search

blakeelias / pandemic_RL

Reinforcement learning for economically optimal pandemic response.
GNU General Public License v3.0
2 stars 1 forks source link

Leverage speed-up for deterministic dynamics #31

Open blakeelias opened 3 years ago

blakeelias commented 3 years ago

Don't have to iterate over entire space of next-states; instead can just iterate over the states that are actually reached -- which could be just 1 state.

blakeelias commented 3 years ago

Of course, if I want to account for superspreaders properly, I need to have some stochasticity at least when in the the small numbers. But once cases get above a minimal threshold, can remove the stochasticity and get a big speedup.