Tidy code

- Don't mutate the elite
- Remove unused methods
- Remove roulette selection

src/lifecycle.py

@@ -16,7 +16,9 @@ class Lifecycle(object):
             self.iterations += 1
             elite = self.params["elite"]
             crossover = self.params["crossover"]
-            self.population.advance_generation(elite, crossover_rate=crossover, n_arena=4)
+            mutation = self.params["mutation"]
+            self.population.advance_generation(elite, crossover_rate=crossover,
+                                               n_arena=4, mutation=mutation)
             self.best_fit.append(self.population.best_fitness())
             self.average_fit.append(self.population.avg_fitness())
             if epoch > 50:
@@ -24,7 +26,6 @@ class Lifecycle(object):
                 if max(recent_best) - min(recent_best) < 0.02:
                     break
 
-
     def best_fitness(self):
         return self.population.best_fitness()
 
@@ -32,11 +33,19 @@ class Lifecycle(object):
         plt.plot(self.best_fit)
         plt.xlabel("Epoch")
         plt.ylabel("Best fitness")
+        plt.title(self.get_params())
         if show:
             plt.show()
         if location is not None:
             plt.savefig(location)
 
+    def get_params(self):
+        return_string = ("mutation: %d " % self.params['mutation'])
+        return_string += ("crossover: %d " % self.params['crossover'])
+        return_string += ("elite: %d " % self.params['elite'])
+        return_string += ("iter: %d" % self.iterations)
+        return return_string
+
     def get_csv(self):
         id = {'id': self.id}
         best_fitness = {'best_fit': self.best_fitness()}

src/main.py

@@ -1,4 +1,3 @@
-import os
 from multiprocessing import Process, Queue
 from lifecycle import Lifecycle
 from file import File
@@ -8,14 +7,16 @@ import time
 param_example = {'population_size': 10,
                  'elite': round(0.1 * 10),
                  'crossover': round(0.6 * 10),
+                 'mutation': 0,
                  'epochs': 1000,
                  'iter': 0}
 
 
-def gen_param(pop, elite, crossover, epochs):
+def gen_param(pop, elite, crossover, epochs, mutation=0):
     return {'population_size': pop,
-            'elite': round(elite * pop),
-            'crossover': round(crossover * pop),
+            'elite': round((elite / 100) * pop),
+            'crossover': round((crossover / 100) * pop),
+            'mutation': mutation,
             'epochs': epochs}
 
 
@@ -25,25 +26,23 @@ outputs = Queue()
 def run_instance(instance, output_list):
     instance.start()
     output_list.put(instance.get_csv())
+    instance.generate_graph(location="images/%d.png" % instance.id, show=False)
     print(instance.id)
 
 
-def cls():
-    os.system('cls' if os.name == 'nt' else 'clear')
-
-
 output_file = File("output.csv", ['id', 'best_fit'] + list(param_example))
-
 output_file.write_header()
 
 instances = list()
+instance_id = 0
 
-id = 0
-
-for i in [x * 1 for x in range(0, 100)]:
-    instances.append(Process(name="Thread-%d" % i, target=run_instance,
-                             args=[Lifecycle(id, gen_param(50, 0, 0.5, 500)), outputs]))
-    id += 1
+for i in range(20, 90, 10):
+    for j in range(0, 3):
+        instances.append(
+            Process(name="Thread-%d" % i, target=run_instance,
+                    args=[Lifecycle(instance_id,
+                                    gen_param(50, j, i, 500, mutation=5)), outputs]))
+        instance_id += 1
 
 start_time = time.time()
 

src/population.py

@@ -72,38 +72,32 @@ class Population(object):
 
         return sorted_members
 
-    def advance_generation(self, n_elite, crossover_rate=0.5, n_arena=4):
+    def advance_generation(self, n_elite, crossover_rate=0.5, n_arena=4, mutation=0):
         new_generation = list()
 
-        # elitism
-        for member in self.elite(n_elite):
-            new_generation.append(member)
-
-        # parent
-        # for ind in range(0, n_crossover):
-            # self.roulette_crossover(new_generation)
-
-        while len(new_generation) < len(self.members):
+        while len(new_generation) < len(self.members) - n_elite:
             if n_arena > 0:
                 x, y = self.tournament_selection(n_arena, crossover_rate)
-            else:
-                x, y = self.roulette_crossover()
 
             new_generation.append(x)
             new_generation.append(y)
 
-        self.members = new_generation
+        # Remove excess members
+        while len(new_generation) > len(self.members) - n_elite:
+            new_generation.pop()
 
-        self.mutate(10)
+        for member in new_generation:
+            if rand.random() < mutation/100:
+                member.mutate()
+
+        # elitism
+        new_generation += self.elite(n_elite)
+
+        self.members = new_generation
 
     def remove_member(self, member):
         self.members.remove(member)
 
-    def roulette_crossover(self):
-        parent_one = self.roulette()
-        parent_two = self.roulette()
-        return parent_one.crossover(parent_two)
-
     def tournament_selection(self, arena_size, rate):
         parents = list()
         for i in range(arena_size):