burlap.mdp.singleagent.pomdp.beliefstate

Class TabularBeliefUpdate

java.lang.Object

burlap.mdp.singleagent.pomdp.beliefstate.TabularBeliefUpdate

All Implemented Interfaces:

BeliefUpdate

public class TabularBeliefUpdate
extends java.lang.Object
implements BeliefUpdate

A BeliefUpdate that operates on TabularBeliefState instances. Computation is exhaustive and performs the exact Bayesian update.

Author:

James MacGlashan.

Field Summary

Fields

Modifier and Type	Field and Description
protected PODomain	domain
protected StateEnumerator	stateEnumerator

Constructor Summary

Constructors

Constructor and Description
TabularBeliefUpdate(PODomain domain)
TabularBeliefUpdate(PODomain domain, StateEnumerator stateEnumerator)

Method Summary

Modifier and Type	Method and Description
PODomain	getDomain()
StateEnumerator	getStateEnumerator()
void	setDomain(PODomain domain)
void	setStateEnumerator(StateEnumerator stateEnumerator)
BeliefState	update(BeliefState belief, State observation, Action a) Computes a new belief distribution from a previous belief and given a new observation received after taking a specific action.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

domain

protected PODomain domain

stateEnumerator

protected StateEnumerator stateEnumerator

Constructor Detail

TabularBeliefUpdate

public TabularBeliefUpdate(PODomain domain)

TabularBeliefUpdate

public TabularBeliefUpdate(PODomain domain,
                           StateEnumerator stateEnumerator)

Method Detail

getDomain

public PODomain getDomain()

setDomain

public void setDomain(PODomain domain)

getStateEnumerator

public StateEnumerator getStateEnumerator()

setStateEnumerator

public void setStateEnumerator(StateEnumerator stateEnumerator)

update

public BeliefState update(BeliefState belief,
                          State observation,
                          Action a)

Description copied from interface: BeliefUpdate

Computes a new belief distribution from a previous belief and given a new observation received after taking a specific action.

Specified by:

update in interface BeliefUpdate

Parameters:

belief - the prior belief

observation - the conditioned POMDP observation defined by a State instance.

a - the conditioned action selection in the previous time step.

Returns:

the new belief state distribution represented by a new BeliefState instance.