bolero.optimizer.ACMESOptimizer¶

class bolero.optimizer.ACMESOptimizer(initial_params=None, variance=1.0, covariance=None, n_samples_per_update=None, n_pre_samples_per_update=500, active=False, n_start_iter=100, n_train_max=None, n_iter_per_sample=1000, bounds=None, maximize=True, log_to_file=False, log_to_stdout=False, random_state=None)[source]¶

CMA-ES with ranking SVM as surrogate model.

CMA-ES with comparison-based surrogate model (ranking SVM).

For details, see

• presentation
• paper [1]

The authors of the algorithm were creative with the name: the letters of ACM-ES have the same meaning as in CMA-ES but the order of the first three letters is alphabetically to indicate that it uses a comparison-based surrogate model.

Parameters:

initial_params : array-like, shape = (n_params,), optional (default: 0s) Initial parameter vector. variance : float, optional (default: 1.0) Initial exploration variance. covariance : array-like, optional (default: None) Either a diagonal (with shape (n_params,)) or a full covariance matrix (with shape (n_params, n_params)). A full covariance can contain information about the correlation of variables. n_samples_per_update : integer, optional (default: 4+int(3*log(n_params))) Number of roll-outs that are required for a parameter update. n_pre_samples_per_update : int, optional (default: 500) Number of samples that will be compared with the surrogate model before we select the individuals for a generation active : bool, optional (default: False) Active CMA-ES (aCMA-ES) with negative weighted covariance matrix update n_start_iter : int, optional (default: 100) Number of iterations before we use the surrogate model n_train_max : int, optional (default: 40 + int(4 * n_params ** 1.7)) Maximum number of training samples for surrogate model n_iter_per_sample : int, optional (default: 50000) Number of iterations per sample for the quadratic programming solver for surrogate model bounds : array-like, shape (n_samples, 2), optional (default: None) Upper and lower bounds for each parameter. maximize : optional, boolean (default: True) Maximize return or minimize cost? log_to_file: optional, boolean or string (default: False) Log results to given file, it will be located in the $BL_LOG_PATH log_to_stdout: optional, boolean (default: False) Log to standard output random_state : int or RandomState, optional (default: None) Seed for the random number generator or RandomState object.

References

[1]	(1, 2) Loshchilov, I.; Schoenauer, M.; Sebag, M. Comparison-Based Optimizers Need Comparison-Based Surrogates, Parallel Problem Solving from Nature, 2010.

__init__(initial_params=None, variance=1.0, covariance=None, n_samples_per_update=None, n_pre_samples_per_update=500, active=False, n_start_iter=100, n_train_max=None, n_iter_per_sample=1000, bounds=None, maximize=True, log_to_file=False, log_to_stdout=False, random_state=None)[source]¶

get_args()¶

Get parameters for this estimator.

Returns:	params : mapping of string to any Parameter names mapped to their values.

get_best_parameters(method='best')[source]¶

Get the best parameters.

Parameters:	method : string, optional (default: ‘best’) Either ‘best’ or ‘mean’
Returns:	best_params : array-like, shape (n_params,) Best parameters