API Reference - EligibilityTraces - BaseEligibilityTrace
Constructors
new()
Creates a new eligibility trace object. If any of the arguments are not given, default argument values for that argument will be used.
BaseEligibilityTrace.new({lambda: number}): EligibilityTraceObject
Parameters:
- lambda: lambda: At 0, the model acts like the Temporal Difference algorithm. At 1, the model acts as Monte Carlo algorithm. Between 0 and 1, the model acts as both. [Default: 0]
Returns:
- EligibilityTraceObject: The generated eligibility trace object.
Functions
increment()
BaseEligibilityTrace:increment(actionIndex: number, discountFactor: number, dimensionSizeArray: {number})
Parameters:
-
actionIndex: The action index to be incremented.
-
discountFactor: The discount factor to be used to modify the eligibility trace.
-
dimensionSizeArray: The dimension size array for generating the eligibility trace.
calculate()
BaseEligibilityTrace:calculate(temporalDifferenceErrorVector: tensor): tensor
Parameters:
- temporalDifferenceErrorVector: A temporal difference error vector.
Returns:
- temporalDifferenceErrorVector: A temporal difference error vector.
setCalculateFunction()
BaseEligibilityTrace:setIncrementFunction(IncrementFunction: function)
Parameters:
- IncrementFunction: The increment function to be set.
setLambda()
Set the lambda to the eligibility trace object.
BaseEligibilityTrace:setLambda(lambda: number)
Parameters:
- lambda: The regularization factor. Recommended values are between 0 to 1.
getLambda()
Get the lambda from the eligibility trace object.
BaseEligibilityTrace:getLambda(): number
Returns:
- lambda: The regularization factor.
reset()
Resets the stored eligibility traces.
BaseEligibilityTrace:reset()