burlap.mdp.singleagent.pomdp.observations

Interface ObservationFunction

All Known Subinterfaces:

DiscreteObservationFunction

All Known Implementing Classes:

TigerObservations

public interface ObservationFunction

Defines the observation function and observations associated with a POMDP domain (PODomain). Requires two methods: probability(State, State, Action) and sample(State, Action). The former defines the probability mass/density function for an observation given a hidden state and action that led the hidden state. The latter samples and observation given a hidden state and action that led to the hidden state.

Method Summary

Modifier and Type	Method and Description
double	probability(State observation, State state, Action action) Returns the probability that an observation will be observed conditioned on the MDP state and previous action taken that led to the state.
State	sample(State state, Action action) Samples an observation given the true MDP state and action taken in the previous step that led to the MDP state.

Method Detail

probability

double probability(State observation,
                   State state,
                   Action action)

Returns the probability that an observation will be observed conditioned on the MDP state and previous action taken that led to the state.

Parameters:

observation - the observation, represented by a State

state - The true MDP state that generated the observation.

action - the action that led to the MDP state and which generated the observation

Returns:

the probability of observing the observation.

sample

State sample(State state,
             Action action)

Samples an observation given the true MDP state and action taken in the previous step that led to the MDP state.

Parameters:

state - the true MDP state

action - the action that led to the MDP state

Returns:

an observation represented with a State.