Action Noise

class stable_baselines3.common.noise.ActionNoise

The action noise base class

reset()

Call end of episode reset for the noise

Return type:

None

class stable_baselines3.common.noise.NormalActionNoise(mean, sigma, dtype=<class 'numpy.float32'>)

A Gaussian action noise.

Parameters:

• mean (ndarray) – Mean value of the noise

• sigma (ndarray) – Scale of the noise (std here)

• dtype (Union[dtype, None, type, _SupportsDType[dtype], str, Tuple[Any, int], Tuple[Any, Union[SupportsIndex, Sequence[SupportsIndex]]], List[Any], _DTypeDict, Tuple[Any, Any]]) – Type of the output noise

class stable_baselines3.common.noise.OrnsteinUhlenbeckActionNoise(mean, sigma, theta=0.15, dt=0.01, initial_noise=None, dtype=<class 'numpy.float32'>)

An Ornstein Uhlenbeck action noise, this is designed to approximate Brownian motion with friction.

Based on http://math.stackexchange.com/questions/1287634/implementing-ornstein-uhlenbeck-in-matlab

Parameters:

• mean (ndarray) – Mean of the noise

• sigma (ndarray) – Scale of the noise

• theta (float) – Rate of mean reversion

• dt (float) – Timestep for the noise

• initial_noise (Optional[ndarray]) – Initial value for the noise output, (if None: 0)

• dtype (Union[dtype, None, type, _SupportsDType[dtype], str, Tuple[Any, int], Tuple[Any, Union[SupportsIndex, Sequence[SupportsIndex]]], List[Any], _DTypeDict, Tuple[Any, Any]]) – Type of the output noise

reset()

reset the Ornstein Uhlenbeck noise, to the initial position

Return type:

None

class stable_baselines3.common.noise.VectorizedActionNoise(base_noise, n_envs)

A Vectorized action noise for parallel environments.

Parameters:

• base_noise (ActionNoise) – Noise generator to use

• n_envs (int) – Number of parallel environments

reset(indices=None)

Reset all the noise processes, or those listed in indices.

Parameters:

indices (Optional[Iterable[int]]) – The indices to reset. Default: None. If the parameter is None, then all processes are reset to their initial position.

Return type:

None