burlap.behavior.learningrate

Interface LearningRate

All Known Implementing Classes:

ConstantLR, ExponentialDecayLR, FourierBasisLearningRateWrapper, SoftTimeInverseDecayLR

public interface LearningRate

Provides an interface for different methods of learning rate decay schedules. Learning rates can be defined either by state-action pairs, or by state(-action) features. The learnign rate can also be peeked at without affecting its value using the peek methods. Poll methods will cause the learning rate to decay. Note that the poll methods also take as a parmater the agent time at which the learning rate is being polled. This is because it's possible for the agent to poll the learning rate for multiple state-actions/features in any given time step, but if the learning rate function is indepdent of the states/features, then it should only be decayed once for that time index.

Author:

James MacGlashan

Method Summary

Modifier and Type	Method and Description
double	peekAtLearningRate(int featureId) A method for looking at the current learning rate for a state (-action) feature without having it altered.
double	peekAtLearningRate(State s, Action ga) A method for looking at the current learning rate for a state-action pair without having it altered.
double	pollLearningRate(int agentTime, int featureId) A method for returning the learning rate for a given state (-action) feature and then decaying the learning rate as defined by this class.
double	pollLearningRate(int agentTime, State s, Action ga) A method for returning the learning rate for a given state action pair and then decaying the learning rate as defined by this class.
void	resetDecay() Causes any learnign rate decay to reset to where it started.

Method Detail

peekAtLearningRate

double peekAtLearningRate(State s,
                          Action ga)

A method for looking at the current learning rate for a state-action pair without having it altered.

Parameters:

s - the state for which the learning rate should be returned

ga - the action from which the learning rate should be returned

Returns:

the current learning rate for the given state-action pair

pollLearningRate

double pollLearningRate(int agentTime,
                        State s,
                        Action ga)

A method for returning the learning rate for a given state action pair and then decaying the learning rate as defined by this class.

Parameters:

agentTime - the time index of the agent when polling.

s - the state for which the learning rate should be returned

ga - the action from which the learning rate should be returned

Returns:

the current learning rate for the given state-action pair

peekAtLearningRate

double peekAtLearningRate(int featureId)

A method for looking at the current learning rate for a state (-action) feature without having it altered.

Parameters:

featureId - the state feature for which the learning rate should be returned

Returns:

the current learning rate for the given state feature-action pair

pollLearningRate

double pollLearningRate(int agentTime,
                        int featureId)

A method for returning the learning rate for a given state (-action) feature and then decaying the learning rate as defined by this class.

Parameters:

agentTime - the time index of the agent when polling.

featureId - the state feature for which the learning rate should be returned

Returns:

the current learning rate for the given state feature-action pair

resetDecay

void resetDecay()

Causes any learnign rate decay to reset to where it started.