Grid Search

does this and that…

class olympus.planners.Grid(goal='minimize', levels=2, budget=None, exceed_budget=True, shuffle=False, random_seed=None)

Grid search.

Note that the number of samples grow exponentially with the number of dimensions. E.g. for a 2-dimensional parameter space, with 2 levels, the grid will contain 4 samples; for a 3-dimensional space, it will contain 8 samples; for a 6-dimensional space, 64 samples.

Parameters

• goal (str) – The optimization goal, either ‘minimize’ or ‘maximize’. Default is ‘minimize’.

• levels (int, list) – How many locations in parameter space to sample per parameter/dimension. If an integer is provided, the same number of levels will be used for each dimension. Provide a list to use different levels for different dimensions. Default is 2.

• budget (int) – Maximum number of samples you want to evaluate. From the specified budget an adequate value for the levels will be derived. Note that, if provided, the argument levels will be discarded. Default is None.

• exceed_budget (bool) – Whether to allow building the grid with more samples then budget. This means some points might not be evaluated, but ensures there will be enough grid points to run as many evaluations as defined in budget. If False, the number of grid points will be less or equal to budget; this guarantees the budget is enough to guarantee the exploration of the whole grid. Default is True.

• shuffle (bool) – Whether to randomize the order of the samples in the grid. Default is False.

• random_seed (int) – Random seed. Set a random seed for a reproducible randomization of the grid if shuffle was set to True.

Methods

tell([observations])	Provide the planner with all previous observations.
ask([return_as])	suggest new set of parameters
recommend([observations, return_as])	Consecutively executes tell and ask: tell the planner about all previous observations, and ask about the next query point.
optimize(emulator[, num_iter, verbose])	Optimizes a surface for a fixed number of iterations.
reset()	Clears the remaining samples in the grid and prepares the planner for re-initialisation.

ask(return_as=None)

suggest new set of parameters

Parameters

return_as (string) – choose data type for returned parameters allowed options (dict, array)

Returns

newly generated parameters

Return type

ParameterVector

optimize(emulator, num_iter=1, verbose=False)

Optimizes a surface for a fixed number of iterations.

Parameters

• emulator (object) – Emulator or a Surface instance to optimize over.

• num_iter (int) – Maximum number of iterations allowed.

• verbose (bool) – Whether to print information to screen.

Returns

Campaign object with information about the optimization, including all parameters

tested and measurements obtained.

Return type

campaign (Campaign)

recommend(observations=None, return_as=None)

Consecutively executes tell and ask: tell the planner about all previous observations, and ask about the next query point.

Parameters

• observations (list of ???) –

• return_as (string) – choose data type for returned parameters allowed options (dict, array)

Returns

newly generated parameters

Return type

list

reset()

Clears the remaining samples in the grid and prepares the planner for re-initialisation.

set_param_space(param_space)

Defines the parameter space over which the planner will search.

Parameters

param_space (ParameterSpace) – a ParameterSpace object defining the space over which to search.

tell(observations=<olympus.campaigns.observations.Observations object>)

Provide the planner with all previous observations.

Parameters

observations (Observations) – an Observation object containing all previous observations. This defines the history of the campaign seen by the planner. The default is None, i.e. there are no previous observations.