Sunday, April 28, 2013

Virtual Evolution

A research team at Cornell University has created a computer algorithm that simulates evolution. The goal of the creatures is to move as fast and efficiently as possible. The creatures produced by the algorithm worked much better than those designed by humans. This program demonstrates the processes of natural selection because the creatures that can move farther are used as the basis for the next generation of creatures.

NOS Themes:
1. science is based on evidence - the algorithm is based on "developmental biology"
2. role of chance - the creatures that preformed better were used to create new creatures
3. role of curiosity - this program demonstrates evolution that normally takes thousands of years in seconds

Article: http://www.sciencedaily.com/releases/2013/04/130423135843.htm
Video: http://www.youtube.com/watch?v=z9ptOeByLA4&feature=youtu.be
 An interesting program that is similar: http://boxcar2d.com/

2 comments:

  1. This is a very interesting article Michael! I was wondering if this program accounts for factors like predation and fitness level? Also what time period is this species supposed to be from? One thing I found interesting with this is that the this program is goal oriented whereas evolution isn't and I thought it was cool to look at the goal oriented side of it! I found another cool website on this subject if you want to check it out. Link:
    http://media.hhmi.org/biointeractive/vlabs/stickleback/index.html

    ReplyDelete
  2. This sounds like the thing we did in class with the origami birds. It would be interesting to include factors other than just the ability to move in the selection process. The programmers could make an entire imaginary ecosystem, complete with predators, diseases, and famines, to see what the population looked like over time.

    ReplyDelete