Instead of converging state probabilities to steady-state t=inf solution, a
duration could be specified and all possible outcomes up to t=n could be
calculated.
Eg: J>Cons
t=0, 1 initial starting state[CS next]
t=1.5, 2 states: J miss[Cons next], J hit[Cons next]
t=3, 2 states: [J next]
...
t=600
This gives us an exact result without having to run simulations for many
timesteps. Some state coalescence logic will be required to prevent an
explosion of states for high values of t.
Original issue reported on code.google.com by iminmm...@gmail.com on 29 Mar 2012 at 6:03
Original issue reported on code.google.com by
iminmm...@gmail.comon 29 Mar 2012 at 6:03