Omnimaga

General Discussion => Technology and Development => Computer Programming => Topic started by: ruler501 on February 25, 2011, 08:41:36 pm

Title: Genetic Programming
Post by: ruler501 on February 25, 2011, 08:41:36 pm
If you do not know about this there is a website here: http://www.geneticprogramming.us/

I am trying to learn genetic programming/create a new library for it in python.
I thought this would be a good place to create a discussion on it and possible implementations and uses. Please post any thoughts you have on the subject
Title: Re: Genetic Programming
Post by: jnesselr on February 25, 2011, 10:44:22 pm
I like how it essentially randomly chooses stuff, so you can give criteria even though you don't fully know where it's gonna go.
Title: Re: Genetic Programming
Post by: AngelFish on February 25, 2011, 10:45:22 pm
The only problem is that it'll take a supercomputer a week to solve even a simple problem with genetic programming :P
Title: Re: Genetic Programming
Post by: ruler501 on February 25, 2011, 10:49:45 pm
You can get it working well. I believe there are efficient algorithms that will make it in python take less than a minute on an average computer
Title: Re: Genetic Programming
Post by: AngelFish on February 25, 2011, 10:51:58 pm
Depends on the problem. Genetic programming is a method that inherently relies on randomness, so it will only converge to a solution (if it can find one) with quite a bit of time. It's definitely more inefficient than most analytical algorithms.
Title: Re: Genetic Programming
Post by: ruler501 on February 25, 2011, 10:53:48 pm
It can be applied for certain things to work better. I found some good uses for this.
Is there a better way to find polynomial approximations for functions?
Title: Re: Genetic Programming
Post by: AngelFish on February 25, 2011, 10:55:21 pm
Gaussian Quadrature is an excellent method of approximating functions.