burlap.mdp.singleagent.pomdp.observations

Interface DiscreteObservationFunction

All Superinterfaces:

ObservationFunction

All Known Implementing Classes:

TigerObservations

public interface DiscreteObservationFunction
extends ObservationFunction

Defines additional methods for an ObservationFunction for the case when the set of observations are discrete and able to be enumerated. Requires implementing the allObservations() method and probabilities(State, Action) method. The former should return all possible observations; the latter should return the set of observations with non-zero probability and their probabilities, for a given input hidden state and action that led to the hidden state.

Author:

James MacGlashan.

Method Summary

Modifier and Type	Method and Description
java.util.List<State>	allObservations() Returns a List containing all possible observations.
java.util.List<ObservationProbability>	probabilities(State state, Action action) Returns the observation probability mass/density function for all observations that have non-zero mass/density conditioned on the true MDP state and previous action taken that led to the state.

Methods inherited from interface burlap.mdp.singleagent.pomdp.observations.ObservationFunction

probability, sample

Method Detail

allObservations

java.util.List<State> allObservations()

Returns a List containing all possible observations. Observations are represented with a class that implements the State interface, since observations may consist of multiple observations of distinct objects in the world.

Returns:

a List of all possible observations.

probabilities

java.util.List<ObservationProbability> probabilities(State state,
                                                     Action action)

Returns the observation probability mass/density function for all observations that have non-zero mass/density conditioned on the true MDP state and previous action taken that led to the state. The function is represented as a List of ObservationProbability objects, which is a pair of an observation (represented by a State and double specifying its mass/density.

Parameters:

state - the true MDP state that generated the observations

action - the action that led to the MDP state and which generated the observations.

Returns:

the observation probability mass/density function represented by a List of ObservationProbability objects.