This simulation was created as part of course CS-263, taught by
Lee
Spector and Jon Klein. All materials in this page, except the video
below, were generated by that course. Original course materials can be
found at http://hampshire.edu/lspector/cs263/cs263s04.html
There are four versions of the simulation, each adding newer features.
Here is the text describing each version, which is also included in the
ABOUT.txt file in the tar file provided above.
WUBWorld_1
A breve simulation introducing
"WUBWorld", where students can
program autonomous agents to interact
with a dynamic environment.
In this first version, students learn to
direct the agent to
collect pieces of food.
WUBWorld_2
An update to WUBWorld which introduces
obstacles and enemies
into the environment. The
agent must now eat food in order to
survive, while avoiding enemies and
navigating around obstacles.
WUBWorld_3
A version of WUBWorld which includes
code for building neural
networks. It also allows the
agents to attack enemies by
firing towards them. Students
can use the neuron class to make
decisions about when to attack or how to
move.
WUBWorld_4
The final update to WUBWorld introduces
a whole mess of
complications: smarter enemies, enemy
rebirth, and deadly "shock
waves". To cope with the more
complicated world, agents are
given the ability to evolve.
When an agent dies, it can be
reborn, selecting traits and behaviors
from more successful
agents.