physbo.search.score_multi module

physbo.search.score_multi.EHVI(fmean, fstd, pareto)[ソース]

Calculate Expected Hyper-Volume Improvement (EHVI).

Reference: (Couckuyt et al., 2014) Fast calculation of multiobjective probability of improvement and expected improvement criteria for Pareto optimization

パラメータ:
  • fmean (numpy.ndarray) -- mean of the predictive distribution

  • fstd (numpy.ndarray) -- standard deviation of the predictive distribution

  • pareto (Pareto object) -- Pareto object

戻り値:

score -- score of the candidate

戻り値の型:

numpy.ndarray

physbo.search.score_multi.HVPI(fmean, fstd, pareto)[ソース]

Calculate Hypervolume-based Probability of Improvement (HVPI).

Reference: (Couckuyt et al., 2014) Fast calculation of multiobjective probability of improvement and expected improvement criteria for Pareto optimization

パラメータ:
  • fmean (numpy.ndarray) -- mean of the predictive distribution

  • fstd (numpy.ndarray) -- standard deviation of the predictive distribution

  • pareto (Pareto object) -- Pareto object

戻り値:

score -- score of the candidate

戻り値の型:

numpy.ndarray

physbo.search.score_multi.TS(predictor_list, training_list, test, alpha=1, reduced_candidate_num=None)[ソース]

Thompson Sampling for multi-objective optimization.

パラメータ:
  • predictor_list (list of Predictor) -- List of predictors for each objective.

  • training_list (list of Variable) -- List of training data for each objective.

  • test (Variable) -- Test points.

  • alpha (float, optional) -- Scaling parameter for sampling from posterior distribution. Default is 1.

  • reduced_candidate_num (int, optional) -- Number of candidates to randomly select for Pareto front calculation. If None or less than number of test points, all points are used. Default is None.

戻り値:

score_res -- Score array with 1 at the chosen index and 0 elsewhere.

戻り値の型:

numpy.ndarray

physbo.search.score_multi.score(mode, predictor_list, test, training_list, **kwargs)[ソース]