BayesianBatchDeviationOED

class pyapprox.expdesign.BayesianBatchDeviationOED(ndesign_candidates, obs_fun, noise_std, prior_variable, out_quad_opts, in_quad_opts, qoi_fun=None, deviation_fun=<function oed_standard_deviation>, pred_risk_fun=<function oed_prediction_average>, data_risk_fun=<function oed_data_expectation>, nprocs=1, max_ncollected_obs=2, ndata_per_candidate=1)

Bases: AbstractBayesianOED

Compute open-loop OED by minimizing the deviation on the push forward of the posterior through a QoI model.

Methods Summary

compute_expected_utility(...[, return_all])	Compute the negative expected deviation in predictions of QoI
compute_utilities(ncandidates, ...)
populate()	Compute the data needed to initialize the OED algorithm.

Methods Documentation

compute_expected_utility(collected_design_indices, new_design_indices, return_all=False)

Compute the negative expected deviation in predictions of QoI

Parameters:

collected_design_indicesnp.ndarray (nobs)

The indices into the qoi vector associated with the collected observations

new_design_indicesnp.ndarray (nnew_obs)

The indices into the qoi vector associated with new design locations under consideration

return_allboolean

False - return the utilities True - used for debugging returns utilities and itermediate data useful for testing

Returns:

utilityfloat

The negative expected deviation

compute_utilities(ncandidates, collected_design_indices, new_indices, return_all)

populate()

Compute the data needed to initialize the OED algorithm.