Uses of Class
net.pakl.rl.State

Packages that use State

net.pakl.rl	These are the basic reinforcement learning model classes -- every reinforcement learning problem can be described (minimally) as a World (collection of states), Policy, ValueFunction, and Actions.
net.pakl.rl.maze	World and policy for a 2D problem with impassable obstacles.
org.eyelanguage.rl.reading	Code for the Adaptive Reading Agent; see ReadingMain for parameters and default values.

Uses of State in net.pakl.rl

Subclasses of State in net.pakl.rl

class

StateActionPair Can be used to convert a ValueFunction into a Q-function.

Fields in net.pakl.rl declared as State

protected State

Agent.state

Methods in net.pakl.rl that return State

State	World.getNewState(State oldState, Action action) Given a state and an action, returns the resulting state.
State	World.getStartingState()

Methods in net.pakl.rl that return types with arguments of type State

java.util.Set<State>

ValueFunctionHashMap.getKeySet()

Methods in net.pakl.rl with parameters of type State

void	ActionSet.allowAction(State state, Action newAction) This method allows you to program in a possible action given a state.
void	Agent.experience(ValueFunction newValueFunction, ValueFunction valueFunction, State startState, State nextState, double reinforcement)
void	PolicyExtractor.forceInitialState(State s)
java.util.List	ActionSet.getAllPossibleActions(State state) This method returns a list of all possible actions an agent could take given a State.
Action	Agent.getBestActionForValueFrom(State s, ValueFunction vf, ActionSet p) This is a function which should probably be called more often to reduce duplicated code.
State	World.getNewState(State oldState, Action action) Given a state and an action, returns the resulting state.
Action	ActionSet.getRandomAllowedAction(State state)
double	ReinforcementFunction.getReward(State state, Action action) Return the reward given an action from a particular state.
double	ReinforcementFunction.getReward(State state, Action action, State newState) Sometimes a reward is also based on the resulting state.
double	ValueFunctionResidualAlgorithmPerceptron.getValue(State state)
double	ValueFunctionResidualAlgorithmLinear.getValue(State state)
double	ValueFunctionPerceptron.getValue(State state)
double	ValueFunctionHashMap.getValue(State state) The value of a state is defined as the sum of the terinforcements received when starting in that state and following some fixed policy to a terminal state; the optimal policy would map states to actions that maximizes the sum of reinforcements received when starting in an arbitrary state and performing actions until the terminal state is reached
double	ValueFunction.getValue(State state) Retrieve the value associated with a state (which may be different for non-stored states depending on the actual class implementing this value function).
boolean	IsWinnable.isDrawState(State s)
boolean	IsWinnable.isLoseState(State s)
boolean	World.isTerminalState(State state) Reports whether this particular state is a terminal state.
boolean	IsWinnable.isWinState(State s)
protected void	Agent.performValueIterationUpdateOnState(ValueFunction newValueFunction, ValueFunction valueFunction, State currentState)
void	ActionSet.removeAction(State state, Action oldAction) This method allows you to deprogram an action from being possible
void	ReinforcementFunction.setReward(State state, Action action, double newReward)
void	Agent.setState(State newState)
void	ValueFunctionResidualAlgorithmPerceptron.setValue(State state, double newValue)
void	ValueFunctionResidualAlgorithmLinear.setValue(State state, double newValue)
void	ValueFunctionPerceptron.setValue(State state, double newValue)
void	ValueFunctionHashMap.setValue(State state, double newValue)
void	ValueFunction.setValue(State state, double newValue)
void	ValueFunctionResidualAlgorithmPerceptron.setValue(State thisState, State nextState, double newValue, double discountFactor)
void	ValueFunctionResidualAlgorithmLinear.setValue(State thisState, State nextState, double newValue, double discountFactor)

Constructors in net.pakl.rl with parameters of type State

StateActionPair(State s, Action a)

Uses of State in net.pakl.rl.maze

Subclasses of State in net.pakl.rl.maze

class

State2D This class represents a state which corresponds to a position in a two-dimensional World

Methods in net.pakl.rl.maze that return State

State	MazeWorld.getNewState(State oldState, Action action) This critical method returns the new state given an action from an old state, and in this case, simply adds state to position to simulate movement.
State	MazeWorld.getRandomState()
State	MazeWorld.getStartingState()

Methods in net.pakl.rl.maze with parameters of type State

void	MazeWorld.addTeleporter(State location, State destination)
int	MazeWorld.distance(State state1, State state2)
State	MazeWorld.getNewState(State oldState, Action action) This critical method returns the new state given an action from an old state, and in this case, simply adds state to position to simulate movement.
boolean	MazeWorld.isTerminalState(State state)

Uses of State in org.eyelanguage.rl.reading

Subclasses of State in org.eyelanguage.rl.reading

class	ReadingState This class represents the State of a reading Agent -- scroll down to Field Detail to see more detailed on information for each dimension of the state.
class	ReadingStateParallelRelative In parallel states, we have to keep track of which words of the sentence have been identified already but for default compatability with earlier states we return information about the first word in the window for length, ID and whether the word is identified.
class	ReadingStateRelative Inherits from ReadingState, but several methods have been altered so that the model of the environment is word-centric instead of absolute in the sentence (this allows for generalization to new circumstances).

Methods in org.eyelanguage.rl.reading that return State

State	SentenceWorldParallel.getNewState(State oldGenericState, Action genericAction) Call the same code that is in the serial SentenceWorld newState(state, action), but also calls a new function called SentenceWorldParallel.performParallelLexicalProcessing(org.eyelanguage.rl.reading.ReadingStateParallelRelative) on the state.
State	SentenceWorld.getNewState(State oldGenericState, Action genericAction) This critical method returns the new state given an action from an old state, and in this case, simply adds state to position to simulate movement.
State	SentenceWorld.getNewState(State oldGenericState, Action genericAction, java.lang.String r)
State	SentenceWorld.getRandomState()
State	SentenceWorldParallel.getStartingState()
State	SentenceWorld.getStartingState()

Methods in org.eyelanguage.rl.reading with parameters of type State

int	SentenceWorld.distance(State state1, State state2)
java.util.List	ReadingPolicyParallel.getAllPossibleActions(State state)
java.util.List	ReadingPolicy.getAllPossibleActions(State state)
State	SentenceWorldParallel.getNewState(State oldGenericState, Action genericAction) Call the same code that is in the serial SentenceWorld newState(state, action), but also calls a new function called SentenceWorldParallel.performParallelLexicalProcessing(org.eyelanguage.rl.reading.ReadingStateParallelRelative) on the state.
State	SentenceWorld.getNewState(State oldGenericState, Action genericAction) This critical method returns the new state given an action from an old state, and in this case, simply adds state to position to simulate movement.
State	SentenceWorld.getNewState(State oldGenericState, Action genericAction, java.lang.String r)
double	ReadingReinforcementFunction.getReward(State state, Action action)
double	ReadingParallelReinforcementFunction.getReward(State state, Action action)
double	ReadingReinforcementFunction.getReward(State state, Action action, State resultingState)
double	ReadingParallelReinforcementFunction.getReward(State state, Action action, State resultingState)
double	ParallelToSerialVFAdapter.getValue(State state)
void	SentenceWorld.growBounds(State state)
boolean	SentenceWorldParallel.isTerminalState(State state) The final word in the sentence, and all preceding words, must be identified.
boolean	SentenceWorld.isTerminalState(State state)
void	ParallelToSerialVFAdapter.setValue(State state, double newValue)