We closed this forum 18 June 2010. It has served us well since 2005 as the ALPHA forum did before it from 2002 to 2005. New discussions are ongoing at the new URL http://forum.processing.org. You'll need to sign up and get a new user account. We're sorry about that inconvenience, but we think it's better in the long run. The content on this forum will remain online.
Index › Programming Questions & Help › Programs › easiest way to evaluate a 3d structure
‹ Previous Topic | Next Topic ›
Page Index Toggle Pages: 1
easiest way to evaluate a 3d structure (Read 1166 times)
sebastiano
easiest way to evaluate a 3d structure
Feb 22nd, 2010, 8:05am  
Hi i would like to know which is the easiest way to evaluate 3d structures.
I made some code in processing that generates random 3d structures, and i would like to experiment with 3d artificial evolution, so i need to evaluate my structures in order to evolve them,   it doenst mind to have functional structures or evaluate using a functional criteria ,   i just want to know which is the easiest way to evaluate 3d structures.

thanks

seb.
Back to top
 
 
Cedric
Re: easiest way to evaluate a 3d structure
Reply #1 - Feb 22nd, 2010, 8:28am  
evaluate regarding what criterias?
you cant just say, this structure is better than the other one...
Back to top
 
... twitter.com/CedricKiefer
WWW CedricKiefer 9570393 CedricKiefer  
Cedric
Re: easiest way to evaluate a 3d structure
Reply #2 - Feb 22nd, 2010, 9:03am  
This Topic was moved here from Syntax Questions by Cedric.
Back to top
 
... twitter.com/CedricKiefer
WWW CedricKiefer 9570393 CedricKiefer  
Chrisir
Re: easiest way to evaluate a 3d structure
Reply #3 - Feb 22nd, 2010, 11:03am  
Hello,

I like the idea.

Sounds like
Evolutionary algorithm
http://en.wikipedia.org/wiki/Evolutionary_algorithm

Now you need a
http://en.wikipedia.org/wiki/Fitness_function

What I can think of is:
  • size (add volume of all parts of one object)
  • Symmetry (e.g. first approach: take one plane through the object and calculate volume left and right of the plane, do this for x- and y- and z-plane)
  • complexity (hm... number of parts?)
  • the inverse of the similarity towards e.g. a simple cube
  • sheer beauty (hmmm... Wink )

You also could do random for a while and then let the program stop and ask you for the 5 best of the last 50 objects and then continue random with those and ask you again after a while and so forth...
see http://en.wikipedia.org/wiki/Interactive_genetic_algorithms

Greetings,

Chrisir
Back to top
 
 
Cedric
Re: easiest way to evaluate a 3d structure
Reply #4 - Feb 22nd, 2010, 11:26am  
i still think you need to answer "what is best" in this case.
does it have to fulfill a special task?
anyway. there is this GA library that is definitely a good start for exploring GA. http://robotacid.com/PBeta/AILibrary/GA/index.html

and this post by dan shiffmann http://www.shiffman.net/teaching/nature/ga/
Back to top
 
... twitter.com/CedricKiefer
WWW CedricKiefer 9570393 CedricKiefer  
sebastiano
Re: easiest way to evaluate a 3d structure
Reply #5 - Feb 23rd, 2010, 1:22am  
hello thanks, for the replies , cedric , Chrisir.
Cedrid i think you dont undesrtand what i said , i dont want a specific result for something. or solving a real problem, i dont need to fulfill a special task.

I just was thinking about different ways of evaluating 3d structures and i just was curious about which of all possibilities was the easiest to implement.
So im not interested in the result but im intersted in the easiest way to  implement a evaluating function. So i dont have any criteria, and my question should be, which could be the best criteria for making the easiest evaluating 3d structure function?

Before i  was thinking about comparing 2 3d structures, the one that is generated by my program and other that could be the "ideal" one.

but i think the easiest way must be size or complexity (number of parts) as Chrisir says .
Symetry is a really nice idea idea for evaluating function i didnt think about that.
And Chrisir what do you mean with the inverse of the similarity towards? i cant understand that one.
Back to top
 
 
Chrisir
Re: easiest way to evaluate a 3d structure
Reply #6 - Feb 23rd, 2010, 4:55am  


Hello,

By
Quote:
inverse of the similarity towards

I meant:
to measure how much "unsimilar" your new object is from a plain box.

How much your new object is different from a box.

But I don't know a pretty formular for that one...  Cry

Greetings,

Chrisir



Back to top
 
 
Cedric
Re: easiest way to evaluate a 3d structure
Reply #7 - Feb 23rd, 2010, 5:42am  
I did understand you, but you still need to know what you actually want to evaluate.  if it is a special task, fuction, qualities or characteristics of your structure like Symetry or complexity you mentioned. you could also just decide by favour what is the seed for your next generation. but this is also some kind of evaluation... or what else do you mean by "evaluate my structures in order to evolve them," if not that.
Back to top
 
... twitter.com/CedricKiefer
WWW CedricKiefer 9570393 CedricKiefer  
Page Index Toggle Pages: 1
‹ Previous Topic | Next Topic ›