!!! There is a bug in the pyCRLD package regarding the average reward Ri computation. A hack around is to first compute the stationary observation distribution.
I'm not entirely sure what it is, maybe @wbarfuss can point it out.
Here's the work around:
!!! There is a bug in the pyCRLD package regarding the average reward
Ricomputation. A hack around is to first compute the stationary observation distribution.I'm not entirely sure what it is, maybe @wbarfuss can point it out.
Here's the work around:
