bayes_opt.SequentialDomainReductionTransformer¶

See the Sequential Domain Reduction notebook for a complete example.

class bayes_opt.SequentialDomainReductionTransformer(gamma_osc: float = 0.7, gamma_pan: float = 1.0, eta: float = 0.9, minimum_window: ndarray | list[float] | float | dict[str, float] | None = 0.0)¶

Reduce the searchable space.

A sequential domain reduction transformer based on the work by Stander, N. and Craig, K: “On the robustness of a simple domain reduction scheme for simulation-based optimization”

Parameters:¶

gamma_osc : float, default=0.7¶

Parameter used to scale (typically dampen) oscillations.

gamma_pan : float, default=1.0¶

Parameter used to scale (typically unitary) panning.

eta : float, default=0.9¶

Zooming parameter used to shrink the region of interest.

minimum_window : float or np.ndarray or dict, default=0.0¶

Minimum window size for each parameter. If a float is provided, the same value is used for all parameters.

initialize(target_space: TargetSpace) → None¶

Initialize all of the parameters.

Parameters:¶

target_space : TargetSpace¶

TargetSpace this DomainTransformer operates on.

transform(target_space: TargetSpace) → dict¶

Transform the bounds of the target space.

Parameters:¶

target_space : TargetSpace¶

TargetSpace this DomainTransformer operates on.

Returns:¶

dict – The new bounds of each parameter.