alife

The alife sub-project tries to create a virtual population, that is capable of surviving in a virtual environment. this will make the actual world more active and may deliver some interesting results.
status
The current approach uses neural networks to learn different tasks. The Reproduction and Mutation is done by the populationmanager. An agent exploring the world has been created, but due to some missing features, at the moment it is not possible to further improve it to an agent, that tries to survive by consuming food.
current activity
coevolution: one hunter-agent tries to catch a prey-agent.
Hunter hunting prey, that does not move, are working. Next step is to evolve prey, that escape those hunters.
how to
To run the alife subproject, you first need access to the aep CVS server. There you can download the alife project. Build the jar-file and add it to your runtime-workspace. The current files are configured for a two agent scenario, so best use the aepconfig-full-alife-hunt configfile. The Nativemodules have to be under a project named test. You find them under comp.micropsi.modules. The agents now wait for a connection with the populationmanager (port 7720 for hunter-agent, port 7721 for prey-agent). You can dowload it under http://www.informatik.hu-berlin.de/~hild/EvoRob/experimentalkit.html