Python and Agent Simulation/Visualization Environments

Question

I am trying to end my dependence on Matlab and have turned to Python. I am soon to begin work implementing machine learning (neural nets/SVMs) in agents (simulate ship traffic). I need a machine learning toolbox and a simulation environment.

There are many ML/Python options. The remaining challenge then is a simulation environment (2D or 3D).

Netlogo is interesting but seems limited in being able to call external functions. It is also Java based. Breve seems perfect but also seems a dead project. Pybrain mentions some environments (PyODE?) but is unclear on detail.

So has anyone some good recommendations on a simulation/visualization environment with a good Python interface?

Answer 1

Mesa是一个基于代理的模拟框架，包括某种基于Web的可视化。

Answer 2

Have just found Simpy - haven't tried it yet but it seems to have been under development for some time:

http://simpy.sourceforge.net/index.html

Answer 3

如果您只需要三维可视化，那么Visual可以非常轻松地将某些内容组合在一起。

Answer 4

VTK seems a very nice option.

http://www.vtk.org/