def significantly_mutate(self, v: float, arity: int):
"""Randomly drawn a normal value, and redraw until it's different after discretization by the quantiles
1/arity, 2/arity, ..., (arity-1)/arity.
"""
w = v
while discretization.threshold_discretization(
[w], arity) == discretization.threshold_discretization([v], arity):
w = self.random_state.normal(0.0, 1.0)
return w
def coordinatewise_mutation(
self, parent: tp.ArrayLike, velocity: tp.ArrayLike, boolean_vector: tp.ArrayLike, arity: int
) -> tp.ArrayLike:
"""This is the anisotropic counterpart of the classical 1+1 mutations in discrete domains
with tunable intensity: it is useful for anisotropic adaptivity."""
dimension = len(parent)
boolean_vector = np.zeros(dimension, dtype=bool)
while not any(boolean_vector):
boolean_vector = self.random_state.rand(dimension) < (1.0 / dimension)
discrete_data = discretization.threshold_discretization(parent, arity=arity)
discrete_data = np.where(
boolean_vector,
discrete_data + self.random_state.choice([-1.0, 1.0], size=dimension) * velocity,
discrete_data,
)
return discretization.inverse_threshold_discretization(discrete_data)
def hardleadingones5(y: np.ndarray) -> float:
"""Leadingones, with a discretization by 5 with 4 thresholds (quantiles of Gaussian)."""
return _leadingones(discretization.threshold_discretization(y, 5))
def hardjump5(y: np.ndarray) -> float:
"""Jump, with a discretization by 5 with 4 thresholds (quantiles of Gaussian)."""
return _jump(discretization.threshold_discretization(y, 5))
def hardonemax5(y: np.ndarray) -> float:
"""Hardonemax, with a discretization by 5 with 4 thresholds (quantiles of Gaussian)."""
return _onemax(discretization.threshold_discretization(y, 5))
def hardleadingones(y: np.ndarray) -> float:
"""Leading ones, with a discretization in 2 by threshold 0 (>0 or <0)."""
return _leadingones(discretization.threshold_discretization(y))
def hardjump(y: np.ndarray) -> float:
"""Hardjump, with a discretization in 2 by threshold 0 (>0 or <0)."""
return _jump(discretization.threshold_discretization(y))
def hardonemax(y: np.ndarray) -> float:
"""Onemax, with a discretization in 2 by threshold 0 (>0 or <0)."""
return _onemax(discretization.threshold_discretization(y))
