from operator import attrgetter
from pygenalgo.genome.chromosome import Chromosome
from pygenalgo.operators.genetic_operator import increase_counter
from pygenalgo.operators.selection.select_operator import SelectionOperator
[docs]
class TruncationSelector(SelectionOperator):
"""
Description:
Truncation Selection is a straightforward method where only a predetermined
percentage of the best individuals are selected based on fitness, while the
rest are discarded. This ensures a consistent evolution of stronger solutions,
accelerating convergence toward optimal fitness levels. While effective in
honing in on high-quality individuals, truncation selection can create a loss
of genetic diversity, risking premature convergence if the population becomes
too homogeneous. The selection proportion can be adjusted to control the pressure,
making this method adaptable for various scenarios but necessitating monitoring to
retain essential diversity.
"""
def __init__(self, select_probability: float = 1.0, p: float = 0.3) -> None:
"""
Construct a 'TruncationSelector' object with a given probability value.
:param select_probability: (float) in [0, 1].
:param p: proportion of the population that will reproduce (float).
"""
# Call the super constructor with the provided probability value.
super().__init__(select_probability)
# The proportion value should be in [0.1, 0.9].
self._items = max(min(float(p), 0.9), 0.1)
# _end_def_
[docs]
@increase_counter
def select(self, population: list[Chromosome]) -> list[Chromosome]:
"""
Select the individuals, from the input population, that will be
passed on to the next genetic operations of crossover and mutation
to form the new population of solutions.
:param population: a list of chromosomes to select the parents from.
:return: the selected parents population (as list of chromosomes).
"""
# Get the population size.
pop_size = len(population)
# Sort the population in descending order using their fitness value.
sorted_population = sorted(population, key=attrgetter("fitness"), reverse=True)
# Select tne new parents using only the higher percentage '%' of
# the old population (indexes).
index = self.rng.choice(int(pop_size*self._items), size=pop_size,
replace=True, shuffle=True)
# Return the new parents (individuals).
return [sorted_population[i] for i in index]
# _end_def_
# _end_class_