Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Given a computer program generating the Mandelbrot Set - using this one for example, which uses a module called mandel.js - what would be the pseudocode to find the complex coordinates, capable of creating a highlighted point representing the tangent where that co-prime's bulb connects, from two entered p and q values?

Refer to the passage here and the accompanying equation beginning C p/q = ...

So far I have tried:

part1 = (E^(2.0 * PI * i * (p div q))) div 2.0
part2 = 1.0 - (E^(2.0 * PI * i * (p div q))) div 2.0

Where p and q are co-primes, PI and E are the expected constants, and i is set to 1 or -1 (not sure of the correct handling of i in this context).

This simply tries to solve two parts of the equation given p and q, and does not seem to work. I am having trouble (1) programmatically translating the fundamental equation to code, and (2) conceptually relating the resulting parameter solution set for co-primes given p and q in the equation referenced, which I don't fully understand, to the complex plane as the standard screen geometry x,y pixel-painting used in the program, which I can understand.

SOLUTION:

Thanks to Jim's helpfulness, here is the pseudocode I came up with and tested in a working fork of the M-set program mentioned:

1 epower = ((p * 2 * PI)/q)
2 expr = CPLX(COS(epower),SIN(epower))
3 divBy = CPLX(2,0)
4 subFrom = CPLX(1,0)
5 term1 = CPLX.divide(expr,divBy)
6 term2 = CPLX.subtract(subFrom,term1)
7 result = CPLX.multiply(term1,term2)
8 x = result.real
9 y = result.imag

Line 1 is the complex power of e in the equation. Line 2 required a bit of research - it is apparently an application of de Moivre's formula. I imported and stitched together a custom complex number object from two examples I found on the web. This library is indicated as CPLX. It's constructor takes two parts, a real and imaginary, to create a complex variable. It then supports the basic operations. Lines 3 and 4 are just real numbers turned into complex number form to support the operations in lines 5, 6, 7. Lines 8 and 9 return the resulting complex number parts, which can be translated into the screen geometry of the M-set display.

share|improve this question
    
Please do not post in the imperative. This seems to me a huge assignment to give free help. What have you done? Where are you stuck? –  Ross Millikan Jun 8 '11 at 20:20
    
@Ross - I have tried to fix the question. Yes it is significant effort. If it helps, I plan to cross-post a good answer to the referenced Wikipedia article - there is very good pseudocode to generate M-sets given there already, and I believe both equations and pseudocode are essential to the many interest and skill levels that might be approaching this. –  KTys Jun 8 '11 at 21:02
    
From what I understand, you want the program to take $p$ and $q$ as input, and then draw a point at $c_{p/q}$, where $c_{p/q}$ is defined by the formula in the Wikipedia article. What I don't understand is your description of what you're having trouble with. Shouldn't this just be a matter of plugging the formula into the program and then drawing the point? –  Jim Belk Jun 8 '11 at 21:52
    
@Jim - I tried plugging and solving. The expected range for the M-set is -2 to 1 on the real and -1 to 1 on the imaginary axes. I could not get values in that range - so presumably I did not set up or perform your recommended calculation correctly. An example, based on the referenced equation or other method, of finding the real and imaginary coordinates for the p/q ratios 1/4, 2/5, 1/3 (these can be readily checked visually) would be a good answer to the question. –  KTys Jun 8 '11 at 23:51
add comment

1 Answer

up vote 1 down vote accepted

Here's an example of how to compute a $c_{p/q}$ value in the case where $p/q=1/3$. The formula given by Wikipedia is $$ c_{p/q} \;=\; \frac{e^{2\pi i p/q}}{2}\left(1-\frac{e^{2\pi i p/q}}{2}\right)\text{,} $$ so \begin{align*} c_{1/3} \;&=\; \frac{e^{2\pi i/3}}{2}\left(1-\frac{e^{2\pi i/3}}{2}\right) \\ \\ &=\; \frac{-0.5 + 0.866025 i}{2}\left(1-\frac{-0.5 + 0.866025 i}{2}\right) \\ \\ &=\; (-0.25 + 0.433013 i)(1.25-0.433013 i) \\ \\ &=\; -0.125 + 0.64952 i. \end{align*}

Edit: In general, the formula for $c_{p/q}$ using real numbers is $$ c_{p/q} \;=\; \frac{2 \cos (2 \pi p/q)-\cos (4 \pi p/q)}{4} \;+\; 2 \sin ^3(\pi p/q) \cos (\pi p/q)\; i $$

share|improve this answer
    
This is correct, and it definitely counts as half the answer. How would one construct an algorithm to automate this procedure? –  KTys Jun 9 '11 at 3:36
    
Are you using a programming language that doesn't support complex numbers? Why isn't the formula itself an algorithm? –  Jim Belk Jun 9 '11 at 5:16
    
Yes - javascript needs an outside library for complex numbers. –  KTys Jun 9 '11 at 16:44
    
OK, I've added a formula that avoids complex numbers. –  Jim Belk Jun 9 '11 at 16:56
    
Thanks Jim, I really appreciate the help. As you can see I worked from your first solution and eventually got it, but I will also test the second. –  KTys Jun 9 '11 at 17:05
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.