Framework working

First tests being done using inductive q-learning.
Policy can be learned and backtested.

Update function needs to be extended to do recursive state (t, i) visiting instead of running the entire period (t) at once.