Greedy

Greedy acquisition function selects candidates with the highest predicted mean values. This is a pure exploitation strategy that does not consider uncertainty, making it suitable when the surrogate model is highly confident or when pure exploitation is desired.

class alf_tools.optimizer.acquisition_functions.greedy.Greedy

Bases: AcquisitionFunction

Greedy acquisition function.

The Greedy acquisition value is given by: Greedy = μ, where μ is the mean prediction. Higher Greedy values indicate more promising candidates. This is a maximising acquisition function.