ray.rllib.utils.exploration.random.Random#

class ray.rllib.utils.exploration.random.Random(action_space: gymnasium.spaces.Space, *, model: ModelV2, framework: str | None, **kwargs)[source]#

A random action selector (deterministic/greedy for explore=False).

If explore=True, returns actions randomly from self.action_space (via Space.sample()). If explore=False, returns the greedy/max-likelihood action.

Methods

`__init__`	Initialize a Random Exploration object.
`before_compute_actions`	Hook for preparations before policy.compute_actions() is called.
`get_exploration_optimizer`	May add optimizer(s) to the Policy's own `optimizers`.
`get_state`	Returns the current exploration state.
`on_episode_end`	Handles necessary exploration logic at the end of an episode.
`on_episode_start`	Handles necessary exploration logic at the beginning of an episode.
`postprocess_trajectory`	Handles post-processing of done episode trajectories.
`set_state`	Sets the Exploration object's state to the given values.