cma.restricted_gaussian_sampler.GaussVkDSampler

class documentation

class GaussVkDSampler(StatisticalModelSamplerWithZeroMeanBaseClass):

Constructor: GaussVkDSampler(dimension, randn, kadapt, **kwargs)

Restricted Gaussian Sampler for VkD-CMA O(N*k^2 + k^3) Time/Space Variant of CMA-ES with C = D * (I + V * V^T) * D

References

[1] Youhei Akimoto and Nikolaus Hansen. Online Model Selection for Restricted Covariance Matrix Adaptation. In Proc. of PPSN 2016, pp. 3--13 (2016) [2] Youhei Akimoto and Nikolaus Hansen. Projection-Based Restricted Covariance Matrix Adaptation for High Dimension. In Proc. of GECCO 2016, pp. 197--204 (2016)

Static Method	`extend_cma_options`	return correct options to run `cma.fmin` or initialize `cma.CMAEvolutionStrategy` using the `GaussVkDSampler` AKA VkD-CMA-ES
Method	`__imul__`	Undocumented
Method	`__init__`	pass dimension of the underlying sample space
Method	`get_condition_numbers`	get the condition numbers of D**2 and (I + VV')
Method	`inverse_hessian_scalar_correction`	return scalar correction `alpha` such that `X` and `f` fit to `f(x) = (x-mean) (alpha * C)**-1 (x-mean)`
Method	`norm`	return Mahalanobis norm of `x` w.r.t. the statistical model
Method	`parameters`	return `dict` with (default) parameters, e.g., `c1` and `cmu`.
Method	`parameters_old`	return `dict` with (default) parameters, e.g., `c1` and `cmu`.
Method	`sample`	return list of i.i.d. samples.
Method	`to_linear_transformation`	return associated linear transformation
Method	`to_linear_transformation_inverse`	return inverse of associated linear transformation
Method	`transform`	transform `x` as implied from the distribution parameters
Method	`transform_inverse`	Undocumented
Method	`update`	`vectors` is a list of samples, `weights` a corrsponding list of learning rates
Instance Variable	`accepted_slowdown`	Undocumented
Instance Variable	`cs`	Undocumented
Instance Variable	`D`	Undocumented
Instance Variable	`ds`	Undocumented
Instance Variable	`dx`	Undocumented
Instance Variable	`ema_log_d`	Undocumented
Instance Variable	`ema_log_s`	Undocumented
Instance Variable	`ema_log_sigma`	Undocumented
Instance Variable	`factor_diag_slope`	Undocumented
Instance Variable	`factor_sigma_slope`	Undocumented
Instance Variable	`flg_injection`	Undocumented
Instance Variable	`itr_after_k_inc`	Undocumented
Instance Variable	`k`	Undocumented
Instance Variable	`k_active`	Undocumented
Instance Variable	`k_adapt_decay`	Undocumented
Instance Variable	`k_adapt_factor`	Undocumented
Instance Variable	`k_adapt_wait`	Undocumented
Instance Variable	`k_dec_cond`	Undocumented
Instance Variable	`k_inc_cond`	Undocumented
Instance Variable	`kadapt`	Undocumented
Instance Variable	`kmax`	Undocumented
Instance Variable	`kmin`	Undocumented
Instance Variable	`last_log_cond_corr`	Undocumented
Instance Variable	`last_log_d`	Undocumented
Instance Variable	`last_log_sigma`	Undocumented
Instance Variable	`lndiag_change`	Undocumented
Instance Variable	`lnlambda_change`	Undocumented
Instance Variable	`lnsigma_change`	Undocumented
Instance Variable	`N`	Undocumented
Instance Variable	`opt_conv`	Undocumented
Instance Variable	`pc`	Undocumented
Instance Variable	`ps`	Undocumented
Instance Variable	`randn`	Undocumented
Instance Variable	`S`	Undocumented
Instance Variable	`sigma`	Undocumented
Instance Variable	`sigma_fac`	Undocumented
Instance Variable	`U`	Undocumented
Instance Variable	`V`	Undocumented
Instance Variable	`weights`	Undocumented
Property	`condition_number`	Undocumented
Property	`correlation_matrix`	Undocumented
Property	`covariance_matrix`	Undocumented
Property	`variances`	vector of coordinate-wise (marginal) variances
Method	`_get_log_determinant_of_cov`	Undocumented
Method	`_get_params`	Return the learning rate cone, cmu, cc depending on k
Method	`_get_params2`	Undocumented
Instance Variable	`_debug`	Undocumented
Instance Variable	`_mueff`	Undocumented
Instance Variable	`_parameters`	Undocumented

Inherited from StatisticalModelSamplerWithZeroMeanBaseClass:

Instance Variable _lam Undocumented

@staticmethod
def extend_cma_options(opts=None): ¶

return correct options to run cma.fmin or initialize cma.CMAEvolutionStrategy using the GaussVkDSampler AKA VkD-CMA-ES

def __imul__(self, factor): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.__imul__

Undocumented

def __init__(self, dimension, randn=np.random.randn, kadapt=True, **kwargs): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.__init__

pass dimension of the underlying sample space

def get_condition_numbers(self): ¶

get the condition numbers of D**2 and (I + VV')

Theoretically, the condition number of the covariance matrix can be at most the product of the return values. It might be safe to stop a run if the product of the return values reaches 1e14.

Returns

float: condition number of D
float: condition number of I + VV'

def inverse_hessian_scalar_correction(self, mean, X, f): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.inverse_hessian_scalar_correction

return scalar correction alpha such that X and f fit to f(x) = (x-mean) (alpha * C)**-1 (x-mean)

def norm(self, x): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.norm

return Mahalanobis norm of x w.r.t. the statistical model

def parameters(self, mueff=None, **kwargs): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.parameters

return dict with (default) parameters, e.g., c1 and cmu.

See Also
`RecombinationWeights`

def parameters_old(self, weights): ¶

return dict with (default) parameters, e.g., c1 and cmu.

See Also
`RecombinationWeights`

def sample(self, number, update=None): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.sample

return list of i.i.d. samples.

Parameters
number	is the number of samples.
update	controls a possibly lazy update of the sampler.

def to_linear_transformation(self, reset=False): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.to_linear_transformation

return associated linear transformation

def to_linear_transformation_inverse(self, reset=False): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.to_linear_transformation_inverse

return inverse of associated linear transformation

def transform(self, x): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.transform

transform x as implied from the distribution parameters

def transform_inverse(self, x): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.transform_inverse

Undocumented

def update(self, vectors, weights): ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.update

vectors is a list of samples, weights a corrsponding list of learning rates

accepted_slowdown = ¶

Undocumented

cs = ¶

Undocumented

D = ¶

Undocumented

ds = ¶

Undocumented

dx = ¶

Undocumented

ema_log_d = ¶

Undocumented

ema_log_s = ¶

Undocumented

ema_log_sigma = ¶

Undocumented

factor_diag_slope = ¶

Undocumented

factor_sigma_slope = ¶

Undocumented

flg_injection: bool = ¶

Undocumented

itr_after_k_inc: int = ¶

Undocumented

k = ¶

Undocumented

k_active = ¶

Undocumented

k_adapt_decay = ¶

Undocumented

k_adapt_factor = ¶

Undocumented

k_adapt_wait = ¶

Undocumented

k_dec_cond = ¶

Undocumented

k_inc_cond = ¶

Undocumented

kadapt = ¶

Undocumented

kmax = ¶

Undocumented

kmin = ¶

Undocumented

last_log_cond_corr = ¶

Undocumented

last_log_d = ¶

Undocumented

last_log_sigma = ¶

Undocumented

lndiag_change = ¶

Undocumented

lnlambda_change = ¶

Undocumented

lnsigma_change = ¶

Undocumented

N = ¶

Undocumented

opt_conv = ¶

Undocumented

pc = ¶

Undocumented

ps: float = ¶

Undocumented

randn = ¶

Undocumented

S = ¶

Undocumented

sigma: float = ¶

Undocumented

sigma_fac: float = ¶

Undocumented

U = ¶

Undocumented

V = ¶

Undocumented

weights = ¶

Undocumented

@property
condition_number = ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.condition_number

Undocumented

@property
correlation_matrix = ¶

Undocumented

@property
covariance_matrix = ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.covariance_matrix

Undocumented

@property
variances = ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass.variances

vector of coordinate-wise (marginal) variances

def _get_log_determinant_of_cov(self): ¶

Undocumented

def _get_params(self, weights, k): ¶

Return the learning rate cone, cmu, cc depending on k

Parameters

weights : list of float: the weight values for vectors used to update the distribution
k : int: the number of vectors for covariance matrix

Returns

cone, cmu, cc : float in [0, 1]. Learning rates for rank-one, rank-mu,: and the cumulation factor for rank-one.

def _get_params2(self, mueff, k): ¶

Undocumented

_debug = ¶

Undocumented

_mueff = ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass._mueff

Undocumented

_parameters = ¶

overrides cma.interfaces.StatisticalModelSamplerWithZeroMeanBaseClass._parameters

Undocumented