Source code for metaheuristic_designer.initializers.seed_initializer

"""
Initializers that insert predefined solutions into the population.
"""

from __future__ import annotations
from typing import Iterable, Optional
from ..population import Population
from ..utils import MatrixLike, VectorLike
from ..initializer import Initializer




class SeedProbInitializer(Initializer):
    """
    Initializer that inserts a predefined solution with a given probability.

    With probability `insert_prob`, a randomly chosen solution from the
    provided set is used; otherwise a random individual is generated
    by the fallback initializer.

    Parameters
    ----------
    default_init : Initializer
        Fallback initializer for random individuals.
    solutions : Population, Iterable[VectorLike] or MatrixLike
        Set of predefined solutions to draw from.
    insert_prob : float, optional
        Probability of using a predefined solution (default 0.1).
    random_state : RNGLike, optional
        Random number generator.
    """

    def __init__(
        self, default_init: Initializer, solutions: Population | Iterable[VectorLike] | MatrixLike, insert_prob: float = 0.1, random_state=None
    ):
        assert len(solutions) > 0, "The solution set should not be empty."
        if isinstance(solutions, Population):
            infered_dimension = solutions.genotype_matrix.shape[1]
        else:
            infered_dimension = solutions[0].shape[0]

        super().__init__(dimension=infered_dimension, population_size=default_init.population_size, random_state=random_state)

        self.default_init = default_init
        self.solutions = solutions
        self.insert_prob = insert_prob



    def generate_random(self) -> VectorLike:
        """Generate a random individual using the fallback initializer.

        Returns
        -------
        VectorLike
            A 1-D array generated by the fallback initializer.
        """

        return self.default_init.generate_random()




    def generate_individual(self) -> VectorLike:
        """Generate an individual, possibly replacing it with a seeded solution.

        With probability `insert_prob` a randomly chosen predefined solution
        is returned; otherwise a new random individual is created by the
        fallback initializer.

        Returns
        -------
        ndarray
            A 1-D array representing the individual.
        """

        new_indiv = None
        if self.random_state.random() < self.insert_prob:
            new_solution = self.random_state.choice(self.solutions, axis=0)
            new_indiv = new_solution
        else:
            new_indiv = self.default_init.generate_individual()

        return new_indiv






class SeedDetermInitializer(Initializer):
    """
    Initializer that inserts a fixed number of predefined solutions.

    The first `n_to_insert` individuals generated are taken from the
    solution set (cycled if necessary); the remaining are created by
    the fallback initializer.

    Parameters
    ----------
    default_init : Initializer
        Fallback initializer for random individuals.
    solutions : Population, Iterable[VectorLike] or MatrixLike
        Set of predefined solutions to draw from.
    n_to_insert : int, optional
        Exact number of predefined solutions to insert.  Defaults to
        the size of the solution set.
    random_state : RNGLike, optional
        Random number generator.
    """

    def __init__(
        self, default_init: Initializer, solutions: Population | Iterable[VectorLike] | MatrixLike, n_to_insert: int = None, random_state=None
    ):
        assert len(solutions) > 0, "The solution set should not be empty."
        if isinstance(solutions, Population):
            infered_dimension = solutions.genotype_matrix.shape[1]
        else:
            infered_dimension = solutions[0].shape[0]
        super().__init__(dimension=infered_dimension, population_size=default_init.population_size, random_state=random_state)

        self.default_init = default_init
        self.solutions = solutions

        self.number_to_insert = n_to_insert
        if n_to_insert is None:
            self.number_to_insert = len(solutions)

        self.inserted = 0



    def generate_random(self) -> VectorLike:
        """Generate a random individual using the fallback initializer.

        Returns
        -------
        VectorLike
            A 1-D array generated by the fallback initializer.
        """

        return self.default_init.generate_random()




    def generate_individual(self) -> VectorLike:
        """Return a predefined solution until the insertion quota is met.

        The first `n_to_insert` individuals are taken from the solution set
        (cycling if necessary).  After that, random individuals are generated
        by the fallback initializer.

        Returns
        -------
        VectorLike
            A 1-D array representing the individual.
        """

        new_indiv = None
        if self.inserted < self.number_to_insert:
            new_solution = self.solutions[self.inserted % len(self.solutions)]
            new_indiv = new_solution
        else:
            new_indiv = self.default_init.generate_individual()

        self.inserted += 1
        return new_indiv




    def generate_population(self, objfunc, n_individuals: Optional[int] = None) -> Population:
        """Create the population, resetting the insertion counter first.

        Parameters
        ----------
        objfunc : ObjectiveFunc
            The objective function.
        n_individuals : int, optional
            Number of individuals to generate.  Defaults to
            :attr:`population_size`.

        Returns
        -------
        Population
            A population with the predefined solutions inserted at the
            beginning.
        """

        self.inserted = 0

        return super().generate_population(objfunc, n_individuals)