def es(population,x_train, y_train,x_validate, y_validate, x_test, y_test):
mut = 0.3
cross = 0.7
#fitness_all = ft.calculate_pop_ft(population, x_train, y_train, x_test, y_test)
best_fit = ft.find_best(population,x_train, y_train, x_validate, y_validate)
num_generations = 50
decendants = []
for i in range(num_generations):
number_decendants = 0
decendants = []
while number_decendants < (len(population)+5):
#print(population)
father = np.random.random_integers(0,len(population)-1)
mother = np.random.random_integers(0,len(population)-1)
mother2 = np.random.random_integers(0,len(population)-1)
father = population[father]
mother = population[mother]
mother2 = population[mother2]
monft = ft.fitness(mother, x_train, y_train, x_validate, y_validate)
monft2 = ft.fitness(mother2, x_train, y_train, x_validate, y_validate)
if monft[0] < monft2[0]:
mother = mother2
crossed = crossover(father,mother,cross)
mutated = mutation(crossed,mut)
#print(crossed)
decendants.append(mutated)
number_decendants += 1
ft_decendants,model_decendants = ft.calculate_pop_ft(decendants, x_train, y_train, x_validate, y_validate)
new_population,ft_new_pop,model_new = ft.fetch_n_better(decendants,ft_decendants,model_decendants,len(population))
ft_best = (new_population[0],(ft_new_pop[0],model_new[0]))
population = new_population
if ft_best[1][0] > best_fit[1][0]:
best_fit = ft_best
ee_accu = ft.fitness_best(best_fit[0], best_fit[1][1],x_test,y_test)
return ee_accu
def ga(population,x_train, y_train,x_validate, y_validate, x_test, y_test):
mut = 0.3
cross = 0.7
best = ft.find_best(population, x_train, y_train, x_validate, y_validate)
#treshold = 0.5
new_population = []
#parents_mated = []
num_generations = 50
for k in range(num_generations):
fitness_all, models_all = ft.calculate_pop_ft(population,x_train, y_train,x_validate, y_validate)
for j in range(int((len(population)/2))):
parents =roulette_wheel_selection(fitness_all,2,len(population))
parents = (population[parents[0]],population[parents[1]])
offspring = crossover(parents[0],parents[1],cross)
offspring2 = crossover(parents[0],parents[1],cross)
offspring = mutation(offspring,mut)
offspring2 = mutation(offspring2,mut)
new_population.append(offspring)
new_population.append(offspring2)
#parents_mated.append(parents[0])
#parents_mated.append(parents[1])
#new_population = np.concatenate((new_population,parents_mated[:int((len(population)/2))]))
population = new_population
gen_best = ft.find_best(population, x_train, y_train, x_validate, y_validate)
if gen_best[1][0] > best[1][0]:
best = gen_best
#------- Clean aux Lists------
#parents_mated = []
new_population = []
#best = ft.find_best(population,x_train, y_train, x_test, y_test)
best = ft.fitness_best(best[0],best[1][1],x_test,y_test)
return best
def de(population,x_train, y_train,x_validate, y_validate, x_test, y_test):
fitness_all, models_all = ft.calculate_pop_ft(population, x_train, y_train, x_validate, y_validate)
weight_dimention = (1200)
mut_constant = 0.5
limiar_crossover = 0.6
best_idx = 0
generations = 50
for n in range(generations):
for i in range(len(population)): #Iterate over all cromossomes
indices = [indices for indices in range(len(population)) if indices != i]
a, b, c = population[np.random.choice(indices, 3, replace=False)]
mutant = np.clip(a + mut_constant * (b - c), -1, 1)
cross_points = np.random.uniform(low = 0, high= 1,size = weight_dimention) < limiar_crossover
trial = np.where(cross_points, mutant, population[i])
trial_ft, trial_model = ft.fitness(trial,x_train, y_train, x_test, y_test)
if trial_ft > fitness_all[i]:
fitness_all[i] = trial_ft
population[i] = trial_ft
if trial_ft > fitness_all[best_idx]:
best_idx = i
#best_weight = trial
return fitness_all[best_idx]
def bfo(population, x_train, y_train, x_test, y_test, x_validate, y_validate):
#print("BFO")
Celln = 20
Ned = 10
Nre = 10
Nc = 20
Ns = 10
Ped = 0.25
StepSize = 0.1
best_fit = ft.find_best(population, x_train, y_train, x_validate,
y_validate)
for l in range(Ned):
for k in range(Nre):
for j in range(Nc):
population, cellHealth_all = chemotaxixAndSwim(
population, 1200, 20, Ns, StepSize, x_train, y_train,
x_validate, y_validate)
local_best = ft.find_best(population, x_train, y_train,
x_validate, y_validate)
if local_best[1][0] > best_fit[1][0]:
best_fit = local_best
#DUVIDA DEVERIA USAR O HEALTH AO INVES DO FITNESS AQUI?
all_fitness, all_models = ft.calculate_pop_ft(
population, x_train, y_train, x_validate, y_validate)
new_pop, ft_new, model_new = ft.fetch_n_better(
population, all_fitness, all_models, 10)
#DIVIDA COMO GARANTIR QUE NO FINAL A POPULAÇÃO VAI TER 20 BACTERIAS???
for cell in new_pop:
if (rd.uniform(0, 1) >= Ped):
cell = np.random.uniform(low=0, high=1.0, size=1200)
new_pop.append(cell)
population = new_pop
bfo_accu = ft.fitness_best(best_fit[0], best_fit[1][1], x_test, y_test)
return bfo_accu