garrafas

from random import Random
from time import time
from math import cos
from math import pi
from inspyred import ec
from inspyred.ec import terminators
import numpy as np
import os


# Generate Population
def generate_(random, args):
    size = args.get('num_inputs', 2)
    return [random.randint(0, 800) for i in range(size)]


# Evaluate Fitness
def evaluate_(candidates, args):
    fitness = []
    for cs in candidates:
        fit = perform_fitness(cs[0], cs[1])
        fitness.append(fit)
    return fitness


# Perform Fitness, calculate the fitness of each candidate
def perform_fitness(L, S):
    L = np.round(L)
    S = np.round(S)

    fit = float((5*L + 4.5*S) / 7375)
    h1 = np.maximum(0, float(((6*L+5*S)/100)-60)) / 15
    h2 = np.maximum(0, float(((10*L+20*S)-15000))) / 3750
    h3 = np.maximum(0, float(L-800)) / 200
    h4 = np.maximum(0, float(S-750)) / 187.5

    fit = fit - (h1 + h2 + h3 + h4)
    return fit

# Avaliação final do melhor indivíduo(objetivo)


def solution_evaluation(L, S):
    #L = cs[0]
    #S = cs[1]

    L = np.round(L)
    S = np.round(S)

    print
    print("..RESUDO DA PRODUÇÃO DE GARRAFAS..")
    print("Lucro total:", float(5*L+4.5*S))
    print("Tempo de utilização semanal", float(10*L+20*S))
    print("Garrafas de leite:", L)
    print("Garrafas de suco:", S)


def main():
    rand = Random()
    rand.seed(int(time()))

    ea = ec.GA(rand)  # Instancia o algoritmo genético
    ea.selector = ec.selectors.tournament_selection  # Meteodo de seleção por torneio
    ea.variator = [ec.variators.uniform_crossover,
                   ec.variators.gaussian_mutation]  # Metodo cruzamento uniforme
    ea.replacer = ec.replacers.steady_state_replacement  # Metodo de substituição

    # Função que determina o critério  de parada
    # Critério de parada por geração
    ea.terminator = terminators.generation_termination

    # Função para gerar estatistica da evolução
    ea.observer = [ec.observers.stats_observer,
                   ec.observers.file_observer]  # Estatistica

    final_pop = ea.evolve(generator=generate_,
                          evaluator=evaluate_,
                          pop_size=1000,
                          maximize=True,
                          bounder=ec.Bounder(0, 800),
                          max_generations=10000,
                          num_inputs=2,
                          crossover_rate=0.25,
                          mutation_rate=0.25,
                          num_elites=1,
                          num_selected=2,
                          tournament_size=2,
                          statistcs_fize=open("statistics.csv", "w"),
                          individuals_file=open("individuals.csv", "w")
                          )

    final_pop.sort(reverse=True)
    #print(final_pop[0])

    perform_fitness(final_pop[0].candidate[0], final_pop[1].candidate[1])
    solution_evaluation(final_pop[0].candidate[0], final_pop[1].candidate[1])


main()