Search Space

gpjax.decision_making.search_space

AbstractSearchSpace dataclass

Bases: ABC

The AbstractSearchSpace class is an abstract base class for search spaces, which are used to define domains for sampling and optimisation functionality in GPJax.

dimensionality: int abstractmethod property

Dimensionality of the search space. Returns: int: Dimensionality of the search space.

sample(num_points: int, key: KeyArray) -> Float[Array, 'N D'] abstractmethod

Sample points from the search space. Args: num_points (int): Number of points to be sampled from the search space. key (KeyArray): JAX PRNG key. Returns: Float[Array, "N D"]: num_points points sampled from the search space.

ContinuousSearchSpace dataclass

Bases: AbstractSearchSpace

The ContinuousSearchSpace class is used to bound the domain of continuous real functions of dimension $D$.

lower_bounds: Float[Array, ' D'] instance-attribute

upper_bounds: Float[Array, ' D'] instance-attribute

dimensionality: int property

__post_init__()

sample(num_points: int, key: KeyArray) -> Float[Array, 'N D']

Sample points from the search space using a Halton sequence.

Parameters:

Name	Type	Description	Default
num_points	int	Number of points to be sampled from the search space.	required
key	KeyArray	JAX PRNG key.	required

Returns: Float[Array, "N D"]: num_points points sampled using the Halton sequence from the search space.