Share with us how are you using the powerrrr of the force
Moderator: Alain
by Marmakoide » Wed May 03, 2006 9:45 am
Thomas wrote:wow, very cool! i am looking forward to the new creatures.
![Smile :)](./images/smilies/icon_smile.gif)
(the vlc player (
www.videolan.org) played the videos without the need for any codec.)
i am just curious. how far do such evolutionary simulations go currently? what would be possible with the world's fastest super computer for example? i find this very interesting.
![Smile :)](./images/smilies/icon_smile.gif)
If I use all the computing power of the lab, I can find how to control any robotic creature in a few minutes of computing, because genetic algorithm are very easy to run on computer cluster. Here it is the steps of my research :
1) Fast evolution of static morphologies (tower, bridges, etc...)
on progress
2) Fast evolution of the control for a given morphology (robot)
on progress
3) Simultaneous evolution of control and morphology
The videos are the beginning of the point 2). Wit the point 3) I hope I can produce autonomous robots in an automatic way. Robots that could produce robots, robots that could survive in a famous red desert planet, etc... And be, of course the master of the Earth and get my vengeance against.... a well, I will find what after.
"Let's rock baby..."
-
![User avatar](./download/file.php?avatar=700.jpg)
Marmakoide
-
- Posts: 6
- Joined: Tue Jun 14, 2005 5:38 pm
- Location: Paris (France)
-
by grod » Fri Sep 15, 2006 11:25 am
Those videos were amazing!
I'm currently working on a similiar project for school, with some kind of walker, or maye even a "jumper" if the chromosomes would be given the time that would take to evolve in that way. =)
Anyway, my problem is that I have a population of about 40-50 chromosomes, and each cycle of simulation is about 3-4 seconds, and I want to let them evolve for something like 200 generations.. That will take too much time to simulate in real-time. Is there any way to simulate the physics with newton faster than real-time?
I was thinking of just render the fittest of all chromosomes in each generation while the program simulates the next generation in the backround.
-
grod
-
- Posts: 1
- Joined: Fri Sep 15, 2006 11:12 am
by JernejL » Fri Sep 15, 2006 1:54 pm
grod wrote:Those videos were amazing!
I'm currently working on a similiar project for school, with some kind of walker, or maye even a "jumper" if the chromosomes would be given the time that would take to evolve in that way. =)
Anyway, my problem is that I have a population of about 40-50 chromosomes, and each cycle of simulation is about 3-4 seconds, and I want to let them evolve for something like 200 generations.. That will take too much time to simulate in real-time. Is there any way to simulate the physics with newton faster than real-time?
I was thinking of just render the fittest of all chromosomes in each generation while the program simulates the next generation in the backround.
what about you pre-generating genetics and put them into game later?
-
![User avatar](./download/file.php?avatar=446_1516040955.png)
JernejL
-
- Posts: 1578
- Joined: Mon Dec 06, 2004 2:00 pm
- Location: Slovenia
-
by bones » Thu Nov 23, 2006 7:46 am
The hexapod is brilliant. I'm trying to make a simular model for a game I'm working on but I'm rubbish with joints being new to newton. You could save me alot of time if you could supply maybe a small example program so I can learn how you did it. I don't want to steal any ideas though.
![Very Happy :D](./images/smilies/icon_biggrin.gif)
-
bones
-
- Posts: 1
- Joined: Fri Nov 17, 2006 10:02 am
Return to User Gallery
Who is online
Users browsing this forum: No registered users and 24 guests