EDAspy.optimization.custom package

Subpackages

Submodules

EDAspy.optimization.custom.eda_custom module

class EDAspy.optimization.custom.eda_custom.EDACustom(size_gen: int, max_iter: int, dead_iter: int, n_variables: int, alpha: float, elite_factor: float, disp: bool, pm: int, init: int, bounds: tuple)[source]

Bases: EDA

This class allows the user to define an EDA by custom. This implementation is thought to be extended and extend the methods to allow different implementations. Moreover, the probabilistic models and initializations can be combined to invent or design a custom EDA.

The class allows the user to export and load the settings of previous EDA configurations, so this favours the implementation of auto-tuning approaches, for example.

Example

This example uses some very well-known benchmarks from CEC14 conference to be solved using a custom implementation of EDAs.

from EDAspy.optimization.custom import EDACustom, GBN, UniformGenInit
from EDAspy.benchmarks import ContinuousBenchmarkingCEC14

n_variables = 10
my_eda = EDACustom(size_gen=100, max_iter=100, dead_iter=n_variables, n_variables=n_variables, alpha=0.5,
                   elite_factor=0.2, disp=True, pm=4, init=4, bounds=(-50, 50))

benchmarking = ContinuousBenchmarkingCEC14(n_variables)

my_gbn = GBN([str(i) for i in range(n_variables)])
my_init = UniformGenInit(n_variables)

my_eda.pm = my_gbn
my_eda.init = my_init

eda_result = my_eda.minimize(cost_function=benchmarking.cec14_4)
export_settings() → dict[source]

Export the settings of the EDA. :return: dictionary with the configuration. :rtype dict

EDAspy.optimization.custom.eda_custom.read_settings(settings: dict) → EDACustom[source]

This function is implemented to automatic implement the EDA custom by importing the configuration of a previous implementation. The function accepts the configuration exported from a previous EDA.

Parameters:

settings (dict) – dictionary with the previous configuration.

Returns:

EDA custom automatic built.

Return type:

EDACustom

Module contents