from hyponic.optimizers.base_optimizer import BaseOptimizer
import numpy as np
import numexpr as ne
[docs]class GA(BaseOptimizer):
"""
Genetic Algorithm(GA)
Example
~~~~~~~
>>> from hyponic.optimizers.genetic_based.GA import GA
>>> import numpy as np
>>>
>>> def sphere(x):
>>> return np.sum(x ** 2)
>>>
>>> problem_dict = {
>>> 'fit_func': sphere,
>>> 'lb': [-5.12, -5, -14, -6, -0.9],
>>> 'ub': [5.12, 5, 14, 6, 0.9],
>>> 'minmax': 'min'
>>> }
>>>
>>> ga = GA(epoch=40, population_size=100, verbose=True, early_stopping=4)
>>> ga.solve(problem_dict)
>>> print(ga.get_best_score())
>>> print(ga.get_best_solution())
"""
def __init__(self, epoch: int = 10, population_size: int = 10, minmax: str = None,
verbose: bool = False, mode: str = 'single', n_workers: int = 4, early_stopping: int | None = None,
**kwargs):
"""
:param epoch: number of iterations
:param population_size: number of individuals in the population
:param minmax: 'min' or 'max', depending on whether the problem is a minimization or maximization problem
:param verbose: whether to print the progress, default is False
:param mode: 'single' or 'multithread', depending on whether to use multithreading or not
:param n_workers: number of workers to use in multithreading mode
:param early_stopping: number of epochs to wait before stopping the optimization process. If None, then early
stopping is not used
"""
super().__init__(**kwargs)
self.epoch = epoch
self.population_size = population_size
self.minmax = minmax
self.verbose = verbose
self.mode = mode
self.n_workers = n_workers
self.early_stopping = early_stopping
self.population = None
self.scores = None
self.best_score = None
self.best_solution = None
[docs] def initialize(self, problem_dict):
super().initialize(problem_dict)
self.population = np.random.uniform(low=self.lb, high=self.ub, size=(self.population_size, self.dimensions))
self.scores = np.array([self.function(self.population[i]) for i in range(self.population_size)])
best_idx = self._argminmax()(self.scores)
self.best_score = self.scores[best_idx]
self.best_solution = self.population[best_idx]
[docs] def evolve(self, epoch):
next_population = np.zeros_like(self.population)
next_scores = np.zeros(self.population_size)
# Elitism: keep the best solution from the previous generation
best_idx = self._argminmax()(self.scores)
next_population[0] = self.population[best_idx]
next_scores[0] = self.scores[best_idx]
# Roulette Wheel Selection
fitness = self.scores - np.min(self.scores) if self.minmax == 'min' else np.max(self.scores) - self.scores
total_fitness = np.sum(fitness)
if total_fitness == 0:
probs = np.ones(self.population_size) / self.population_size
else:
probs = fitness / total_fitness
cum_probs = np.cumsum(probs)
# Crossover and mutation
for i in range(1, self.population_size):
# Select two parents using roulette wheel selection
parent1_idx = np.searchsorted(cum_probs, np.random.rand())
parent2_idx = np.searchsorted(cum_probs, np.random.rand())
# Single point crossover
crossover_point = np.random.randint(1, self.dimensions)
next_population[i, :crossover_point] = self.population[parent1_idx, :crossover_point]
next_population[i, crossover_point:] = self.population[parent2_idx, crossover_point:]
# Mutation
mutation_strength = ne.evaluate("0.5 * (ub - lb)", local_dict={"ub": self.ub, "lb": self.lb})
next_population[i] += np.random.normal(0, mutation_strength, size=self.dimensions)
next_population[i] = np.clip(next_population[i], self.lb, self.ub)
# evaluate the new population
next_scores = np.array([self.function(next_population[i]) for i in range(self.population_size)])
# update the best solution and score
best_idx = self._argminmax()(next_scores)
if self._minmax()(next_scores) < self._minmax()(self.scores):
self.best_solution = next_population[best_idx]
self.best_score = next_scores[best_idx]
# replace the old population with the new one
self.population = next_population
self.scores = next_scores
[docs] def get_best_score(self):
return self.best_score
[docs] def get_best_solution(self):
return self.best_solution
[docs] def get_current_best_score(self):
return self._minmax()(self.scores)
[docs] def get_current_best_solution(self):
return self.population[self._argminmax()(self.scores)]