MDPs and POMDPs in Julia - An interface for defining, solving, and simulating fully and partially observable Markov decision processes on discrete and continuous spaces.
Currently we have MDP{S, A} and POMDP{S, A, O}. The type parameters don't serve much of a purpose.
This was convenient when the only way to define an MDP or POMDP was using the object-oriented method because it saved users from having to specify statetype, actiontype and obstype, but now QuickPOMDPs can detect this automatically.
Instead, we should just require that new problem types implement the XXXtype functions directly.
Though this was probably a mistake in hindsight, it would be difficult to fix and only offer a very small benefit, so I'm going to classify it as not-planned for now.
Currently we have
MDP{S, A}
andPOMDP{S, A, O}
. The type parameters don't serve much of a purpose.This was convenient when the only way to define an MDP or POMDP was using the object-oriented method because it saved users from having to specify
statetype
,actiontype
andobstype
, but now QuickPOMDPs can detect this automatically.Instead, we should just require that new problem types implement the
XXXtype
functions directly.