Action

class qctrl.dynamic.types.reinforcement_learning_step.Action(*, discrete_index=None)

The action the Agent has chosen.

Variables

discrete_index (int, optional) – The index into a discrete action space. It is in the range 0 <= discreteIndex < discreteActionSpaceSize. You could use this as an index to select from a constant list of samples from a continuous action space. Note: The reason this is nullable is that in the future we may have other Agents that return actions in a continuous action space.