DataPredict

API Reference - Models - ReinforcementLearningBaseModel

ReinforcementLearningBaseModel is a base class for reinforcement learning neural network models.

Constructors

new()

Creates a new base model object. If any of the arguments are nil, default argument values for that argument will be used.

ReinforcementLearningBaseModel.new(discountFactor: number): ModelObject

Parameters:

• discountFactor: The higher the value, the more likely it focuses on long-term outcomes. The value must be set between 0 and 1.

Returns:

• ModelObject: The generated model object.

Functions

setParameters()

Set model’s parameters. When any of the arguments are nil, previous argument values for that argument will be used.

ReinforcementLearningBaseModel:setParameters(discountFactor: number)

setModel()

Sets the model.

ReinforcementLearningBaseModel:setModel(Model: ModelObject)

Parameters:

• Model: The model to be used.

getModel()

Gets the model.

ReinforcementLearningBaseModel:getModel(): ModelObject

Returns:

• Model: The model that was set.

setCategoricalUpdateFunction()

Sets the model’s categorical policy update function.

ReinforcementLearningBaseModel:setCategoricalUpdateFunction(categoricalUpdateFunction)

Parameters:

• categoricalUpdateFunction: The function to run when categoricalUpdate() is called.

setDiagonalGaussianUpdateFunction()

Sets the model’s diagonal Gausian policy update function.

ReinforcementLearningBaseModel:setDiagonalGaussianUpdateFunction(diagonalGaussianUpdateFunction)

Parameters:

• diagonalGaussianUpdateFunction: The function to run when diagonalGaussianUpdate() is called.

setEpisodeUpdateFunction()

Sets the model’s episode update function.

ReinforcementLearningBaseModel:setEpisodeUpdateFunction(episodeUpdateFunction)

Parameters:

• episodeUpdateFunction: The function to run when episodeUpdate() is called.

categoricalUpdate()

Updates the model parameters using categoricalUpdateFunction().

ReinforcementLearningBaseModel:categoricalUpdate(previousFeatureVector: featureVector, action: number/string, rewardValue: number, currentFeatureVector: featureVector)

Parameters:

• previousFeatureVector: The previous state of the environment.

• action: The action selected.

• rewardValue: The reward gained at current state.

• currentFeatureVector: The currrent state of the environment.

diagonalGaussianUpdate()

Updates the model parameters using diagonalGaussianUpdateFunction().

ReinforcementLearningBaseModel:diagonalGaussianUpdate(previousFeatureVector: featureVector, actionMeanVector: vector, actionStandardDeviationVector, rewardValue: number, currentFeatureVector: featureVector)

Parameters:

• previousFeatureVector: The previous state of the environment.

• actionMeanVector: The vector containing mean values for all actions.

• actionStandardDeviationVector: The vector containing standard deviation values for all actions.

• rewardValue: The reward gained at current state.

• currentFeatureVector: The currrent state of the environment.

episodeUpdate()

Updates the model parameters using episodeUpdateFunction().

ReinforcementLearningBaseModel:episodeUpdate()

setResetFunction()

Sets a new function on reset alongside with the current reset() function.

ReinforcementLearningBaseModel:setResetFunction(resetFunction)

Parameters:

• resetFunction: The function to run when reset() is called.

reset()

Reset model’s stored values (excluding the parameters).

ReinforcementLearningBaseModel:reset()