Clean separation between system parameters and state variables

[achimstruve]

I suggest that we introduce a clean separation between system parameters, which are fixed values and remain fixed throughout the whole simulation and state variables, which potentially change throughout the simulation.

This ties in to issue #1

[BlockBoy32]

That makes sense, I still figuring out the best strategies for building these models so I through I would use them sort of hand and hand in case to keep it simple. I think building a more robust class for each agent would solve a lot of these problems, I just wanted to stand something up first then look back to iterate after. There is a lot of inefficiencies when it comes to specific agents and the conditions in policies because of how the data is stored.