AI and Design of an El-fish like simulator environment?

Question

Hi there, first post here on stack overflow, hoping to get some advice on how to construct a simulation program akin to the 1993 maxis simulator known as El-Fish wiki here , Also, game info here .

Are there known "Simulation system" algorithm groups that can function and create real life interaction etc... e.g. the visualization known as 'flocking' ? Or, is there an open-source code base to study off of already in construction?

Programming wise, would this also be able to be easily done in a purely functional language? if done in an OOP way, i was thinking of prototyping it in python.

Anyways thanks for any direction in pointing me towards a good starting place. I hope to build a graphical view of an idea/data world. It will be hopefully controlled by underlying simulation AI(heuristics maybe?)

Answer 1

In terms of simulation systems, I recommend you search for "agent-based modeling" software. There are a lot of free toolkits available. The two I like the most are NetLogo and Repast.

Also, it looks like you are implementing a "genetic algorithm". There are many good books and pages on that topic.

Python is good, but so are many other languages.

Most of your time will be spent doing the graphics: animating the fish so they look realistic. Unless you can find a free fish-animation-library.

Answer 2

I recommend you try my own GarlicSim framework. It's written in Python and you'll be writing your specific simulation in Python.

It can definitely handle the kind of simulation you want. There are tutorials available which will teach you the basics of GarlicSim in 30 minutes. I'll be happy to help you build your simulation package, just say hello on the mailing list and I'll guide you from there.

Answer 3

I'm not sure about "real life", but there is a flocking algorithm called boids that might be a good example to start from. There are a couple python versions of it as well. There's one that is an example in the owyl project on google code.