Temporal Structure

In order to deterministically generate events over time, we need a model that does a few things:

• generates from a seed and keeps things deterministic all the way down
• generates discrete events over time in realistic patterns by incorporating:
  • Day/Night
  • Daily/Weekly Work Hours
  • Independent factor (ARMA)
  • dependent factor (group dynamic, etc)
• generates unique activity for each actor
  • that is still beholden to dependent factors

At the event level, we'll need the following:

• when did the event occur?
• who is the actor for this event?
• a long representing the seed for generating this event.
  • should be a function of
    • the simulation (base) seed
    • the time
    • the actor
• ~~a double representing the coarse "effectiveness" of the actor when the event occurs~~
  • ~~harder to reason about this one, as no value/desirability is available~~