States and accentuators
States and deterministic environments
Intertemporal decision making
Lotteries and risk aversion
Fully observable Markov Decision Processes (MDP)
Decision making with unknown uncertainty
Sensors and Partially Observable Markov Decision Processes (POMDP)
Deep sequential games
Repeated episodic games
Games with imperfect or incomplete information
For environment with no actors, transition model is matrix (for finite state).
Can multiply matrix n times to get state in n periods.
Transition matrix is perumation matrix. 1 or 0 only.
Does time stop to allow decisions. eg chess v driving.