Início do AG

ppa/config.py

@@ -17,3 +17,16 @@ def get_config():
     config['follow_chance'] = 0.8
 
     return config
+
+
+def get_config_ga():
+    config = {}
+
+    config['population_size'] = 20
+    config['num_generations'] = 100
+
+    config['top_selection_ratio'] = 0.2
+    config['bottom_selection_ratio'] = 0.1
+    config['mutation_chance'] = 0.01
+
+    return config

ppa/ga.py

@@ -0,0 +1,49 @@
+import numpy as np
+import random
+
+from objective import fitness, fitness_population
+from config import get_config_ga
+from instance import Instance, print_instance
+
+
+# TODO(felipe:2018-06-04): Adicionar outros métodos de crossover além do two-point
+def crossover(individuals):
+    cut_point1 = random.randint(0, individuals[0].size - 1)
+    cut_point2 = random.randint(0, individuals[0].size - 1)
+
+    while cut_point2 == cut_point1:
+        if individuals[0].size == 1:
+            break
+
+        cut_point2 = random.randint(0, individuals[0].size - 1)
+
+    if cut_point1 > cut_point2:
+        cut_point1, cut_point2 = cut_point2, cut_point1
+
+    new_individual = np.empty(individuals[0])
+
+    for i in range(0, cut_point1):
+        new_individual[i] = individuals[0]
+
+    for i in range(cut_point1, cut_point2):
+        new_individual[i] = individuals[1]
+
+    for i in range(cut_point2, individuals[0].size):
+        new_individual[i] = individuals[0]
+
+    return new_individual
+
+
+instance_test = Instance.load_from_file("../instance_files/config.txt")
+print_instance(instance_test)
+print("")
+config = get_config_ga()
+
+population_size = config['population_size']
+
+population = np.random.randint(2, size=(population_size, instance_test.num_materials), dtype=bool)
+
+selection_quant = 2
+
+for generation in range(config['num_generations']):
+    pass

ppa/individual.py

@@ -0,0 +1,23 @@
+import random
+
+# TODO Mudar (talvez) representação de vetor de inteiros pra um único inteiro em binário
+
+
+class Individual:
+    def __init__(self, paramId, size):
+        self.size = size
+        self.id = paramId
+        self.prey = [0] * self.size
+
+        self.survival_value = 0.0
+
+    def generate_random_prey(self):
+        self.prey.clear()
+        for i in range(self.size):
+            self.prey.append(random.randint(0, 1))
+
+    def __str__(self):
+        returned = ''
+        for position in self.prey:
+            returned += str(position) + ' '
+        return returned + '\nSurvival value: ' + str(self.survival_value) + '\n'

ppa/population.py

@@ -0,0 +1,25 @@
+# TODO Mudar (talvez) representação de vetor de inteiros pra um único inteiro em binário
+
+
+class Population:
+    def __init__(self, individuals=None):
+        self.individuals = individuals
+
+        self.size_population = 0
+        self.best_preys_id = None
+        self.predator_id = 0
+        self.ordinary_preys_ids = None
+
+    def __str__(self):
+        returned = ''
+        for individual in self.individuals:
+            returned += 'ID = ' + str(individual.id) + '\n'
+            for position in individual.prey:
+                returned += str(position) + ' '
+            returned += '\n'
+            returned += 'Survival value = ' + str(individual.survival_value) + '\n'
+        returned += '\nBest prey = ' + str(self.best_preys_id) + '\n'
+        returned += 'Predator = ' + str(self.predator_id) + '\n'
+        returned += 'Ordinary preys = ' + str(self.ordinary_preys_ids) + '\n'
+
+        return returned

ppa/singleton_print.py

@@ -0,0 +1,21 @@
+class SingletonPrint:
+    class __SingletonPrint:
+        def __init__(self):
+            pass
+
+    __instance = None
+    __out = []
+
+    def __init__(self):
+        if not SingletonPrint.__instance:
+            SingletonPrint.__instance = SingletonPrint.__SingletonPrint()
+
+    def add_string(self, text):
+        self.__out.append(text)
+
+    def out(self):
+        for text in self.__out:
+            print(text, end='')
+
+    def free(self):
+        self.__out.clear()