N-Bodies Simulator

Overview
  The idea is similar to the N-Body Simulator, except instead of dealing with
  stars, it would instead deal with living organisms. There is no problem being
  solved, but rather just a cool idea. There would be a basic framework in
  in which creatures would exist, with food sources and other terrain.
  
  The simulation would run for some number of time units with some random set
  of animals and choose the most viable (based on some score function) of those
  animals. These would have random mutations and a new round of simulation
  would be run. It will not require that much single-step computation, but the
  goal would be to run the simulation for many, many iterations and see what
  animals it creates.
  
Solution/System Architecture
  The hardest part of the project, far and away, would be modeling the animals
  and their possible interactions in the word in a way that is not artificially
  limiting. Coming up with a good set of building blocks of both physical parts
  and interaction parts (eat, hunt, mate, sleep, etc) will be a large design
  challenge.
  
  The rest of the code will be fairly simple, involving spawning many different
  worlds, each with a different set of creatures. These worlds would run for
  some number of steps and constitute the map phase. In the reduce phase, the
  highest scoring creatures (surivivability? destruction of opposing lifeforms?
  any other metric we can think of) will be selected, and random mutations of
  those plus more random creatures will form the backbone of the next round of
  simulation.
  
  Basic things creatures would have are some form of energy and sustenance
  system, such that they need to rest and eat, and having more body parts would
  increase these needs. Optionally, they would also be able to have social
  interactions, but this is admittedly much tougher. They should also, if time
  permits, maintain a list of their evolutions to see how they got to their
  current state.
  
Development Plan
  Week 1: Have in place a generic system of actions/body parts that can be
          put together and react in reasonable, but inventive ways.
  
  Week 2: Continue evolving the possible actions and other rules of the
          simulation.
          
  Week 3: Continue development on infrastructure and have a finished world
          simulation running. Try several iterations and see what happens.
          
  Week 4: Fine tune and code like hell. Run the world for a long period of time
          and find both very succesful and very pitiful creatures to present.
          
Feasability Rationale
  A basic system is implementable in a day. A cool system could take much, much
  longer than anyone can imagine. The goal would be to get something moderately
  cool without failing your other classes.