Acquisition optimizer

The module optimizes the acquisition function of the main objective of the study by a combination of a heuristic global optimization followed by a local convergence of the best result.

num_initial_samples (int)

Number of initial samples for maximizing the acquisition function.

Default: Automatic choice depending on the dimensionality of the parameter space and the required time for the client to return an observation for a given suggestion.

max_num_model_evals (int)

Maximum number of evaluations of the surrogates for finding the maximum of the acquisition function.

Default: Automatic choice depending on the dimensionality of the parameter space and the required time for the client to return an observation for a given suggestion.

adaptive_local_search (bool)

The maximization of the acquisition function consists of a global heuristic search followed by a local convergence of the best samples. By default, the local search effort adapts to the value of max_num_model_evals. In some cases far better samples are computed, if the local search is only stopped when convergence to a local minimum was obtained.

Default: true

num_training_samples (int)

Number of pseudo-random initial samples before the samples are drawn according to the acquisition function.

Default: Automatic choice depending depending on dimensionality of design space.

compute_suggestion_in_advance (bool)

If true, a suggestion is computed in advance while waiting for the next request to speed up the provision of new suggestions. The pre-computation is only used if the mismatch to the current environment of the pre-computed sample is small.

Default: true