any n)
|
|
|
Logged
|
|
|
|
jdebord
|
|
« Reply #12 on: April 02, 2014, 09:05:34 AM » |
|
If f(z) denotes the iterated formula, the critical points are the roots of the equation f'(z) = 0
For the classical Mandelbrot set, f(z) = z^2 + c and f'(z) = 2z.
The equation is 2z = 0, hence z = 0
So, z0 = 0 is a critical point (the only one, in this case).
|
|
|
Logged
|
|
|
|
TheRedshiftRider
|
|
« Reply #13 on: June 10, 2014, 05:48:45 PM » |
|
I tried something else, I tried to divide the two parts of the basic formula and the results are nice: (z^10+c)/(z^2+c) (z^10+c)/(z^8+c)
|
|
|
Logged
|
Motivation is like a salt, once it has been dissolved it can react with things it comes into contact with to form something interesting.
|
|
|
jdebord
|
|
« Reply #14 on: June 11, 2014, 10:07:54 AM » |
|
Nice pictures, especially the second one ! I have tried to find a formula for the derivatives, assuming p > q > 1 : The formula for f'(c) is needed if you want to apply the distance estimator method (here z' is the value of the derivative at the previous iteration). The formula for f'(z) shows that 0 is always a critical point, but the other critical points depend on c and need to solve a polynomial. I assume that the pictures are for the critical point zero ?
|
|
« Last Edit: June 12, 2014, 09:20:34 AM by jdebord »
|
Logged
|
|
|
|
|