Tuesday, 14 October 2008

My Artificial Life Simulator Shortlist


Polyworld is another artificial life simulator.

I was going to take it off my AL simulator short list until I saw a video presentation on Google Tech Talk presented by Virgil Griffith, a student who is continuing the project. It looks like the project is still going on.

I find it interesting that people are working on adaptations through physical mutations and personal learning but not through cultural development. By cultural development I am referring to having language emerge and to having creatures modify the environment.

Larry Yaeger, is the original developer.

I wonder how easy it is to set up the system with two agents who communicate ...

I might still take it out of my short list because I prefer the world to have physics. That leaves breve and Framsticks on my list. Framsticks seems to include a neural network component but breve requires a neural network component to be tacked on.

I like how breve can display the a view of the world from the perspective of a selected agent.


The Framsticks GUI is shareware which makes me feel iffy.



  1. Hello,

    I've started using breve myself for artificial evolution experiments. I did not do as much research as you did before settling for this particular environment, but I'm quite happy with it. You can use it with the NEAT neural network package (neat-python on google code).

  2. Hi wigglything!

    Thanks for letting me know about your experiences with breve. I'ld like to hear more about what you're doing.

    I chose Polyworld over breve because Polyworld has a rich digital ecosystem aspect that allows for pure natural selection which is more useful for what I am trying to do which is coaxing language development.

    Since breve does not seem to have an ecosystem, my hunch is that it requires a user defined fitness function. Since that is just my hunch, I'll definitely give breve a go.