burlap.behavior.singleagent.planning.stochastic.rtdp

Enum BoundedRTDP.StateSelectionMode

java.lang.Object

java.lang.Enum<BoundedRTDP.StateSelectionMode>

burlap.behavior.singleagent.planning.stochastic.rtdp.BoundedRTDP.StateSelectionMode

All Implemented Interfaces:

java.io.Serializable, java.lang.Comparable<BoundedRTDP.StateSelectionMode>

Enclosing class:

BoundedRTDP

public static enum BoundedRTDP.StateSelectionMode
extends java.lang.Enum<BoundedRTDP.StateSelectionMode>

The different ways that states can be selected for expansion. That is, after an action is selected in the rollout, the next state from the possible outcome states may be selected in these different ways. The default, and reccomended mode is MODELBASED.

MODELBASED is a standard RTDP-like sampling from the transition dynamics. This is the default mode and in practice, we found it gives faster performance than the reccomeneded approach in the Bounded RTDP paper. Additionally, if the domain has an efficient performAction state sampler, this mode will provide some computational gains over the other approaches.

WEIGHTEDMARGIN is the state selection method suggested in the original Bounded RTDP paper, which we found to be somewhat slower than MODELBASED. In this approach, each possible outcome state is selected randomly from a distribution that is the transition dynamics probability weighted by the margin in the lower bound and upper bound of the next state's value function. This promotes exploration towards states that are more uncertain.

MAXMARGIN selects the state with the maximum margin between the lower bound and upperbound of the value function. Ties are broken randomly. This method is much more exploratory, but results in significantly longer initial rollouts.

Author:

James MacGlashan

Enum Constant Summary

Enum Constants

Enum Constant and Description
MAXMARGIN
MODELBASED
WEIGHTEDMARGIN

Method Summary

Modifier and Type	Method and Description
static BoundedRTDP.StateSelectionMode	valueOf(java.lang.String name) Returns the enum constant of this type with the specified name.
static BoundedRTDP.StateSelectionMode[]	values() Returns an array containing the constants of this enum type, in the order they are declared.

Methods inherited from class java.lang.Enum

clone, compareTo, equals, finalize, getDeclaringClass, hashCode, name, ordinal, toString, valueOf

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Enum Constant Detail

MODELBASED

public static final BoundedRTDP.StateSelectionMode MODELBASED

WEIGHTEDMARGIN

public static final BoundedRTDP.StateSelectionMode WEIGHTEDMARGIN

MAXMARGIN

public static final BoundedRTDP.StateSelectionMode MAXMARGIN

Method Detail

values

public static BoundedRTDP.StateSelectionMode[] values()

Returns an array containing the constants of this enum type, in the order they are declared. This method may be used to iterate over the constants as follows:

for (BoundedRTDP.StateSelectionMode c : BoundedRTDP.StateSelectionMode.values())
    System.out.println(c);

Returns:

an array containing the constants of this enum type, in the order they are declared

valueOf

public static BoundedRTDP.StateSelectionMode valueOf(java.lang.String name)

Returns the enum constant of this type with the specified name. The string must match exactly an identifier used to declare an enum constant in this type. (Extraneous whitespace characters are not permitted.)

Parameters:

name - the name of the enum constant to be returned.

Returns:

the enum constant with the specified name

Throws:

java.lang.IllegalArgumentException - if this enum type has no constant with the specified name

java.lang.NullPointerException - if the argument is null