burlap.domain.singleagent.gridworld

Class GridWorldDomain.GridWorldModel

java.lang.Object

burlap.domain.singleagent.gridworld.GridWorldDomain.GridWorldModel

All Implemented Interfaces:

FullStateModel, SampleStateModel

Enclosing class:

GridWorldDomain

public static class GridWorldDomain.GridWorldModel
extends java.lang.Object
implements FullStateModel

Nested Class Summary

Nested classes/interfaces inherited from interface burlap.mdp.singleagent.model.statemodel.FullStateModel

FullStateModel.Helper

Field Summary

Fields

Modifier and Type	Field and Description
protected java.util.Random	rand
protected double[][]	transitionDynamics Matrix specifying the transition dynamics in terms of movement directions.

Constructor Summary

Constructors

Constructor and Description
GridWorldModel(int[][] map, double[][] transitionDynamics)

Method Summary

Modifier and Type	Method and Description
protected int	actionInd(java.lang.String name)
protected State	move(State s, int xd, int yd) Attempts to move the agent into the given position, taking into account walls and blocks
State	sample(State s, Action a) Samples and returns a State from a state transition function.
java.util.List<StateTransitionProb>	stateTransitions(State s, Action a) Returns the set of possible transitions when Action is applied in State s.

Methods inherited from class java.lang.Object

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

Field Detail

transitionDynamics

protected double[][] transitionDynamics

Matrix specifying the transition dynamics in terms of movement directions. The first index indicates the action direction attempted (ordered north, south, east, west) the second index indicates the actual resulting direction the agent will go (assuming there is no wall in the way). The value is the probability of that outcome. The existence of walls does not affect the probability of the direction the agent will actually go, but if a wall is in the way, it will affect the outcome. For instance, if the agent selects north, but there is a 0.2 probability of actually going east and there is a wall to the east, then with 0.2 probability, the agent will stay in place.

rand

protected java.util.Random rand

Constructor Detail

GridWorldModel

public GridWorldModel(int[][] map,
                      double[][] transitionDynamics)

Method Detail

stateTransitions

public java.util.List<StateTransitionProb> stateTransitions(State s,
                                                            Action a)

Description copied from interface: FullStateModel

Returns the set of possible transitions when Action is applied in State s. The returned list only needs to include transitions that have non-zero probability of occurring.

Specified by:

stateTransitions in interface FullStateModel

Parameters:

s - the source state

a - the action to be applied in the source state

Returns:

the probability distribution of the state transition function specified as a list of StateTransitionProb objects.

sample

public State sample(State s,
                    Action a)

Description copied from interface: SampleStateModel

Samples and returns a State from a state transition function.

Specified by:

sample in interface SampleStateModel

Parameters:

s - the source state

a - the action to be executed in the source state

Returns:

a sample State

move

protected State move(State s,
                     int xd,
                     int yd)

Attempts to move the agent into the given position, taking into account walls and blocks

Parameters:

s - the current state

xd - the attempted new X position of the agent

yd - the attempted new Y position of the agent

Returns:

input state s, after modification

actionInd

protected int actionInd(java.lang.String name)