ActionSelector

Protocol for action selection strategies.

Defines the interface for selecting actions from environment states. Implemented by DQNAgent (greedy Q-value selection) and PPOAgent (greedy policy logit selection).

Source

Methods

get_actions

get_actions(states: np.ndarray, action_masker: ActionMasker | None = None) -> list[int]

Choose actions for a batch of states.

Parameters:

• states (ndarray) – A batch of state observations with shape (N, *state_shape).
• action_masker (ActionMasker | None) – Optional ActionMasker for filtering allowed actions.

Returns:

• list[int] – A list of selected actions.

Source