If the state space of a Markov decision process has size 4, the action space has size

Question:

If the state space of a Markov decision process has size 4, the action space has size 3, and all actions are admissible at all states, then how many stationary policies are there? How many admissible feedback policies are there for a finite horizon problem with terminal time $T=5$ ?

Fantastic news! We've Found the answer you've been seeking!