Welcome to Fractal Forums

Fractal Math, Chaos Theory & Research => IFS - Iterated Function Systems => Topic started by: Endemyon on March 19, 2014, 08:54:16 PM




Title: Overlap on an IFS
Post by: Endemyon on March 19, 2014, 08:54:16 PM
Hello everybody !
I have created a program which creates IFS from L-system principle.
I start from a segment and i replace it by a some connex segments with a scaling process.
After that i replace each segment created by the same set of segments and i keep going.

How can i be sure that the curve i create this way isn't gonna overlap on itself  ?


With all respect,
Endemyon


Title: Re: Overlap on an IFS
Post by: youhn on March 19, 2014, 08:58:32 PM
Easy. Visualize and check.

But I think you mean before actually doing stuff, like a general rule to prevent all situations that lead to self-intersect? I don't think there is. Can you show us the program? Looking at specific stuff leads to more specific answers.


Title: Re: Overlap on an IFS
Post by: Endemyon on March 19, 2014, 09:18:36 PM
Yes, I mean before doing the stuff !
Because, how can I know, if it's just by looking, know if it's gonna intersect at the iteration after which I stopped !
It's a Matlab program, so it could not be easy for you to use it.
But it does pretty much the same thing as the program on this website :
http://www.kevs3d.co.uk/dev/lsystems/

If you try with angle=60, Axiom : F and rule 1: F=F+F--F+F
you get Von Koch's curve (no intersections)

But if you try the same angle and axiom but with rule 1 : F=F++F--F--F++F
you get a Sierpinski Triangle wich intersect itself (for this it's easy to see it's gonna intersect, but for more complicated stuff it may not be)