aPRAM is a framework for building probabilistic, relational agent-based models (PRAMs). The central ideas in aPRAM are:
- To vectorize computation for speed;
- To make agents behave stochastically rather than deterministically;
- To make behaviors depend on relations between agents
aPRAM is built on Numpy, so most operations are vectorized and relatively fast. You don't need to know Numpy to build aPRAM models. The tutorial contains a complete aPRAM model of an epidemic process and illustrates most of aPRAM's classes and methods.
- Python 3
- Numpy
- Pandas
- Matplotlib