closes #97
Changed incremental_update() such that it always uses the last appended reward (except at t0).
Also added the argument terminal_state. This was done to prevent a possible bug with the old check for clean-up. In the special case where not_ended_reward is not unique (e.g. equal to tie_reward) the clean-up would initiate prematurely.
closes #97 Changed
incremental_update()
such that it always uses the last appended reward (except at t0).Also added the argument
terminal_state
. This was done to prevent a possible bug with the old check for clean-up. In the special case where not_ended_reward is not unique (e.g. equal to tie_reward) the clean-up would initiate prematurely.